From 0d9e844c465b889855ed7950f579956275239b63 Mon Sep 17 00:00:00 2001 From: Alejandro Acosta Date: Mon, 7 Aug 2023 14:03:26 +0100 Subject: [PATCH 1/3] Rename SYCL-BLAS to portBLAS --- .github/workflows/build-and-test.yml | 8 +- CMakeLists.txt | 90 +++++------ Dockerfile | 12 +- NOTICE | 2 +- README.md | 58 +++---- benchmark/CMakeLists.txt | 6 +- benchmark/README.md | 8 +- benchmark/bench_info.cc | 2 +- benchmark/clBench/acl/CMakeLists.txt | 6 +- benchmark/clBench/acl/blas3/gemm.cpp | 2 +- benchmark/clBench/acl/main.cpp | 2 +- benchmark/clBench/acl/utils.hpp | 2 +- benchmark/clBench/clblas/CMakeLists.txt | 6 +- benchmark/clBench/clblas/blas2/tbsv.cpp | 2 +- benchmark/clBench/clblas/blas2/trsv.cpp | 2 +- benchmark/clBench/clblas/blas3/trsm.cpp | 2 +- benchmark/clBench/clblas/main.cpp | 2 +- benchmark/clBench/clblas/utils.hpp | 2 +- benchmark/clBench/clblast/CMakeLists.txt | 4 +- benchmark/clBench/clblast/blas1/asum.cpp | 2 +- benchmark/clBench/clblast/blas1/axpy.cpp | 2 +- benchmark/clBench/clblast/blas1/dot.cpp | 2 +- benchmark/clBench/clblast/blas1/iamax.cpp | 2 +- benchmark/clBench/clblast/blas1/iamin.cpp | 2 +- benchmark/clBench/clblast/blas1/nrm2.cpp | 2 +- benchmark/clBench/clblast/blas1/scal.cpp | 2 +- benchmark/clBench/clblast/blas2/gbmv.cpp | 2 +- benchmark/clBench/clblast/blas2/gemv.cpp | 2 +- benchmark/clBench/clblast/blas2/ger.cpp | 2 +- benchmark/clBench/clblast/blas2/sbmv.cpp | 2 +- benchmark/clBench/clblast/blas2/spr.cpp | 2 +- benchmark/clBench/clblast/blas2/spr2.cpp | 2 +- benchmark/clBench/clblast/blas2/symv.cpp | 2 +- benchmark/clBench/clblast/blas2/syr.cpp | 2 +- benchmark/clBench/clblast/blas2/syr2.cpp | 2 +- benchmark/clBench/clblast/blas2/tbmv.cpp | 2 +- benchmark/clBench/clblast/blas2/trmv.cpp | 2 +- benchmark/clBench/clblast/blas2/trsv.cpp | 2 +- benchmark/clBench/clblast/blas3/gemm.cpp | 2 +- .../clBench/clblast/blas3/gemm_batched.cpp | 2 +- benchmark/clBench/clblast/blas3/symm.cpp | 2 +- benchmark/clBench/clblast/blas3/trsm.cpp | 2 +- benchmark/clBench/clblast/main.cpp | 2 +- benchmark/clBench/clblast/utils.hpp | 2 +- benchmark/clBench/clwrap.hpp | 2 +- benchmark/computecpp_version_config.h.in | 2 +- benchmark/cublas/CMakeLists.txt | 6 +- benchmark/cublas/blas1/asum.cpp | 2 +- benchmark/cublas/blas1/axpy.cpp | 2 +- benchmark/cublas/blas1/dot.cpp | 2 +- benchmark/cublas/blas1/iamax.cpp | 2 +- benchmark/cublas/blas1/iamin.cpp | 2 +- benchmark/cublas/blas1/nrm2.cpp | 2 +- benchmark/cublas/blas1/rotg.cpp | 2 +- benchmark/cublas/blas1/rotm.cpp | 2 +- benchmark/cublas/blas1/rotmg.cpp | 2 +- benchmark/cublas/blas1/scal.cpp | 2 +- benchmark/cublas/blas2/gbmv.cpp | 2 +- benchmark/cublas/blas2/gemv.cpp | 2 +- benchmark/cublas/blas2/ger.cpp | 2 +- benchmark/cublas/blas2/sbmv.cpp | 2 +- benchmark/cublas/blas2/spmv.cpp | 2 +- benchmark/cublas/blas2/spr.cpp | 2 +- benchmark/cublas/blas2/spr2.cpp | 2 +- benchmark/cublas/blas2/symv.cpp | 2 +- benchmark/cublas/blas2/syr.cpp | 2 +- benchmark/cublas/blas2/syr2.cpp | 2 +- benchmark/cublas/blas2/tbmv.cpp | 2 +- benchmark/cublas/blas2/tbsv.cpp | 2 +- benchmark/cublas/blas2/tpmv.cpp | 2 +- benchmark/cublas/blas2/tpsv.cpp | 2 +- benchmark/cublas/blas2/trmv.cpp | 2 +- benchmark/cublas/blas2/trsv.cpp | 2 +- benchmark/cublas/blas3/gemm.cpp | 2 +- benchmark/cublas/blas3/gemm_batched.cpp | 2 +- .../cublas/blas3/gemm_batched_strided.cpp | 2 +- benchmark/cublas/blas3/symm.cpp | 2 +- benchmark/cublas/blas3/syr2k.cpp | 2 +- benchmark/cublas/blas3/syrk.cpp | 2 +- benchmark/cublas/blas3/trmm.cpp | 2 +- benchmark/cublas/blas3/trsm.cpp | 2 +- benchmark/cublas/blas3/trsm_batched.cpp | 2 +- benchmark/cublas/extension/omatcopy.cpp | 2 +- benchmark/cublas/main.cpp | 4 +- benchmark/cublas/utils.hpp | 4 +- benchmark/gen_param.py | 2 +- .../{syclblas => portblas}/CMakeLists.txt | 14 +- .../{syclblas => portblas}/blas1/asum.cpp | 2 +- .../{syclblas => portblas}/blas1/axpy.cpp | 2 +- .../{syclblas => portblas}/blas1/copy.cpp | 2 +- .../{syclblas => portblas}/blas1/dot.cpp | 2 +- .../{syclblas => portblas}/blas1/iamax.cpp | 2 +- .../{syclblas => portblas}/blas1/iamin.cpp | 2 +- .../{syclblas => portblas}/blas1/nrm2.cpp | 2 +- .../{syclblas => portblas}/blas1/rotg.cpp | 2 +- .../{syclblas => portblas}/blas1/rotm.cpp | 2 +- .../{syclblas => portblas}/blas1/rotmg.cpp | 2 +- .../{syclblas => portblas}/blas1/scal.cpp | 2 +- .../{syclblas => portblas}/blas1/sdsdot.cpp | 2 +- .../{syclblas => portblas}/blas2/gbmv.cpp | 2 +- .../{syclblas => portblas}/blas2/gemv.cpp | 2 +- .../{syclblas => portblas}/blas2/ger.cpp | 2 +- .../{syclblas => portblas}/blas2/sbmv.cpp | 2 +- .../{syclblas => portblas}/blas2/spmv.cpp | 2 +- .../{syclblas => portblas}/blas2/spr.cpp | 2 +- .../{syclblas => portblas}/blas2/spr2.cpp | 2 +- .../{syclblas => portblas}/blas2/symv.cpp | 2 +- .../{syclblas => portblas}/blas2/syr.cpp | 2 +- .../{syclblas => portblas}/blas2/syr2.cpp | 2 +- .../{syclblas => portblas}/blas2/tbmv.cpp | 2 +- .../{syclblas => portblas}/blas2/tbsv.cpp | 2 +- .../{syclblas => portblas}/blas2/tpmv.cpp | 2 +- .../{syclblas => portblas}/blas2/tpsv.cpp | 2 +- .../{syclblas => portblas}/blas2/trmv.cpp | 2 +- .../{syclblas => portblas}/blas2/trsv.cpp | 2 +- .../{syclblas => portblas}/blas3/gemm.cpp | 2 +- .../blas3/gemm_batched.cpp | 2 +- .../blas3/gemm_batched_strided.cpp | 2 +- .../{syclblas => portblas}/blas3/symm.cpp | 2 +- .../{syclblas => portblas}/blas3/trsm.cpp | 2 +- .../extension/omatadd.cpp | 2 +- .../extension/omatcopy.cpp | 2 +- .../extension/omatcopy2.cpp | 2 +- .../extension/reduction.cpp | 2 +- benchmark/{syclblas => portblas}/main.cpp | 6 +- benchmark/{syclblas => portblas}/utils.hpp | 4 +- benchmark/rocblas/CMakeLists.txt | 4 +- benchmark/rocblas/blas1/asum.cpp | 2 +- benchmark/rocblas/blas1/axpy.cpp | 2 +- benchmark/rocblas/blas1/dot.cpp | 2 +- benchmark/rocblas/blas1/iamax.cpp | 2 +- benchmark/rocblas/blas1/iamin.cpp | 2 +- benchmark/rocblas/blas1/nrm2.cpp | 2 +- benchmark/rocblas/blas1/rotg.cpp | 2 +- benchmark/rocblas/blas1/rotm.cpp | 2 +- benchmark/rocblas/blas1/rotmg.cpp | 2 +- benchmark/rocblas/blas1/scal.cpp | 2 +- benchmark/rocblas/blas2/gbmv.cpp | 2 +- benchmark/rocblas/blas2/gemv.cpp | 2 +- benchmark/rocblas/blas2/ger.cpp | 2 +- benchmark/rocblas/blas2/sbmv.cpp | 2 +- benchmark/rocblas/blas2/spmv.cpp | 2 +- benchmark/rocblas/blas2/spr.cpp | 2 +- benchmark/rocblas/blas2/spr2.cpp | 2 +- benchmark/rocblas/blas2/symv.cpp | 2 +- benchmark/rocblas/blas2/syr.cpp | 2 +- benchmark/rocblas/blas2/syr2.cpp | 2 +- benchmark/rocblas/blas2/tbmv.cpp | 2 +- benchmark/rocblas/blas2/tbsv.cpp | 2 +- benchmark/rocblas/blas2/tpmv.cpp | 2 +- benchmark/rocblas/blas2/tpsv.cpp | 2 +- benchmark/rocblas/blas2/trmv.cpp | 2 +- benchmark/rocblas/blas2/trsv.cpp | 2 +- benchmark/rocblas/blas3/gemm.cpp | 2 +- benchmark/rocblas/blas3/gemm_batched.cpp | 2 +- .../rocblas/blas3/gemm_batched_strided.cpp | 2 +- benchmark/rocblas/blas3/symm.cpp | 2 +- benchmark/rocblas/blas3/syr2k.cpp | 2 +- benchmark/rocblas/blas3/syrk.cpp | 2 +- benchmark/rocblas/blas3/trmm.cpp | 2 +- benchmark/rocblas/blas3/trsm.cpp | 2 +- benchmark/rocblas/blas3/trsm_batched.cpp | 2 +- benchmark/rocblas/extension/omatcopy.cpp | 2 +- benchmark/rocblas/main.cpp | 2 +- benchmark/rocblas/utils.hpp | 4 +- cmake/CmakeFunctionHelper.cmake | 148 +++++++++--------- ...SYCLBLAS.cmake => ConfigurePORTBLAS.cmake} | 4 +- cmake/Modules/FindACL.cmake | 2 +- cmake/Modules/FindDPCPP.cmake | 2 +- cmake/Modules/FindSB_CLBlast.cmake | 2 +- cmake/Modules/FindSystemBLAS.cmake | 2 +- cmake/Modules/SYCL.cmake | 4 +- .../include/common/benchmark_identifier.hpp | 8 +- common/include/common/benchmark_names.hpp | 8 +- .../include/common/blas1_state_counters.hpp | 2 +- .../include/common/blas2_state_counters.hpp | 2 +- .../include/common/blas3_state_counters.hpp | 2 +- .../common/blas_extension_state_counters.hpp | 2 +- common/include/common/set_benchmark_label.hpp | 8 +- .../include/common/system_reference_blas.hpp | 2 +- conanfile.py | 6 +- doc/AddingBlas3Op.md | 24 +-- doc/Autotuner.md | 8 +- doc/Doxyfile | 2 +- doc/Gemm.md | 22 +-- doc/Reduction.md | 10 +- include/blas_meta.h | 14 +- include/container/sycl_iterator.h | 10 +- include/interface/blas1_interface.h | 8 +- include/interface/blas2_interface.h | 8 +- include/interface/blas3_interface.h | 8 +- include/interface/extension_interface.h | 10 +- include/interface/gemm_launcher.h | 10 +- include/operations/blas1_trees.h | 6 +- include/operations/blas2_trees.h | 10 +- include/operations/blas3_trees.h | 8 +- include/operations/blas_constants.h | 40 ++--- include/operations/blas_operators.h | 6 +- include/operations/extension/reduction.h | 8 +- include/operations/extension/transpose.h | 8 +- include/{sycl_blas.h => portblas.h} | 8 +- .../{sycl_blas_helper.h => portblas_helper.h} | 10 +- include/sb_handle/kernel_constructor.h | 8 +- .../{sycl_blas_handle.h => portblas_handle.h} | 12 +- include/views/view.h | 6 +- python_generator/py_gen_blas_binary.py | 2 +- .../py_gen_blas_binary_special.py | 2 +- python_generator/py_gen_blas_gemm_launcher.py | 2 +- python_generator/py_gen_blas_reduction.py | 2 +- python_generator/py_gen_blas_rotg.py | 2 +- python_generator/py_gen_blas_rotg_return.py | 2 +- python_generator/py_gen_blas_rotmg.py | 2 +- python_generator/py_gen_blas_ternary.py | 2 +- python_generator/py_gen_blas_unary.py | 2 +- run_docker.sh | 12 +- samples/CMakeLists.txt | 12 +- samples/FindPORTBLAS.cmake | 49 ++++++ samples/FindSyclBLAS.cmake | 49 ------ samples/README.md | 16 +- samples/gemm.cpp | 6 +- samples/gemv.cpp | 6 +- samples/symm.cpp | 6 +- src/CMakeLists.txt | 2 +- src/container/sycl_iterator.hpp | 6 +- src/interface/CMakeLists.txt | 2 +- src/interface/blas1/CMakeLists.txt | 2 +- src/interface/blas1/asum.cpp.in | 4 +- src/interface/blas1/asum_return.cpp.in | 4 +- src/interface/blas1/axpy.cpp.in | 4 +- src/interface/blas1/copy.cpp.in | 4 +- src/interface/blas1/dot.cpp.in | 4 +- src/interface/blas1/dot_return.cpp.in | 4 +- src/interface/blas1/iamax.cpp.in | 4 +- src/interface/blas1/iamax_return.cpp.in | 4 +- src/interface/blas1/iamin.cpp.in | 4 +- src/interface/blas1/iamin_return.cpp.in | 4 +- src/interface/blas1/nrm2.cpp.in | 4 +- src/interface/blas1/nrm2_return.cpp.in | 4 +- src/interface/blas1/rot.cpp.in | 4 +- src/interface/blas1/rotg.cpp.in | 4 +- src/interface/blas1/rotg_return.cpp.in | 4 +- src/interface/blas1/rotm.cpp.in | 4 +- src/interface/blas1/rotmg.cpp.in | 4 +- src/interface/blas1/scal.cpp.in | 4 +- src/interface/blas1/sdsdot.cpp.in | 4 +- src/interface/blas1/sdsdot_return.cpp.in | 4 +- src/interface/blas1/swap.cpp.in | 4 +- src/interface/blas1_interface.hpp | 8 +- src/interface/blas2/CMakeLists.txt | 2 +- src/interface/blas2/backend/amd_gpu.hpp | 6 +- src/interface/blas2/backend/backend.hpp | 2 +- src/interface/blas2/backend/default_cpu.hpp | 6 +- src/interface/blas2/backend/intel_gpu.hpp | 6 +- src/interface/blas2/backend/nvidia_gpu.hpp | 6 +- src/interface/blas2/backend/power_vr.hpp | 6 +- src/interface/blas2/gbmv.cpp.in | 4 +- src/interface/blas2/gemv.cpp.in | 4 +- src/interface/blas2/ger.cpp.in | 4 +- src/interface/blas2/sbmv.cpp.in | 4 +- src/interface/blas2/spmv.cpp.in | 4 +- src/interface/blas2/spr.cpp.in | 4 +- src/interface/blas2/spr2.cpp.in | 4 +- src/interface/blas2/symv.cpp.in | 4 +- src/interface/blas2/syr.cpp.in | 4 +- src/interface/blas2/syr2.cpp.in | 4 +- src/interface/blas2/tbmv.cpp.in | 4 +- src/interface/blas2/tbsv.cpp.in | 4 +- src/interface/blas2/tpmv.cpp.in | 4 +- src/interface/blas2/tpsv.cpp.in | 4 +- src/interface/blas2/trmv.cpp.in | 4 +- src/interface/blas2/trsv.cpp.in | 4 +- src/interface/blas2_interface.hpp | 8 +- src/interface/blas3/CMakeLists.txt | 2 +- src/interface/blas3/backend/amd_gpu.hpp | 6 +- src/interface/blas3/backend/backend.hpp | 2 +- src/interface/blas3/backend/default_cpu.hpp | 6 +- src/interface/blas3/backend/intel_gpu.hpp | 6 +- src/interface/blas3/backend/nvidia_gpu.hpp | 6 +- src/interface/blas3/backend/power_vr.hpp | 6 +- src/interface/blas3/gemm.cpp.in | 4 +- src/interface/blas3/gemm_launcher.cpp.in | 4 +- src/interface/blas3/symm.cpp.in | 4 +- src/interface/blas3/trsm.cpp.in | 4 +- src/interface/blas3_interface.hpp | 8 +- src/interface/extension/CMakeLists.txt | 2 +- src/interface/extension/backend/amd_gpu.hpp | 6 +- src/interface/extension/backend/backend.hpp | 2 +- .../extension/backend/default_cpu.hpp | 6 +- src/interface/extension/backend/intel_gpu.hpp | 6 +- .../extension/backend/nvidia_gpu.hpp | 6 +- src/interface/extension/matcopy.cpp.in | 4 +- src/interface/extension/reduction.cpp.in | 4 +- src/interface/extension/transpose.cpp.in | 4 +- src/interface/extension_interface.hpp | 12 +- src/interface/gemm_interface.hpp | 12 +- src/interface/gemm_launcher.hpp | 8 +- src/interface/symm_interface.hpp | 8 +- src/interface/trsm_interface.hpp | 12 +- src/operations/blas1_trees.hpp | 130 +++++++-------- src/operations/blas2/gbmv.hpp | 14 +- src/operations/blas2/gemv.hpp | 60 +++---- src/operations/blas2/ger.hpp | 34 ++-- src/operations/blas2/sbmv.hpp | 14 +- src/operations/blas2/spr.hpp | 12 +- src/operations/blas2/tbmv.hpp | 14 +- src/operations/blas2/txsv.hpp | 12 +- src/operations/blas2/xpmv.hpp | 12 +- src/operations/blas2_trees.hpp | 6 +- src/operations/blas3/gemm_common.hpp | 14 +- src/operations/blas3/gemm_interleaved.hpp | 44 +++--- src/operations/blas3/gemm_load_store.hpp | 20 +-- src/operations/blas3/gemm_local.hpp | 66 ++++---- .../blas3/gemm_local_joint_matrix.hpp | 52 +++--- .../blas3/gemm_no_local_full_vec.hpp | 58 +++---- .../blas3/gemm_no_local_partial_vec.hpp | 46 +++--- src/operations/blas3/gemm_partial_local.hpp | 24 +-- src/operations/blas3/gemm_ref.hpp | 28 ++-- src/operations/blas3/trsm.hpp | 18 +-- src/operations/blas3_trees.hpp | 8 +- src/operations/blas_constants.hpp | 6 +- src/operations/blas_operators.hpp | 82 +++++----- src/operations/extension/reduction.hpp | 22 +-- src/operations/extension/transpose.hpp | 24 +-- src/{sycl_blas.hpp => portblas.hpp} | 8 +- src/sb_handle/kernel_constructor.hpp | 28 ++-- ...cl_blas_handle.hpp => portblas_handle.hpp} | 14 +- src/views/view.hpp | 8 +- src/views/view_sycl.hpp | 90 +++++------ test/CMakeLists.txt | 4 +- test/blas_test.hpp | 6 +- test/blas_test_macros.hpp | 2 +- test/exprtest/CMakeLists.txt | 16 +- test/exprtest/blas1_axpy_copy_test.cpp | 6 +- test/exprtest/blas1_iface_test.cpp | 2 +- test/exprtest/blas1_scal_asum_test.cpp | 6 +- test/exprtest/collapse_nested_tuple.cpp | 4 +- test/exprtest/main.cpp | 2 +- test/unittest/CMakeLists.txt | 96 ++++++------ test/unittest/blas1/blas1_asum_test.cpp | 2 +- test/unittest/blas1/blas1_axpy_test.cpp | 2 +- test/unittest/blas1/blas1_copy_test.cpp | 2 +- test/unittest/blas1/blas1_dot_test.cpp | 2 +- test/unittest/blas1/blas1_iamin_test.cpp | 2 +- test/unittest/blas1/blas1_iaminmax_common.hpp | 2 +- test/unittest/blas1/blas1_nrm2_test.cpp | 2 +- test/unittest/blas1/blas1_rot_test.cpp | 2 +- test/unittest/blas1/blas1_rotg_test.cpp | 2 +- test/unittest/blas1/blas1_rotm_test.cpp | 2 +- test/unittest/blas1/blas1_rotmg_test.cpp | 8 +- test/unittest/blas1/blas1_scal_test.cpp | 2 +- test/unittest/blas1/blas1_sdsdot_test.cpp | 2 +- test/unittest/blas1/blas1_swap_test.cpp | 2 +- test/unittest/blas2/blas2_gbmv_test.cpp | 2 +- test/unittest/blas2/blas2_gemv_test.cpp | 2 +- test/unittest/blas2/blas2_ger_test.cpp | 2 +- test/unittest/blas2/blas2_sbmv_test.cpp | 2 +- test/unittest/blas2/blas2_spmv_test.cpp | 2 +- test/unittest/blas2/blas2_spr2_test.cpp | 2 +- test/unittest/blas2/blas2_spr_test.cpp | 2 +- test/unittest/blas2/blas2_symv_test.cpp | 2 +- test/unittest/blas2/blas2_syr2_test.cpp | 2 +- test/unittest/blas2/blas2_syr_test.cpp | 2 +- test/unittest/blas2/blas2_tbmv_test.cpp | 2 +- test/unittest/blas2/blas2_tbsv_test.cpp | 2 +- test/unittest/blas2/blas2_tpmv_test.cpp | 2 +- test/unittest/blas2/blas2_tpsv_test.cpp | 2 +- test/unittest/blas2/blas2_trmv_test.cpp | 2 +- test/unittest/blas2/blas2_trsv_test.cpp | 2 +- .../blas3/blas3_gemm_batched_test.cpp | 2 +- test/unittest/blas3/blas3_gemm_common.hpp | 6 +- .../blas3/blas3_gemm_tall_skinny_test.cpp | 2 +- test/unittest/blas3/blas3_gemm_test.cpp | 2 +- test/unittest/blas3/blas3_symm_test.cpp | 4 +- test/unittest/blas3/blas3_trsm_test.cpp | 2 +- test/unittest/buffers/sycl_buffer_test.cpp | 2 +- .../extension/extension_reference.hpp | 6 +- test/unittest/extension/omatcopy2_test.cpp | 2 +- test/unittest/extension/omatcopy_test.cpp | 2 +- test/unittest/extension/reduction_test.cpp | 2 +- test/unittest/extension/transpose_test.cpp | 2 +- test/unittest/main.cpp | 2 +- tools/auto_tuner/CMakeLists.txt | 14 +- tools/auto_tuner/README.md | 2 +- tools/auto_tuner/include/gemm_tuner.hpp | 14 +- tools/auto_tuner/include/reference_gemm.hpp | 2 +- tools/auto_tuner/include/tune.hpp | 8 +- tools/auto_tuner/include/tune_impl.hpp | 12 +- tools/auto_tuner/include/tuner_types.hpp | 12 +- tools/auto_tuner/include/utils.hpp | 16 +- tools/auto_tuner/src/tune_all.cpp | 2 +- tools/auto_tuner/src/tune_nn.cpp | 2 +- tools/auto_tuner/src/tune_nt.cpp | 2 +- tools/auto_tuner/src/tune_tn.cpp | 2 +- tools/auto_tuner/src/tune_tt.cpp | 2 +- 394 files changed, 1423 insertions(+), 1423 deletions(-) rename benchmark/{syclblas => portblas}/CMakeLists.txt (86%) rename benchmark/{syclblas => portblas}/blas1/asum.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/axpy.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/copy.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/dot.cpp (99%) rename benchmark/{syclblas => portblas}/blas1/iamax.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/iamin.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/nrm2.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/rotg.cpp (99%) rename benchmark/{syclblas => portblas}/blas1/rotm.cpp (99%) rename benchmark/{syclblas => portblas}/blas1/rotmg.cpp (99%) rename benchmark/{syclblas => portblas}/blas1/scal.cpp (98%) rename benchmark/{syclblas => portblas}/blas1/sdsdot.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/gbmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/gemv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/ger.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/sbmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/spmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/spr.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/spr2.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/symv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/syr.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/syr2.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/tbmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/tbsv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/tpmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/tpsv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/trmv.cpp (99%) rename benchmark/{syclblas => portblas}/blas2/trsv.cpp (99%) rename benchmark/{syclblas => portblas}/blas3/gemm.cpp (99%) rename benchmark/{syclblas => portblas}/blas3/gemm_batched.cpp (99%) rename benchmark/{syclblas => portblas}/blas3/gemm_batched_strided.cpp (99%) rename benchmark/{syclblas => portblas}/blas3/symm.cpp (99%) rename benchmark/{syclblas => portblas}/blas3/trsm.cpp (99%) rename benchmark/{syclblas => portblas}/extension/omatadd.cpp (99%) rename benchmark/{syclblas => portblas}/extension/omatcopy.cpp (99%) rename benchmark/{syclblas => portblas}/extension/omatcopy2.cpp (99%) rename benchmark/{syclblas => portblas}/extension/reduction.cpp (99%) rename benchmark/{syclblas => portblas}/main.cpp (94%) rename benchmark/{syclblas => portblas}/utils.hpp (96%) rename cmake/Modules/{ConfigureSYCLBLAS.cmake => ConfigurePORTBLAS.cmake} (97%) rename include/{sycl_blas.h => portblas.h} (91%) rename include/{sycl_blas_helper.h => portblas_helper.h} (95%) rename include/sb_handle/{sycl_blas_handle.h => portblas_handle.h} (96%) create mode 100644 samples/FindPORTBLAS.cmake delete mode 100644 samples/FindSyclBLAS.cmake rename src/{sycl_blas.hpp => portblas.hpp} (91%) rename src/sb_handle/{sycl_blas_handle.hpp => portblas_handle.hpp} (98%) diff --git a/.github/workflows/build-and-test.yml b/.github/workflows/build-and-test.yml index 3ee012efb..849c02d84 100644 --- a/.github/workflows/build-and-test.yml +++ b/.github/workflows/build-and-test.yml @@ -76,9 +76,9 @@ jobs: --build-arg git_slug=${GIT_SLUG} --build-arg impl=${IMPL} --build-arg command=${COMMAND} - -t sycl-blas . + -t portblas . - name: Run the Docker image - run: docker run -t sycl-blas + run: docker run -t portblas # Builds the dockerfile in the repo and uses that to build auto-tuner. Build-Auto-Tuner: needs: skip-check @@ -116,6 +116,6 @@ jobs: --build-arg git_slug=${GIT_SLUG} --build-arg impl=${IMPL} --build-arg command=${COMMAND} - -t sycl-blas . + -t portblas . - name: Run the Docker image - run: docker run -t sycl-blas + run: docker run -t portblas diff --git a/CMakeLists.txt b/CMakeLists.txt index 589bfc08b..a6b85f570 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -17,13 +17,13 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * # **************************************************************************/ cmake_minimum_required(VERSION 3.4.3) -project(sycl-blas VERSION 0.1.0 LANGUAGES CXX) +project(portBLAS VERSION 0.1.0 LANGUAGES CXX) if(POLICY CMP0074) cmake_policy(SET CMP0074 NEW) @@ -36,7 +36,7 @@ if(EXISTS ${CMAKE_BINARY_DIR}/conanbuildinfo.cmake) list(APPEND CMAKE_PREFIX_PATH ${CMAKE_BINARY_DIR}) endif() -option(INSTALL_HEADER_ONLY "Install SYCL-BLAS as a header only library" OFF) +option(INSTALL_HEADER_ONLY "Install portBLAS as a header only library" OFF) set(BUILD_SHARED_LIBS ON CACHE BOOL "") set(CMAKE_CXX_STANDARD 14) @@ -58,19 +58,19 @@ list(APPEND CMAKE_MODULE_PATH ) # This turns the long file names that are generated into short versions using a hash -# This is required to build SYCL-BLAS on Windows and on WSL since Windows doesn't +# This is required to build portBLAS on Windows and on WSL since Windows doesn't # have good support for long file names -option(SYCLBLAS_USE_SHORT_NAMES "Whether to use short names for generated files" OFF) +option(PORTBLAS_USE_SHORT_NAMES "Whether to use short names for generated files" OFF) if (WIN32) # On WIN32 we can force this option to be true, but when building on WSL this # must be set in the command line - set(SYCLBLAS_USE_SHORT_NAMES ON CACHE BOOL "Enable short names for generated files in Windows" FORCE) + set(PORTBLAS_USE_SHORT_NAMES ON CACHE BOOL "Enable short names for generated files in Windows" FORCE) endif() # Setup datatypes, workgroup sizes and other options. # NB: This has to be included before CmakeFunctionHelper as it declares various options. if (NOT INSTALL_HEADER_ONLY) - include(ConfigureSYCLBLAS) + include(ConfigurePORTBLAS) include(SYCL) find_package(PythonInterp 3 REQUIRED) endif() @@ -78,7 +78,7 @@ endif() if (MSVC) # The device compiler needs C++14 to parse the Windows headers set(BUILD_SHARED_LIBS FALSE CACHE BOOL - "Force SYCL-BLAS to be built as a static library on Windows" + "Force portBLAS to be built as a static library on Windows" FORCE ) endif() @@ -91,12 +91,12 @@ option(GEMM_VECTORIZATION_SUPPORT "Whether to enable vectorization in Gemm kerne add_definitions(-DCL_TARGET_OPENCL_VERSION=220) set(CBLAS_INCLUDE ${CMAKE_CURRENT_SOURCE_DIR}/external/cblas/include) -set(SYCLBLAS_GENERATED_SRC ${CMAKE_CURRENT_BINARY_DIR}/generated_src) -set(SYCLBLAS_INCLUDE $ +set(PORTBLAS_GENERATED_SRC ${CMAKE_CURRENT_BINARY_DIR}/generated_src) +set(PORTBLAS_INCLUDE $ $) -set(SYCLBLAS_COMMON_INCLUDE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/common/include) -set(SYCLBLAS_SRC ${CMAKE_CURRENT_SOURCE_DIR}/src) -set(SYCLBLAS_SRC_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/python_generator) +set(PORTBLAS_COMMON_INCLUDE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/common/include) +set(PORTBLAS_SRC ${CMAKE_CURRENT_SOURCE_DIR}/src) +set(PORTBLAS_SRC_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/python_generator) list(APPEND THIRD_PARTIES_INCLUDE ${CBLAS_INCLUDE}) if(IMGDNN_DIR) @@ -105,7 +105,7 @@ if(IMGDNN_DIR) list(APPEND THIRD_PARTIES_INCLUDE ${IMGDNN_INCLUDE_DIRS}) endif() -option(BLAS_ENABLE_EXTENSIONS "Whether to enable sycl-blas extensions" ON) +option(BLAS_ENABLE_EXTENSIONS "Whether to enable portBLAS extensions" ON) # CmakeFunctionHelper has to be included after any options that it depends on are declared. # These include: @@ -118,91 +118,91 @@ option(BLAS_ENABLE_EXTENSIONS "Whether to enable sycl-blas extensions" ON) include(CmakeFunctionHelper) if (INSTALL_HEADER_ONLY) - add_library(sycl_blas INTERFACE) - set_target_properties(sycl_blas PROPERTIES - INTERFACE_INCLUDE_DIRECTORIES "${SYCLBLAS_INCLUDE};$" + add_library(portblas INTERFACE) + set_target_properties(portblas PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${PORTBLAS_INCLUDE};$" ) else() add_subdirectory(src) - build_library(sycl_blas ${BLAS_ENABLE_EXTENSIONS}) + build_library(portblas ${BLAS_ENABLE_EXTENSIONS}) if (WIN32) # On Windows, all symbols must be resolved at link time for DLLs. - # The sycl_blas target is just a collection of other objects so + # The portblas target is just a collection of other objects so # the linked libraries are not going to be propagated to this target. # This requires manual linking against SYCL on Windows. if(is_computecpp) - target_link_libraries(sycl_blas PUBLIC ComputeCpp::ComputeCpp) + target_link_libraries(portblas PUBLIC ComputeCpp::ComputeCpp) elseif(is_dpcpp) - target_link_libraries(sycl_blas PUBLIC DPCPP::DPCPP) + target_link_libraries(portblas PUBLIC DPCPP::DPCPP) elseif(is_hipsycl) - target_link_libraries(sycl_blas PUBLIC hipSYCL::hipSYCL-rt) + target_link_libraries(portblas PUBLIC hipSYCL::hipSYCL-rt) endif() endif() if(is_computecpp) set(sycl_impl ComputeCpp::ComputeCpp) elseif(is_dpcpp) set(sycl_impl DPCPP::DPCPP) - add_sycl_to_target(TARGET sycl_blas SOURCES) + add_sycl_to_target(TARGET portblas SOURCES) elseif(is_hipsycl) set(sycl_impl hipSYCL::hipSYCL-rt) - add_sycl_to_target(TARGET sycl_blas SOURCES) + add_sycl_to_target(TARGET portblas SOURCES) endif() if(IMGDNN_DIR) - target_link_libraries(sycl_blas PUBLIC IMGDNN::IMGDNN) + target_link_libraries(portblas PUBLIC IMGDNN::IMGDNN) endif() - set_target_properties(sycl_blas PROPERTIES + set_target_properties(portblas PROPERTIES INTERFACE_LINK_LIBRARIES ${sycl_impl} - INTERFACE_INCLUDE_DIRECTORIES "${SYCLBLAS_INCLUDE}" + INTERFACE_INCLUDE_DIRECTORIES "${PORTBLAS_INCLUDE}" ) - set_target_properties(sycl_blas PROPERTIES + set_target_properties(portblas PROPERTIES VERSION ${PROJECT_VERSION} ) endif() include(CMakePackageConfigHelpers) -set(version_file "${CMAKE_CURRENT_BINARY_DIR}/cmake/sycl_blas-version.cmake") +set(version_file "${CMAKE_CURRENT_BINARY_DIR}/cmake/portblas-version.cmake") write_basic_package_version_file(${version_file} VERSION ${PROJECT_VERSION} COMPATIBILITY AnyNewerVersion ) include(GNUInstallDirs) -install(TARGETS sycl_blas - EXPORT sycl_blas +install(TARGETS portblas + EXPORT portblas RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR} LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR} ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR} ) -install(DIRECTORY ${SYCLBLAS_INCLUDE} +install(DIRECTORY ${PORTBLAS_INCLUDE} DESTINATION ${CMAKE_INSTALL_PREFIX} - COMPONENT sycl_blas + COMPONENT portblas FILES_MATCHING PATTERN "*.h" ) if (INSTALL_HEADER_ONLY) - install(DIRECTORY ${SYCLBLAS_SRC} + install(DIRECTORY ${PORTBLAS_SRC} DESTINATION ${CMAKE_INSTALL_PREFIX} - COMPONENT sycl_blas + COMPONENT portblas FILES_MATCHING PATTERN "*.hpp" ) endif() install(FILES ${version_file} DESTINATION ${CMAKE_INSTALL_PREFIX}) -install(EXPORT sycl_blas +install(EXPORT portblas DESTINATION ${CMAKE_INSTALL_PREFIX} - NAMESPACE SYCL_BLAS:: - FILE sycl_blas-config.cmake + NAMESPACE PORTBLAS:: + FILE portblas-config.cmake ) -export(EXPORT sycl_blas - NAMESPACE SYCL_BLAS:: - FILE sycl_blas-config.cmake +export(EXPORT portblas + NAMESPACE PORTBLAS:: + FILE portblas-config.cmake ) option(BLAS_ENABLE_TESTING "Whether to enable testing" ON) option(ENABLE_EXPRESSION_TESTS "Whether to build expression tree fusion tests" OFF) if (INSTALL_HEADER_ONLY AND BLAS_ENABLE_TESTING) - message(STATUS "Tests are disabled when installing SYCL-BLAS in header only mode") + message(STATUS "Tests are disabled when installing portBLAS in header only mode") set(BLAS_ENABLE_TESTING OFF) endif() @@ -219,13 +219,13 @@ option(BUILD_CLBLAS_BENCHMARKS "Whether to build clBLAS benchmarks" OFF) option(BUILD_CUBLAS_BENCHMARKS "Whether to build cuBLAS benchmarks" OFF) option(BUILD_ROCBLAS_BENCHMARKS "Whether to build rocBLAS benchmarks" OFF) option(BUILD_ACL_BENCHMARKS "Whether to build ARM Compute Library benchmarks" OFF) -option(BLAS_BUILD_SAMPLES "Whether to build SYCL-BLAS samples" ON) +option(BLAS_BUILD_SAMPLES "Whether to build portBLAS samples" ON) if (INSTALL_HEADER_ONLY AND BLAS_ENABLE_BENCHMARK) - message(STATUS "Benchmarks are disabled when installing SYCL-BLAS in header only mode") + message(STATUS "Benchmarks are disabled when installing portBLAS in header only mode") set(BLAS_ENABLE_BENCHMARK OFF) endif() if (INSTALL_HEADER_ONLY AND BLAS_BUILD_SAMPLES) - message(STATUS "Samples are disabled when installing SYCL-BLAS in header only mode") + message(STATUS "Samples are disabled when installing portBLAS in header only mode") set(BLAS_BUILD_SAMPLES OFF) endif() diff --git a/Dockerfile b/Dockerfile index f1a12f30d..549ba3777 100644 --- a/Dockerfile +++ b/Dockerfile @@ -39,23 +39,23 @@ RUN if [ "${c_compiler}" = 'gcc-7' ]; then apt-get install -yq \ RUN apt-get install -yq --allow-downgrades --allow-remove-essential \ --allow-change-held-packages ocl-icd-opencl-dev ocl-icd-dev opencl-headers -RUN git clone https://github.com/${git_slug}.git --recursive -b ${git_branch} /sycl-blas +RUN git clone https://github.com/${git_slug}.git --recursive -b ${git_branch} /portBLAS #OpenBLAS -RUN bash /sycl-blas/.scripts/build_OpenBLAS.sh +RUN bash /portBLAS/.scripts/build_OpenBLAS.sh # Intel OpenCL Runtime -RUN bash /sycl-blas/.scripts/install_intel_opencl.sh +RUN bash /portBLAS/.scripts/install_intel_opencl.sh # SYCL -RUN if [ "${impl}" = 'COMPUTECPP' ]; then cd /sycl-blas && bash /sycl-blas/.scripts/build_computecpp.sh; fi -RUN if [ "${impl}" = 'DPCPP' ]; then cd /sycl-blas && bash /sycl-blas/.scripts/build_dpcpp.sh; fi +RUN if [ "${impl}" = 'COMPUTECPP' ]; then cd /portBLAS && bash /portBLAS/.scripts/build_computecpp.sh; fi +RUN if [ "${impl}" = 'DPCPP' ]; then cd /portBLAS && bash /portBLAS/.scripts/build_dpcpp.sh; fi ENV COMMAND=${command} ENV CC=${c_compiler} ENV CXX=${cxx_compiler} ENV SYCL_IMPL=${impl} -CMD cd /sycl-blas && \ +CMD cd /portBLAS && \ if [ "${COMMAND}" = 'build-test' ]; then \ if [ "${SYCL_IMPL}" = 'COMPUTECPP' ]; then \ /tmp/ComputeCpp-latest/bin/computecpp_info && \ diff --git a/NOTICE b/NOTICE index d06492de9..ac3037791 100644 --- a/NOTICE +++ b/NOTICE @@ -1,2 +1,2 @@ -SYCL-BLAS +portBLAS Copyright (C) Codeplay Software Limited. diff --git a/README.md b/README.md index 3a40d3e61..2d7508b68 100644 --- a/README.md +++ b/README.md @@ -1,24 +1,24 @@ -SYCL-BLAS Implementation +portBLAS Implementation === -[![Build and Test](https://github.com/codeplaysoftware/sycl-blas/actions/workflows/build-and-test.yml/badge.svg?event=push)](https://github.com/codeplaysoftware/sycl-blas/actions/workflows/build-and-test.yml) +[![Build and Test](https://github.com/codeplaysoftware/portBLAS/actions/workflows/build-and-test.yml/badge.svg?event=push)](https://github.com/codeplaysoftware/portBLAS/actions/workflows/build-and-test.yml) -SYCL-BLAS implements BLAS - [Basic Linear Algebra Subroutines](https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprograms) - using [SYCL 1.2]( +portBLAS implements BLAS - [Basic Linear Algebra Subroutines](https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprograms) - using [SYCL 1.2]( https://www.khronos.org/registry/sycl/specs/sycl-1.2.pdf), the [Khronos](http://www.khronos.org) abstraction layer for [OpenCL](https://www.khronos.org/opencl/). -SYCL-BLAS is a current work in progress research project from an ongoing +portBLAS is a current work in progress research project from an ongoing collaboration with the *High Performance Computing & Architectures (HPCA) group* from the Universitat Jaume I [UJI](http://www.hpca.uji.es/). -SYCL-BLAS is written using modern C++. The current implementation uses C++11 +portBLAS is written using modern C++. The current implementation uses C++11 features. See [Roadmap](Roadmap.md) for details on the current status and plans for the project. ## Table of Contents -- [SYCL-BLAS Implementation](#sycl-blas-implementation) +- [portBLAS Implementation](#portBLAS-implementation) - [Table of Contents](#table-of-contents) - [Motivation](#motivation) - [Basic Concepts](#basic-concepts) @@ -36,7 +36,7 @@ the project. - [Compile with ComputeCpp](#compile-with-computecpp) - [Compile with DPC++](#compile-with-dpc) - [Compile with hipSYCL](#compile-with-hipsycl) - - [Instaling SYCL-BLAS](#instaling-sycl-blas) + - [Instaling portBLAS](#instaling-portBLAS) - [POWER\_VR support (ComputeCpp Only)](#power_vr-support-computecpp-only) - [Doxygen](#doxygen) - [CMake options](#cmake-options) @@ -74,7 +74,7 @@ of numerical libraries, such that a good BLAS implementation improves the performances of all the other libraries. The development of numerical libraries on SYCL is one of the most important objectives, because it will improve the performance of other SYCL -applications. Obviously, it makes sense SYCL-BLAS was the first step in this +applications. Obviously, it makes sense portBLAS was the first step in this task. On GPUs, the data communication to/from the device and the grain of the kernels @@ -92,11 +92,11 @@ expression tree on which all the single operation which are required to solve a problem appears. This structure could be analysed by the compiler to decide how to merge the different kernel and the best grid topology to execute the fused kernel. The use of expression trees is one of most important features of -SYCL-BLAS. +portBLAS. ## Basic Concepts -SYCL-BLAS uses C++ Expression Tree templates to generate SYCL Kernels via +portBLAS uses C++ Expression Tree templates to generate SYCL Kernels via kernel composition. Expression Tree templates are a widely used technique to implement expressions on C++, that facilitate development and composition of operations. @@ -105,20 +105,20 @@ In particular, been used in various projects to create efficient domain-specific embedded languages that enable users to easily fuse GPU kernels. -SYCL-BLAS can be used -- either as a header-only framework by including `sycl_blas.hpp` in +portBLAS can be used +- either as a header-only framework by including `portblas.hpp` in an application and passing the `src` folder in the list of include directories -- or as a library by including `sycl_blas.h` in an application. +- or as a library by including `portblas.h` in an application. All the relevant files can be found in the `include` directory. -There are four components in SYCL-BLAS, the *View*, the *Operations*, +There are four components in portBLAS, the *View*, the *Operations*, the *SB_Handle* and the *Interface* itself. ### Views -The input data to all the operations in SYCL-BLAS is passed to the library +The input data to all the operations in portBLAS is passed to the library using *Views*. A *View* represents data on top of a container, passed by reference. Views *do not store data*, they only map a visualization of the data on top @@ -139,7 +139,7 @@ properties of the C++11 standard. Operations among elements of vectors (or matrices) are expressed in the set of Operation Classes. Operations are templated classes that take templated types as input. -Operations form the nodes of the SYCL-BLAS expression tree. +Operations form the nodes of the portBLAS expression tree. Refer to the documentation of each node type for details. Composing these is how the compile-time Expression tree is created: @@ -162,7 +162,7 @@ The different headers on the interface directory implement the traditional BLAS interface. Files are organised per BLAS level (1, 2, 3). -When the SYCL-BLAS BLAS interface is called, the Expression Tree for each +When the portBLAS BLAS interface is called, the Expression Tree for each operation is constructed, and then executed. Some API calls may execute several kernels (e.g, when a reduction is required). The expression trees in the API allow to compile-time fuse operations. @@ -184,7 +184,7 @@ a tuple). The containers for the vectors and matrices (and scalars written by the BLAS operations) are iterator buffers that can be created with `make_sycl_iterator_buffer`. -We recommend checking the [samples](samples) to get started with SYCL-BLAS. It +We recommend checking the [samples](samples) to get started with portBLAS. It is better to be familiar with BLAS: - [Wikipedia](https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprograms) @@ -326,7 +326,7 @@ Other non-official extension operators : | `_transpose*` | `sb_handle`, `M`, `N`, `A`, `lda`, `ldb` | Computes an in-place matrix transpose operation using a general dense matrix, lda & ldb being input and output leading dimensions of A respectively _(*Not implemented)_. | ### Experimental Joint Matrix Support -SYCL-BLAS now supports sub-group based collective GEMM operation using the experimental +portBLAS now supports sub-group based collective GEMM operation using the experimental [`joint_matrix`](https://github.com/intel/llvm/blob/sycl/sycl/doc/extensions/experimental/sycl_ext_oneapi_matrix/sycl_ext_oneapi_matrix.asciidoc) extension provided by DPC++. This support is only accessible for the latest NVIDIA Ampere GPUs and beyond. The requirements for using this experimental support are: @@ -342,7 +342,7 @@ The user should expect erroneous behaviour from the code if both of these requir ## Requirements -SYCL-BLAS is designed to work with any SYCL 1.2.1 implementation. +portBLAS is designed to work with any SYCL 1.2.1 implementation. We do not use any OpenCL interoperability, hence, the code is pure C++. The project is developed using [ComputeCpp CE Edition](http://www.computecpp.com) using Ubuntu 16.04 on Intel OpenCL CPU and Intel GPU. @@ -359,7 +359,7 @@ which can cause random test failures. Any version of OpenBLAS `>= 0.3.0` will no from these issues. When using OpenBLAS or any other BLAS library the installation directory must be -added to the `CMAKE_PREFIX_PATH` when building SYCL-BLAS (see +added to the `CMAKE_PREFIX_PATH` when building portBLAS (see [below](###cmake-options)). ## Setup @@ -370,7 +370,7 @@ been replaced by `TUNING_TARGET`, which accepts the same options. triplet for DPC++ or the hipSYCL target. Please refer to the sections below for setting them. -1. Clone the SYCL-BLAS repository, making sure to pass the `--recursive` option, in order to clone submodule(s). +1. Clone the portBLAS repository, making sure to pass the `--recursive` option, in order to clone submodule(s). 2. Create a build directory 3. Run `CMake` from the build directory (see options in the section below): @@ -407,8 +407,8 @@ ninja ``` To build for other than the default devices (`omp`), set the `HIPSYCL_TARGETS` environment variable or specify `-DHIPSYCL_TARGETS` as [documented](https://github.com/illuhad/hipSYCL/blob/develop/doc/using-hipsycl.md). -### Instaling SYCL-BLAS -To install the SYCL-BLAS library (see `CMAKE_INSTALL_PREFIX` below) +### Instaling portBLAS +To install the portBLAS library (see `CMAKE_INSTALL_PREFIX` below) ```bash ninja install @@ -447,14 +447,14 @@ Some of the supported options are: | `ENABLE_EXPRESSION_TESTS` | `ON`/`OFF` | Build additional tests that use the header-only framework (e.g to test expression trees); `OFF` by default | | `BLAS_VERIFY_BENCHMARK` | `ON`/`OFF` | Verify the results of the benchmarks instead of only measuring the performance. See the documentation of the benchmarks for more details. `ON` by default | | `BLAS_ENABLE_CONST_INPUT` | `ON`/`OFF` | Determines whether to enable kernel instantiation with const input buffer (`ON` by default) | -| `BLAS_ENABLE_EXTENSIONS` | `ON`/`OFF` | Determines whether to enable sycl-blas extensions (`ON` by default) | +| `BLAS_ENABLE_EXTENSIONS` | `ON`/`OFF` | Determines whether to enable portBLAS extensions (`ON` by default) | | `BLAS_DATA_TYPES` | `half;float;double` | Determines the floating-point types to instantiate BLAS operations for. Default is `float` | | `BLAS_INDEX_TYPES` | `int32_t;int64_t` | Determines the type(s) to use for `index_t` and `increment_t`. Default is `int` | ### Cross-Compile (ComputeCpp Only) -To cross-compile SYCL-BLAS first the following environment variables must be +To cross-compile portBLAS first the following environment variables must be set: ```bash @@ -464,7 +464,7 @@ export COMPUTECPP_SYSROOT_DIR="$PATH TO SYSROOT_DIR" ``` Clone the [ComputeCpp-SDK](https://github.com/codeplaysoftware/computecpp-sdk) to retrieve the toolchain file. -The following CMake command can be used to cross-compile SYCL-BLAS: +The following CMake command can be used to cross-compile portBLAS: ```bash cmake -GNinja \ @@ -474,7 +474,7 @@ cmake -GNinja -DComputeCpp_HOST_DIR="${COMPUTECPP_X86_PATH}" \ -DCMAKE_TOOLCHAIN_FILE="/path/to/computecpp-sdk/cmake/toolchains/gcc-generic.cmake" \ -DCMAKE_BUILD_TYPE='Release' \ - -DCMAKE_INSTALL_PREFIX=${CROSS_COMPILED_SYCLBLAS_INSTALL} \ + -DCMAKE_INSTALL_PREFIX=${CROSS_COMPILED_PORTBLAS_INSTALL} \ -DOpenCL_INCLUDE_DIR="${OpenCL_Headers_PATH}" \ -DOpenCL_LIBRARY="${OpenCL_LIBRARY}" \ -DCOMPUTECPP_BITCODE="${DEVICE_BITCODE}" \ @@ -493,7 +493,7 @@ The tests and benchmarks have their own documentation: ## Contributing to the project -SYCL-BLAS is an Open Source project maintained by the HPCA group and +portBLAS is an Open Source project maintained by the HPCA group and Codeplay Software Ltd. Feel free to create an issue on the Github tracker to request features or report bugs. diff --git a/benchmark/CMakeLists.txt b/benchmark/CMakeLists.txt index a5663bb10..1d4f1fa83 100644 --- a/benchmark/CMakeLists.txt +++ b/benchmark/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -59,7 +59,7 @@ add_subdirectory(${CMAKE_CURRENT_BINARY_DIR}/googlebench-src ${CMAKE_CURRENT_BINARY_DIR}/googlebench-build EXCLUDE_FROM_ALL) -# Set SYCL-BLAS options +# Set portBLAS options if(CMAKE_CROSSCOMPILING) set(cmake_toolchain "-DCMAKE_TOOLCHAIN_FILE=${CMAKE_TOOLCHAIN_FILE}" @@ -129,7 +129,7 @@ if(SYCL_DEVICE) endif(SYCL_DEVICE) include_directories(${BLAS_BENCH}) -add_subdirectory(${BLAS_BENCH}/syclblas) +add_subdirectory(${BLAS_BENCH}/portblas) if (BUILD_CLBLAST_BENCHMARKS) add_subdirectory(${BLAS_BENCH}/clBench/clblast) diff --git a/benchmark/README.md b/benchmark/README.md index 213a5e93a..0089ba9d2 100644 --- a/benchmark/README.md +++ b/benchmark/README.md @@ -3,7 +3,7 @@ Benchmarks ## General information -The SYCL-BLAS benchmarks are intended to measure the evolution of the +The portBLAS benchmarks are intended to measure the evolution of the performance of this BLAS implementation and how it compares with other tuned implementations, such as [CLBLAST](https://github.com/CNugteren/CLBlast) (a very performant OpenCL BLAS library). @@ -24,10 +24,10 @@ To install CLBlast, see: https://github.com/CNugteren/CLBlast/blob/master/doc/installation.md)) After the compilation, the binaries will be available: -* in the build folder, in `benchmark/syclblas/` and `benchmark/clblast/` +* in the build folder, in `benchmark/portblas/` and `benchmark/clblast/` * if you provide an installation directory with the CMake variable `CMAKE_INSTALL_PREFIX`, and run the installation command, e.g - `ninja install`, in your installation folder, in `sycl_blas/bin/` + `ninja install`, in your installation folder, in `portblas/bin/` A verification of the results is enabled by default and can be disabled with the CMake option `BLAS_VERIFY_BENCHMARK` set to `OFF` or `0`. The verification will @@ -40,7 +40,7 @@ CMake parameter. ## How to run the benchmarks The benchmarks take two kinds of command-line options: those for the benchmark -library and those specific to the SYCL-BLAS projects. +library and those specific to the portBLAS projects. Essentially, the benchmarks can take a CSV configuration file (or will use defaults), and if your machine has more than one OpenCL device, you can specify diff --git a/benchmark/bench_info.cc b/benchmark/bench_info.cc index 6d48778aa..2bc6a4009 100644 --- a/benchmark/bench_info.cc +++ b/benchmark/bench_info.cc @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename bench_info.cc * diff --git a/benchmark/clBench/acl/CMakeLists.txt b/benchmark/clBench/acl/CMakeLists.txt index f5a381392..3d6dbd9f2 100644 --- a/benchmark/clBench/acl/CMakeLists.txt +++ b/benchmark/clBench/acl/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -30,9 +30,9 @@ endif() function(generate_acl_benchmark name source definition) add_executable(${name} ${source} main.cpp) - target_link_libraries(${name} PRIVATE benchmark acl sycl_blas Clara::Clara) + target_link_libraries(${name} PRIVATE benchmark acl portblas Clara::Clara) target_compile_definitions(${name} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) - target_include_directories(${name} PRIVATE ${CBLAS_INCLUDE} ${BLAS_BENCH} ${SYCLBLAS_INCLUDE} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_include_directories(${name} PRIVATE ${CBLAS_INCLUDE} ${BLAS_BENCH} ${PORTBLAS_INCLUDE} ${PORTBLAS_COMMON_INCLUDE_DIR}) target_compile_definitions(${name} PRIVATE ${definition}) if(BLAS_VERIFY_BENCHMARK) diff --git a/benchmark/clBench/acl/blas3/gemm.cpp b/benchmark/clBench/acl/blas3/gemm.cpp index 25a9dce46..295bf8d88 100644 --- a/benchmark/clBench/acl/blas3/gemm.cpp +++ b/benchmark/clBench/acl/blas3/gemm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp * diff --git a/benchmark/clBench/acl/main.cpp b/benchmark/clBench/acl/main.cpp index 9d066e3c4..3b080d3f5 100644 --- a/benchmark/clBench/acl/main.cpp +++ b/benchmark/clBench/acl/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/benchmark/clBench/acl/utils.hpp b/benchmark/clBench/acl/utils.hpp index 1631cf1d5..10c5b31e2 100644 --- a/benchmark/clBench/acl/utils.hpp +++ b/benchmark/clBench/acl/utils.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * diff --git a/benchmark/clBench/clblas/CMakeLists.txt b/benchmark/clBench/clblas/CMakeLists.txt index 224da0ff0..1b7a391f3 100644 --- a/benchmark/clBench/clblas/CMakeLists.txt +++ b/benchmark/clBench/clblas/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -40,9 +40,9 @@ set(sources foreach(clblas_benchmark ${sources}) get_filename_component(clblas_bench_exec ${clblas_benchmark} NAME_WE) add_executable(bench_clblas_${clblas_bench_exec} ${clblas_benchmark} "main.cpp") - target_link_libraries(bench_clblas_${clblas_bench_exec} PRIVATE benchmark Clara::Clara sycl_blas) + target_link_libraries(bench_clblas_${clblas_bench_exec} PRIVATE benchmark Clara::Clara portblas) target_compile_definitions(bench_clblas_${clblas_bench_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) - target_include_directories(bench_clblas_${clblas_bench_exec} PRIVATE ${CLBLAS_INCLUDE_DIRS} ${BLAS_BENCH} ${SYCLBLAS_INCLUDE} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_include_directories(bench_clblas_${clblas_bench_exec} PRIVATE ${CLBLAS_INCLUDE_DIRS} ${BLAS_BENCH} ${PORTBLAS_INCLUDE} ${PORTBLAS_COMMON_INCLUDE_DIR}) target_link_libraries(bench_clblas_${clblas_bench_exec} PRIVATE ${CLBLAS_LIBRARIES}) if(BLAS_VERIFY_BENCHMARK) diff --git a/benchmark/clBench/clblas/blas2/tbsv.cpp b/benchmark/clBench/clblas/blas2/tbsv.cpp index c6fc9077d..0c3691315 100644 --- a/benchmark/clBench/clblas/blas2/tbsv.cpp +++ b/benchmark/clBench/clblas/blas2/tbsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbsv.cpp * diff --git a/benchmark/clBench/clblas/blas2/trsv.cpp b/benchmark/clBench/clblas/blas2/trsv.cpp index 934d5a4b7..492205d9b 100644 --- a/benchmark/clBench/clblas/blas2/trsv.cpp +++ b/benchmark/clBench/clblas/blas2/trsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp * diff --git a/benchmark/clBench/clblas/blas3/trsm.cpp b/benchmark/clBench/clblas/blas3/trsm.cpp index 19cdce26b..bd604cae5 100644 --- a/benchmark/clBench/clblas/blas3/trsm.cpp +++ b/benchmark/clBench/clblas/blas3/trsm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm.cpp * diff --git a/benchmark/clBench/clblas/main.cpp b/benchmark/clBench/clblas/main.cpp index 5b0f12a7b..d16d846b7 100644 --- a/benchmark/clBench/clblas/main.cpp +++ b/benchmark/clBench/clblas/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/benchmark/clBench/clblas/utils.hpp b/benchmark/clBench/clblas/utils.hpp index 73ce2cb89..e8e6c3f5b 100644 --- a/benchmark/clBench/clblas/utils.hpp +++ b/benchmark/clBench/clblas/utils.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * diff --git a/benchmark/clBench/clblast/CMakeLists.txt b/benchmark/clBench/clblast/CMakeLists.txt index 65dbdc7a8..cf791cef0 100644 --- a/benchmark/clBench/clblast/CMakeLists.txt +++ b/benchmark/clBench/clblast/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -63,7 +63,7 @@ foreach(clblast_benchmark ${sources}) add_executable(bench_clblast_${clblast_bench_exec} ${clblast_benchmark} main.cpp) target_link_libraries(bench_clblast_${clblast_bench_exec} PRIVATE benchmark Clara::Clara clblast) target_compile_definitions(bench_clblast_${clblast_bench_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) - target_include_directories(bench_clblast_${clblast_bench_exec} PRIVATE ${SYCLBLAS_INCLUDE} ${CBLAS_INCLUDE} ${BLAS_BENCH} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_include_directories(bench_clblast_${clblast_bench_exec} PRIVATE ${PORTBLAS_INCLUDE} ${CBLAS_INCLUDE} ${BLAS_BENCH} ${PORTBLAS_COMMON_INCLUDE_DIR}) # Even though CLBlast does not use sycl, the common tools indirectly include sycl headers. add_sycl_to_target( diff --git a/benchmark/clBench/clblast/blas1/asum.cpp b/benchmark/clBench/clblast/blas1/asum.cpp index 70bfb1de9..34985326e 100644 --- a/benchmark/clBench/clblast/blas1/asum.cpp +++ b/benchmark/clBench/clblast/blas1/asum.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename asum.cpp * diff --git a/benchmark/clBench/clblast/blas1/axpy.cpp b/benchmark/clBench/clblast/blas1/axpy.cpp index 660137cca..1bb99ccd2 100644 --- a/benchmark/clBench/clblast/blas1/axpy.cpp +++ b/benchmark/clBench/clblast/blas1/axpy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename axpy.cpp * diff --git a/benchmark/clBench/clblast/blas1/dot.cpp b/benchmark/clBench/clblast/blas1/dot.cpp index a4d48e73d..2c65515b4 100644 --- a/benchmark/clBench/clblast/blas1/dot.cpp +++ b/benchmark/clBench/clblast/blas1/dot.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot.cpp * diff --git a/benchmark/clBench/clblast/blas1/iamax.cpp b/benchmark/clBench/clblast/blas1/iamax.cpp index 9d45f30e5..5b503f318 100644 --- a/benchmark/clBench/clblast/blas1/iamax.cpp +++ b/benchmark/clBench/clblast/blas1/iamax.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax.cpp * diff --git a/benchmark/clBench/clblast/blas1/iamin.cpp b/benchmark/clBench/clblast/blas1/iamin.cpp index ec537b231..9dd2ba653 100644 --- a/benchmark/clBench/clblast/blas1/iamin.cpp +++ b/benchmark/clBench/clblast/blas1/iamin.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin.cpp * diff --git a/benchmark/clBench/clblast/blas1/nrm2.cpp b/benchmark/clBench/clblast/blas1/nrm2.cpp index 17646eadf..1b260180e 100644 --- a/benchmark/clBench/clblast/blas1/nrm2.cpp +++ b/benchmark/clBench/clblast/blas1/nrm2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2.cpp * diff --git a/benchmark/clBench/clblast/blas1/scal.cpp b/benchmark/clBench/clblast/blas1/scal.cpp index 5e0f845cb..b681a941e 100644 --- a/benchmark/clBench/clblast/blas1/scal.cpp +++ b/benchmark/clBench/clblast/blas1/scal.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename scal.cpp * diff --git a/benchmark/clBench/clblast/blas2/gbmv.cpp b/benchmark/clBench/clblast/blas2/gbmv.cpp index dcc3a7530..2e05f6cae 100644 --- a/benchmark/clBench/clblast/blas2/gbmv.cpp +++ b/benchmark/clBench/clblast/blas2/gbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.cpp * diff --git a/benchmark/clBench/clblast/blas2/gemv.cpp b/benchmark/clBench/clblast/blas2/gemv.cpp index 590a71030..5213ff60a 100644 --- a/benchmark/clBench/clblast/blas2/gemv.cpp +++ b/benchmark/clBench/clblast/blas2/gemv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.cpp * diff --git a/benchmark/clBench/clblast/blas2/ger.cpp b/benchmark/clBench/clblast/blas2/ger.cpp index c8c71c8ea..ff12da289 100644 --- a/benchmark/clBench/clblast/blas2/ger.cpp +++ b/benchmark/clBench/clblast/blas2/ger.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.cpp * diff --git a/benchmark/clBench/clblast/blas2/sbmv.cpp b/benchmark/clBench/clblast/blas2/sbmv.cpp index ae7edca76..cf6c38fdc 100644 --- a/benchmark/clBench/clblast/blas2/sbmv.cpp +++ b/benchmark/clBench/clblast/blas2/sbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.cpp * diff --git a/benchmark/clBench/clblast/blas2/spr.cpp b/benchmark/clBench/clblast/blas2/spr.cpp index 00d589021..4003f2b70 100644 --- a/benchmark/clBench/clblast/blas2/spr.cpp +++ b/benchmark/clBench/clblast/blas2/spr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.cpp * diff --git a/benchmark/clBench/clblast/blas2/spr2.cpp b/benchmark/clBench/clblast/blas2/spr2.cpp index 313e7f8cf..9849f43b4 100644 --- a/benchmark/clBench/clblast/blas2/spr2.cpp +++ b/benchmark/clBench/clblast/blas2/spr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr2.cpp * diff --git a/benchmark/clBench/clblast/blas2/symv.cpp b/benchmark/clBench/clblast/blas2/symv.cpp index 703578998..96c676503 100644 --- a/benchmark/clBench/clblast/blas2/symv.cpp +++ b/benchmark/clBench/clblast/blas2/symv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symv.cpp * diff --git a/benchmark/clBench/clblast/blas2/syr.cpp b/benchmark/clBench/clblast/blas2/syr.cpp index a6bdec15c..5f8b8c4ef 100644 --- a/benchmark/clBench/clblast/blas2/syr.cpp +++ b/benchmark/clBench/clblast/blas2/syr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr.cpp * diff --git a/benchmark/clBench/clblast/blas2/syr2.cpp b/benchmark/clBench/clblast/blas2/syr2.cpp index 6e2d7ec2d..07f1b7b45 100644 --- a/benchmark/clBench/clblast/blas2/syr2.cpp +++ b/benchmark/clBench/clblast/blas2/syr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2.cpp * diff --git a/benchmark/clBench/clblast/blas2/tbmv.cpp b/benchmark/clBench/clblast/blas2/tbmv.cpp index 45c1da9c3..e9049a958 100644 --- a/benchmark/clBench/clblast/blas2/tbmv.cpp +++ b/benchmark/clBench/clblast/blas2/tbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp * diff --git a/benchmark/clBench/clblast/blas2/trmv.cpp b/benchmark/clBench/clblast/blas2/trmv.cpp index ebcf9a2f0..b9f0b732a 100644 --- a/benchmark/clBench/clblast/blas2/trmv.cpp +++ b/benchmark/clBench/clblast/blas2/trmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmv.cpp * diff --git a/benchmark/clBench/clblast/blas2/trsv.cpp b/benchmark/clBench/clblast/blas2/trsv.cpp index d4a619d58..9ed2e6b3c 100644 --- a/benchmark/clBench/clblast/blas2/trsv.cpp +++ b/benchmark/clBench/clblast/blas2/trsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp * diff --git a/benchmark/clBench/clblast/blas3/gemm.cpp b/benchmark/clBench/clblast/blas3/gemm.cpp index 045f95757..52ac803e5 100644 --- a/benchmark/clBench/clblast/blas3/gemm.cpp +++ b/benchmark/clBench/clblast/blas3/gemm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp * diff --git a/benchmark/clBench/clblast/blas3/gemm_batched.cpp b/benchmark/clBench/clblast/blas3/gemm_batched.cpp index 26dc68296..f82674d4d 100644 --- a/benchmark/clBench/clblast/blas3/gemm_batched.cpp +++ b/benchmark/clBench/clblast/blas3/gemm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched.cpp * diff --git a/benchmark/clBench/clblast/blas3/symm.cpp b/benchmark/clBench/clblast/blas3/symm.cpp index 8c41fbcc9..a951c232e 100644 --- a/benchmark/clBench/clblast/blas3/symm.cpp +++ b/benchmark/clBench/clblast/blas3/symm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symm.cpp * diff --git a/benchmark/clBench/clblast/blas3/trsm.cpp b/benchmark/clBench/clblast/blas3/trsm.cpp index cf9f2746f..0a81d9ef6 100644 --- a/benchmark/clBench/clblast/blas3/trsm.cpp +++ b/benchmark/clBench/clblast/blas3/trsm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm.cpp * diff --git a/benchmark/clBench/clblast/main.cpp b/benchmark/clBench/clblast/main.cpp index 5b0f12a7b..d16d846b7 100644 --- a/benchmark/clBench/clblast/main.cpp +++ b/benchmark/clBench/clblast/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/benchmark/clBench/clblast/utils.hpp b/benchmark/clBench/clblast/utils.hpp index a60f50ab8..4f84a67fc 100644 --- a/benchmark/clBench/clblast/utils.hpp +++ b/benchmark/clBench/clblast/utils.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * diff --git a/benchmark/clBench/clwrap.hpp b/benchmark/clBench/clwrap.hpp index a9ca13b8b..d02b0a2a1 100644 --- a/benchmark/clBench/clwrap.hpp +++ b/benchmark/clBench/clwrap.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename clwrap.h * diff --git a/benchmark/computecpp_version_config.h.in b/benchmark/computecpp_version_config.h.in index 0a3cf4694..d05439c5f 100644 --- a/benchmark/computecpp_version_config.h.in +++ b/benchmark/computecpp_version_config.h.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename computecpp_version_config.h.in * diff --git a/benchmark/cublas/CMakeLists.txt b/benchmark/cublas/CMakeLists.txt index 872eff646..250278fac 100644 --- a/benchmark/cublas/CMakeLists.txt +++ b/benchmark/cublas/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -78,13 +78,13 @@ set(sources foreach(cublas_bench ${sources}) get_filename_component(bench_cublas_exec ${cublas_bench} NAME_WE) add_executable(bench_cublas_${bench_cublas_exec} ${cublas_bench} main.cpp) - target_link_libraries(bench_cublas_${bench_cublas_exec} PRIVATE benchmark CUDA::toolkit CUDA::cublas CUDA::cudart sycl_blas Clara::Clara bench_info) + target_link_libraries(bench_cublas_${bench_cublas_exec} PRIVATE benchmark CUDA::toolkit CUDA::cublas CUDA::cudart portblas Clara::Clara bench_info) target_compile_definitions(bench_cublas_${bench_cublas_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) add_sycl_to_target( TARGET bench_cublas_${bench_cublas_exec} SOURCES ${cublas_bench} ) -target_include_directories(bench_cublas_${bench_cublas_exec} PRIVATE $ENV{CUBLAS_INCLUDE} ${SYCLBLAS_INCLUDE} ${SYCLBLAS_COMMON_INCLUDE_DIR} ${CBLAS_INCLUDE}) +target_include_directories(bench_cublas_${bench_cublas_exec} PRIVATE $ENV{CUBLAS_INCLUDE} ${PORTBLAS_INCLUDE} ${PORTBLAS_COMMON_INCLUDE_DIR} ${CBLAS_INCLUDE}) add_compile_definitions(BUILD_CUBLAS_BENCHMARKS) if(BLAS_VERIFY_BENCHMARK) diff --git a/benchmark/cublas/blas1/asum.cpp b/benchmark/cublas/blas1/asum.cpp index 2523560f0..3766fb2cf 100644 --- a/benchmark/cublas/blas1/asum.cpp +++ b/benchmark/cublas/blas1/asum.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename asum.cpp * diff --git a/benchmark/cublas/blas1/axpy.cpp b/benchmark/cublas/blas1/axpy.cpp index 170bc9b69..c8c0f6f05 100644 --- a/benchmark/cublas/blas1/axpy.cpp +++ b/benchmark/cublas/blas1/axpy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename axpy.cpp * diff --git a/benchmark/cublas/blas1/dot.cpp b/benchmark/cublas/blas1/dot.cpp index 0475de2ad..75ae06719 100644 --- a/benchmark/cublas/blas1/dot.cpp +++ b/benchmark/cublas/blas1/dot.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot.cpp * diff --git a/benchmark/cublas/blas1/iamax.cpp b/benchmark/cublas/blas1/iamax.cpp index d67a4e8bd..914ea8024 100644 --- a/benchmark/cublas/blas1/iamax.cpp +++ b/benchmark/cublas/blas1/iamax.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax.cpp * diff --git a/benchmark/cublas/blas1/iamin.cpp b/benchmark/cublas/blas1/iamin.cpp index 3a87a094d..98bc332f9 100644 --- a/benchmark/cublas/blas1/iamin.cpp +++ b/benchmark/cublas/blas1/iamin.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin.cpp * diff --git a/benchmark/cublas/blas1/nrm2.cpp b/benchmark/cublas/blas1/nrm2.cpp index 5a16bc4d2..edcae7e88 100644 --- a/benchmark/cublas/blas1/nrm2.cpp +++ b/benchmark/cublas/blas1/nrm2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2.cpp * diff --git a/benchmark/cublas/blas1/rotg.cpp b/benchmark/cublas/blas1/rotg.cpp index 08f0f9a7c..032ca6399 100644 --- a/benchmark/cublas/blas1/rotg.cpp +++ b/benchmark/cublas/blas1/rotg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotg.cpp * diff --git a/benchmark/cublas/blas1/rotm.cpp b/benchmark/cublas/blas1/rotm.cpp index cbb49a169..b92549578 100644 --- a/benchmark/cublas/blas1/rotm.cpp +++ b/benchmark/cublas/blas1/rotm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotm.cpp * diff --git a/benchmark/cublas/blas1/rotmg.cpp b/benchmark/cublas/blas1/rotmg.cpp index aa0bf310c..5e46d85e4 100644 --- a/benchmark/cublas/blas1/rotmg.cpp +++ b/benchmark/cublas/blas1/rotmg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotmg.cpp * diff --git a/benchmark/cublas/blas1/scal.cpp b/benchmark/cublas/blas1/scal.cpp index 0f07aa591..377aebe48 100644 --- a/benchmark/cublas/blas1/scal.cpp +++ b/benchmark/cublas/blas1/scal.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename scal.cpp * diff --git a/benchmark/cublas/blas2/gbmv.cpp b/benchmark/cublas/blas2/gbmv.cpp index fa04767e7..679968e43 100644 --- a/benchmark/cublas/blas2/gbmv.cpp +++ b/benchmark/cublas/blas2/gbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.cpp * diff --git a/benchmark/cublas/blas2/gemv.cpp b/benchmark/cublas/blas2/gemv.cpp index 58e66d1e1..3e3388890 100644 --- a/benchmark/cublas/blas2/gemv.cpp +++ b/benchmark/cublas/blas2/gemv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.cpp * diff --git a/benchmark/cublas/blas2/ger.cpp b/benchmark/cublas/blas2/ger.cpp index 8c21b92ab..77280e2ee 100644 --- a/benchmark/cublas/blas2/ger.cpp +++ b/benchmark/cublas/blas2/ger.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.cpp * diff --git a/benchmark/cublas/blas2/sbmv.cpp b/benchmark/cublas/blas2/sbmv.cpp index 51d333fda..ad8219b4b 100644 --- a/benchmark/cublas/blas2/sbmv.cpp +++ b/benchmark/cublas/blas2/sbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.cpp * diff --git a/benchmark/cublas/blas2/spmv.cpp b/benchmark/cublas/blas2/spmv.cpp index a1bd717ad..73ac3b131 100644 --- a/benchmark/cublas/blas2/spmv.cpp +++ b/benchmark/cublas/blas2/spmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spmv.cpp * diff --git a/benchmark/cublas/blas2/spr.cpp b/benchmark/cublas/blas2/spr.cpp index e8df09fa4..be82f07e6 100644 --- a/benchmark/cublas/blas2/spr.cpp +++ b/benchmark/cublas/blas2/spr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.cpp * diff --git a/benchmark/cublas/blas2/spr2.cpp b/benchmark/cublas/blas2/spr2.cpp index a5fbf1518..aac781902 100644 --- a/benchmark/cublas/blas2/spr2.cpp +++ b/benchmark/cublas/blas2/spr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr2.cpp * diff --git a/benchmark/cublas/blas2/symv.cpp b/benchmark/cublas/blas2/symv.cpp index 0e5c1eb80..a788bfd50 100644 --- a/benchmark/cublas/blas2/symv.cpp +++ b/benchmark/cublas/blas2/symv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symv.cpp * diff --git a/benchmark/cublas/blas2/syr.cpp b/benchmark/cublas/blas2/syr.cpp index 5ec8b8f52..fb1b8bf5a 100644 --- a/benchmark/cublas/blas2/syr.cpp +++ b/benchmark/cublas/blas2/syr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr.cpp * diff --git a/benchmark/cublas/blas2/syr2.cpp b/benchmark/cublas/blas2/syr2.cpp index 1f1f4d26e..8cfd269a2 100644 --- a/benchmark/cublas/blas2/syr2.cpp +++ b/benchmark/cublas/blas2/syr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2.cpp * diff --git a/benchmark/cublas/blas2/tbmv.cpp b/benchmark/cublas/blas2/tbmv.cpp index 41aed9b4a..f298707c2 100644 --- a/benchmark/cublas/blas2/tbmv.cpp +++ b/benchmark/cublas/blas2/tbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp * diff --git a/benchmark/cublas/blas2/tbsv.cpp b/benchmark/cublas/blas2/tbsv.cpp index 0110163d1..54a530508 100644 --- a/benchmark/cublas/blas2/tbsv.cpp +++ b/benchmark/cublas/blas2/tbsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbsv.cpp * diff --git a/benchmark/cublas/blas2/tpmv.cpp b/benchmark/cublas/blas2/tpmv.cpp index dfc87374b..0d38e486b 100644 --- a/benchmark/cublas/blas2/tpmv.cpp +++ b/benchmark/cublas/blas2/tpmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpmv.cpp * diff --git a/benchmark/cublas/blas2/tpsv.cpp b/benchmark/cublas/blas2/tpsv.cpp index b0c84adfd..8e3c7d670 100644 --- a/benchmark/cublas/blas2/tpsv.cpp +++ b/benchmark/cublas/blas2/tpsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpsv.cpp * diff --git a/benchmark/cublas/blas2/trmv.cpp b/benchmark/cublas/blas2/trmv.cpp index 6d2a7a053..665e14c72 100644 --- a/benchmark/cublas/blas2/trmv.cpp +++ b/benchmark/cublas/blas2/trmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmv.cpp * diff --git a/benchmark/cublas/blas2/trsv.cpp b/benchmark/cublas/blas2/trsv.cpp index b82dae4d7..a6057f7bd 100644 --- a/benchmark/cublas/blas2/trsv.cpp +++ b/benchmark/cublas/blas2/trsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp * diff --git a/benchmark/cublas/blas3/gemm.cpp b/benchmark/cublas/blas3/gemm.cpp index 8e442bbe4..5a103d032 100644 --- a/benchmark/cublas/blas3/gemm.cpp +++ b/benchmark/cublas/blas3/gemm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp * diff --git a/benchmark/cublas/blas3/gemm_batched.cpp b/benchmark/cublas/blas3/gemm_batched.cpp index e575e18c5..4cce28ff5 100644 --- a/benchmark/cublas/blas3/gemm_batched.cpp +++ b/benchmark/cublas/blas3/gemm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched.cpp * diff --git a/benchmark/cublas/blas3/gemm_batched_strided.cpp b/benchmark/cublas/blas3/gemm_batched_strided.cpp index 52cefc411..d96b7adfe 100644 --- a/benchmark/cublas/blas3/gemm_batched_strided.cpp +++ b/benchmark/cublas/blas3/gemm_batched_strided.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched_strided.cpp * diff --git a/benchmark/cublas/blas3/symm.cpp b/benchmark/cublas/blas3/symm.cpp index a1c119f53..1c0a1d6c2 100644 --- a/benchmark/cublas/blas3/symm.cpp +++ b/benchmark/cublas/blas3/symm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symm.cpp * diff --git a/benchmark/cublas/blas3/syr2k.cpp b/benchmark/cublas/blas3/syr2k.cpp index 79247fae5..62ee63151 100644 --- a/benchmark/cublas/blas3/syr2k.cpp +++ b/benchmark/cublas/blas3/syr2k.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2k.cpp * diff --git a/benchmark/cublas/blas3/syrk.cpp b/benchmark/cublas/blas3/syrk.cpp index 4555655d2..5a86e318b 100644 --- a/benchmark/cublas/blas3/syrk.cpp +++ b/benchmark/cublas/blas3/syrk.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syrk.cpp * diff --git a/benchmark/cublas/blas3/trmm.cpp b/benchmark/cublas/blas3/trmm.cpp index 47bfb36e7..fdbc13da5 100644 --- a/benchmark/cublas/blas3/trmm.cpp +++ b/benchmark/cublas/blas3/trmm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmm.cpp * diff --git a/benchmark/cublas/blas3/trsm.cpp b/benchmark/cublas/blas3/trsm.cpp index 4d39b712f..68eca24b5 100644 --- a/benchmark/cublas/blas3/trsm.cpp +++ b/benchmark/cublas/blas3/trsm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm.cpp * diff --git a/benchmark/cublas/blas3/trsm_batched.cpp b/benchmark/cublas/blas3/trsm_batched.cpp index 6bdb6bffc..544043415 100644 --- a/benchmark/cublas/blas3/trsm_batched.cpp +++ b/benchmark/cublas/blas3/trsm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm_batched.cpp * diff --git a/benchmark/cublas/extension/omatcopy.cpp b/benchmark/cublas/extension/omatcopy.cpp index f4d8c1dfa..2704ffe1b 100644 --- a/benchmark/cublas/extension/omatcopy.cpp +++ b/benchmark/cublas/extension/omatcopy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy.cpp * diff --git a/benchmark/cublas/main.cpp b/benchmark/cublas/main.cpp index d3d7c9cf5..83f276729 100644 --- a/benchmark/cublas/main.cpp +++ b/benchmark/cublas/main.cpp @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename main.cpp # * @@ -40,7 +40,7 @@ int main(int argc, char** argv) { // Initialize googlebench benchmark::Initialize(&argc, argv); - // Create a sycl blas sb_handle from the queue + // Create a portBLAS sb_handle from the queue cublasHandle_t cublas_handle = NULL; CUBLAS_CHECK(cublasCreate(&cublas_handle)); diff --git a/benchmark/cublas/utils.hpp b/benchmark/cublas/utils.hpp index 7b0bec507..eeaee7371 100644 --- a/benchmark/cublas/utils.hpp +++ b/benchmark/cublas/utils.hpp @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename utils.hpp # * @@ -30,7 +30,7 @@ #include #include "benchmark/benchmark.h" -#include "sycl_blas.h" +#include "portblas.h" #include #include diff --git a/benchmark/gen_param.py b/benchmark/gen_param.py index 5ee4cb20d..a8d4019bc 100644 --- a/benchmark/gen_param.py +++ b/benchmark/gen_param.py @@ -18,7 +18,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename gen_param.py # * diff --git a/benchmark/syclblas/CMakeLists.txt b/benchmark/portblas/CMakeLists.txt similarity index 86% rename from benchmark/syclblas/CMakeLists.txt rename to benchmark/portblas/CMakeLists.txt index 244155da9..00e73379d 100644 --- a/benchmark/syclblas/CMakeLists.txt +++ b/benchmark/portblas/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -74,16 +74,16 @@ if(${BLAS_ENABLE_EXTENSIONS}) endif() # Add individual benchmarks for each method -foreach(syclblas_bench ${sources}) - get_filename_component(bench_exec ${syclblas_bench} NAME_WE) - add_executable(bench_${bench_exec} ${syclblas_bench} main.cpp) - target_link_libraries(bench_${bench_exec} PRIVATE benchmark Clara::Clara sycl_blas bench_info) +foreach(portblas_bench ${sources}) + get_filename_component(bench_exec ${portblas_bench} NAME_WE) + add_executable(bench_${bench_exec} ${portblas_bench} main.cpp) + target_link_libraries(bench_${bench_exec} PRIVATE benchmark Clara::Clara portblas bench_info) target_compile_definitions(bench_${bench_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) add_sycl_to_target( TARGET bench_${bench_exec} - SOURCES ${syclblas_bench} + SOURCES ${portblas_bench} ) - target_include_directories(bench_${bench_exec} PRIVATE ${SYCLBLAS_INCLUDE} ${CBLAS_INCLUDE} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_include_directories(bench_${bench_exec} PRIVATE ${PORTBLAS_INCLUDE} ${CBLAS_INCLUDE} ${PORTBLAS_COMMON_INCLUDE_DIR}) if(BLAS_VERIFY_BENCHMARK) target_compile_definitions(bench_${bench_exec} PRIVATE BLAS_VERIFY_BENCHMARK) diff --git a/benchmark/syclblas/blas1/asum.cpp b/benchmark/portblas/blas1/asum.cpp similarity index 98% rename from benchmark/syclblas/blas1/asum.cpp rename to benchmark/portblas/blas1/asum.cpp index f42502c2e..473d192db 100644 --- a/benchmark/syclblas/blas1/asum.cpp +++ b/benchmark/portblas/blas1/asum.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename asum.cpp * diff --git a/benchmark/syclblas/blas1/axpy.cpp b/benchmark/portblas/blas1/axpy.cpp similarity index 98% rename from benchmark/syclblas/blas1/axpy.cpp rename to benchmark/portblas/blas1/axpy.cpp index db29313d5..4ccfaec52 100644 --- a/benchmark/syclblas/blas1/axpy.cpp +++ b/benchmark/portblas/blas1/axpy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename axpy.cpp * diff --git a/benchmark/syclblas/blas1/copy.cpp b/benchmark/portblas/blas1/copy.cpp similarity index 98% rename from benchmark/syclblas/blas1/copy.cpp rename to benchmark/portblas/blas1/copy.cpp index af5657066..bb8aeacdd 100644 --- a/benchmark/syclblas/blas1/copy.cpp +++ b/benchmark/portblas/blas1/copy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename copy.cpp * diff --git a/benchmark/syclblas/blas1/dot.cpp b/benchmark/portblas/blas1/dot.cpp similarity index 99% rename from benchmark/syclblas/blas1/dot.cpp rename to benchmark/portblas/blas1/dot.cpp index b3d516454..3ad379b2f 100644 --- a/benchmark/syclblas/blas1/dot.cpp +++ b/benchmark/portblas/blas1/dot.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot.cpp * diff --git a/benchmark/syclblas/blas1/iamax.cpp b/benchmark/portblas/blas1/iamax.cpp similarity index 98% rename from benchmark/syclblas/blas1/iamax.cpp rename to benchmark/portblas/blas1/iamax.cpp index 35e920d9f..f0aa50bca 100644 --- a/benchmark/syclblas/blas1/iamax.cpp +++ b/benchmark/portblas/blas1/iamax.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax.cpp * diff --git a/benchmark/syclblas/blas1/iamin.cpp b/benchmark/portblas/blas1/iamin.cpp similarity index 98% rename from benchmark/syclblas/blas1/iamin.cpp rename to benchmark/portblas/blas1/iamin.cpp index 14facae00..37364d20f 100644 --- a/benchmark/syclblas/blas1/iamin.cpp +++ b/benchmark/portblas/blas1/iamin.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin.cpp * diff --git a/benchmark/syclblas/blas1/nrm2.cpp b/benchmark/portblas/blas1/nrm2.cpp similarity index 98% rename from benchmark/syclblas/blas1/nrm2.cpp rename to benchmark/portblas/blas1/nrm2.cpp index 9f7d40243..10365d5b4 100644 --- a/benchmark/syclblas/blas1/nrm2.cpp +++ b/benchmark/portblas/blas1/nrm2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2.cpp * diff --git a/benchmark/syclblas/blas1/rotg.cpp b/benchmark/portblas/blas1/rotg.cpp similarity index 99% rename from benchmark/syclblas/blas1/rotg.cpp rename to benchmark/portblas/blas1/rotg.cpp index 33075ddb7..e09276f70 100644 --- a/benchmark/syclblas/blas1/rotg.cpp +++ b/benchmark/portblas/blas1/rotg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotg.cpp * diff --git a/benchmark/syclblas/blas1/rotm.cpp b/benchmark/portblas/blas1/rotm.cpp similarity index 99% rename from benchmark/syclblas/blas1/rotm.cpp rename to benchmark/portblas/blas1/rotm.cpp index 414a2a31e..8da49302a 100644 --- a/benchmark/syclblas/blas1/rotm.cpp +++ b/benchmark/portblas/blas1/rotm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotm.cpp * diff --git a/benchmark/syclblas/blas1/rotmg.cpp b/benchmark/portblas/blas1/rotmg.cpp similarity index 99% rename from benchmark/syclblas/blas1/rotmg.cpp rename to benchmark/portblas/blas1/rotmg.cpp index b89072afd..1e009d8b0 100644 --- a/benchmark/syclblas/blas1/rotmg.cpp +++ b/benchmark/portblas/blas1/rotmg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotmg.cpp * diff --git a/benchmark/syclblas/blas1/scal.cpp b/benchmark/portblas/blas1/scal.cpp similarity index 98% rename from benchmark/syclblas/blas1/scal.cpp rename to benchmark/portblas/blas1/scal.cpp index 3fcf6dbf9..fee25a4d2 100644 --- a/benchmark/syclblas/blas1/scal.cpp +++ b/benchmark/portblas/blas1/scal.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename scal.cpp * diff --git a/benchmark/syclblas/blas1/sdsdot.cpp b/benchmark/portblas/blas1/sdsdot.cpp similarity index 99% rename from benchmark/syclblas/blas1/sdsdot.cpp rename to benchmark/portblas/blas1/sdsdot.cpp index 649f422ca..963b28f0d 100644 --- a/benchmark/syclblas/blas1/sdsdot.cpp +++ b/benchmark/portblas/blas1/sdsdot.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sdsdot.cpp * diff --git a/benchmark/syclblas/blas2/gbmv.cpp b/benchmark/portblas/blas2/gbmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/gbmv.cpp rename to benchmark/portblas/blas2/gbmv.cpp index 1c4fc0cc7..2c878e9af 100644 --- a/benchmark/syclblas/blas2/gbmv.cpp +++ b/benchmark/portblas/blas2/gbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.cpp * diff --git a/benchmark/syclblas/blas2/gemv.cpp b/benchmark/portblas/blas2/gemv.cpp similarity index 99% rename from benchmark/syclblas/blas2/gemv.cpp rename to benchmark/portblas/blas2/gemv.cpp index 2d2f5e822..70e359f87 100644 --- a/benchmark/syclblas/blas2/gemv.cpp +++ b/benchmark/portblas/blas2/gemv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.cpp * diff --git a/benchmark/syclblas/blas2/ger.cpp b/benchmark/portblas/blas2/ger.cpp similarity index 99% rename from benchmark/syclblas/blas2/ger.cpp rename to benchmark/portblas/blas2/ger.cpp index 1968f1b55..088444357 100644 --- a/benchmark/syclblas/blas2/ger.cpp +++ b/benchmark/portblas/blas2/ger.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.cpp * diff --git a/benchmark/syclblas/blas2/sbmv.cpp b/benchmark/portblas/blas2/sbmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/sbmv.cpp rename to benchmark/portblas/blas2/sbmv.cpp index 7ae2ff4e5..efacb6bcb 100644 --- a/benchmark/syclblas/blas2/sbmv.cpp +++ b/benchmark/portblas/blas2/sbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.cpp * diff --git a/benchmark/syclblas/blas2/spmv.cpp b/benchmark/portblas/blas2/spmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/spmv.cpp rename to benchmark/portblas/blas2/spmv.cpp index b003acce7..b81e06f02 100644 --- a/benchmark/syclblas/blas2/spmv.cpp +++ b/benchmark/portblas/blas2/spmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spmv.cpp * diff --git a/benchmark/syclblas/blas2/spr.cpp b/benchmark/portblas/blas2/spr.cpp similarity index 99% rename from benchmark/syclblas/blas2/spr.cpp rename to benchmark/portblas/blas2/spr.cpp index ee7f2ebd0..b0ac4d16a 100644 --- a/benchmark/syclblas/blas2/spr.cpp +++ b/benchmark/portblas/blas2/spr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.cpp * diff --git a/benchmark/syclblas/blas2/spr2.cpp b/benchmark/portblas/blas2/spr2.cpp similarity index 99% rename from benchmark/syclblas/blas2/spr2.cpp rename to benchmark/portblas/blas2/spr2.cpp index 8a1b3f324..057744148 100644 --- a/benchmark/syclblas/blas2/spr2.cpp +++ b/benchmark/portblas/blas2/spr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr2.cpp * diff --git a/benchmark/syclblas/blas2/symv.cpp b/benchmark/portblas/blas2/symv.cpp similarity index 99% rename from benchmark/syclblas/blas2/symv.cpp rename to benchmark/portblas/blas2/symv.cpp index fc4eedfb3..26023d942 100644 --- a/benchmark/syclblas/blas2/symv.cpp +++ b/benchmark/portblas/blas2/symv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symv.cpp * diff --git a/benchmark/syclblas/blas2/syr.cpp b/benchmark/portblas/blas2/syr.cpp similarity index 99% rename from benchmark/syclblas/blas2/syr.cpp rename to benchmark/portblas/blas2/syr.cpp index d7d3ec30a..14f862482 100644 --- a/benchmark/syclblas/blas2/syr.cpp +++ b/benchmark/portblas/blas2/syr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr.cpp * diff --git a/benchmark/syclblas/blas2/syr2.cpp b/benchmark/portblas/blas2/syr2.cpp similarity index 99% rename from benchmark/syclblas/blas2/syr2.cpp rename to benchmark/portblas/blas2/syr2.cpp index cad0f537b..a72222c1b 100644 --- a/benchmark/syclblas/blas2/syr2.cpp +++ b/benchmark/portblas/blas2/syr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2.cpp * diff --git a/benchmark/syclblas/blas2/tbmv.cpp b/benchmark/portblas/blas2/tbmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/tbmv.cpp rename to benchmark/portblas/blas2/tbmv.cpp index bf4880414..f7ee22a11 100644 --- a/benchmark/syclblas/blas2/tbmv.cpp +++ b/benchmark/portblas/blas2/tbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp * diff --git a/benchmark/syclblas/blas2/tbsv.cpp b/benchmark/portblas/blas2/tbsv.cpp similarity index 99% rename from benchmark/syclblas/blas2/tbsv.cpp rename to benchmark/portblas/blas2/tbsv.cpp index 7b8cdeffc..6d3107cd5 100644 --- a/benchmark/syclblas/blas2/tbsv.cpp +++ b/benchmark/portblas/blas2/tbsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbsv.cpp * diff --git a/benchmark/syclblas/blas2/tpmv.cpp b/benchmark/portblas/blas2/tpmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/tpmv.cpp rename to benchmark/portblas/blas2/tpmv.cpp index fab71fac4..116f65128 100644 --- a/benchmark/syclblas/blas2/tpmv.cpp +++ b/benchmark/portblas/blas2/tpmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpmv.cpp * diff --git a/benchmark/syclblas/blas2/tpsv.cpp b/benchmark/portblas/blas2/tpsv.cpp similarity index 99% rename from benchmark/syclblas/blas2/tpsv.cpp rename to benchmark/portblas/blas2/tpsv.cpp index c9039a4b3..75011515a 100644 --- a/benchmark/syclblas/blas2/tpsv.cpp +++ b/benchmark/portblas/blas2/tpsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpsv.cpp * diff --git a/benchmark/syclblas/blas2/trmv.cpp b/benchmark/portblas/blas2/trmv.cpp similarity index 99% rename from benchmark/syclblas/blas2/trmv.cpp rename to benchmark/portblas/blas2/trmv.cpp index c4a95d975..1a635b8bc 100644 --- a/benchmark/syclblas/blas2/trmv.cpp +++ b/benchmark/portblas/blas2/trmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmv.cpp * diff --git a/benchmark/syclblas/blas2/trsv.cpp b/benchmark/portblas/blas2/trsv.cpp similarity index 99% rename from benchmark/syclblas/blas2/trsv.cpp rename to benchmark/portblas/blas2/trsv.cpp index a5e9f143b..2ee7b604c 100644 --- a/benchmark/syclblas/blas2/trsv.cpp +++ b/benchmark/portblas/blas2/trsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp * diff --git a/benchmark/syclblas/blas3/gemm.cpp b/benchmark/portblas/blas3/gemm.cpp similarity index 99% rename from benchmark/syclblas/blas3/gemm.cpp rename to benchmark/portblas/blas3/gemm.cpp index a87c10559..1a49ac3e6 100644 --- a/benchmark/syclblas/blas3/gemm.cpp +++ b/benchmark/portblas/blas3/gemm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp * diff --git a/benchmark/syclblas/blas3/gemm_batched.cpp b/benchmark/portblas/blas3/gemm_batched.cpp similarity index 99% rename from benchmark/syclblas/blas3/gemm_batched.cpp rename to benchmark/portblas/blas3/gemm_batched.cpp index 64c928d3b..9cde87cc0 100644 --- a/benchmark/syclblas/blas3/gemm_batched.cpp +++ b/benchmark/portblas/blas3/gemm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched.cpp * diff --git a/benchmark/syclblas/blas3/gemm_batched_strided.cpp b/benchmark/portblas/blas3/gemm_batched_strided.cpp similarity index 99% rename from benchmark/syclblas/blas3/gemm_batched_strided.cpp rename to benchmark/portblas/blas3/gemm_batched_strided.cpp index c47ab55ad..d741f3e8a 100644 --- a/benchmark/syclblas/blas3/gemm_batched_strided.cpp +++ b/benchmark/portblas/blas3/gemm_batched_strided.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched_strided.cpp * diff --git a/benchmark/syclblas/blas3/symm.cpp b/benchmark/portblas/blas3/symm.cpp similarity index 99% rename from benchmark/syclblas/blas3/symm.cpp rename to benchmark/portblas/blas3/symm.cpp index 1448ab7d4..46f493421 100644 --- a/benchmark/syclblas/blas3/symm.cpp +++ b/benchmark/portblas/blas3/symm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symm.cpp * diff --git a/benchmark/syclblas/blas3/trsm.cpp b/benchmark/portblas/blas3/trsm.cpp similarity index 99% rename from benchmark/syclblas/blas3/trsm.cpp rename to benchmark/portblas/blas3/trsm.cpp index bdafd4558..a716e4f13 100644 --- a/benchmark/syclblas/blas3/trsm.cpp +++ b/benchmark/portblas/blas3/trsm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm.cpp * diff --git a/benchmark/syclblas/extension/omatadd.cpp b/benchmark/portblas/extension/omatadd.cpp similarity index 99% rename from benchmark/syclblas/extension/omatadd.cpp rename to benchmark/portblas/extension/omatadd.cpp index dc671f5a6..4ef561c62 100644 --- a/benchmark/syclblas/extension/omatadd.cpp +++ b/benchmark/portblas/extension/omatadd.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatadd.cpp * diff --git a/benchmark/syclblas/extension/omatcopy.cpp b/benchmark/portblas/extension/omatcopy.cpp similarity index 99% rename from benchmark/syclblas/extension/omatcopy.cpp rename to benchmark/portblas/extension/omatcopy.cpp index 45213681e..8043fda6c 100644 --- a/benchmark/syclblas/extension/omatcopy.cpp +++ b/benchmark/portblas/extension/omatcopy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy.cpp * diff --git a/benchmark/syclblas/extension/omatcopy2.cpp b/benchmark/portblas/extension/omatcopy2.cpp similarity index 99% rename from benchmark/syclblas/extension/omatcopy2.cpp rename to benchmark/portblas/extension/omatcopy2.cpp index 28fa66547..0b8e4b5c4 100644 --- a/benchmark/syclblas/extension/omatcopy2.cpp +++ b/benchmark/portblas/extension/omatcopy2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy2.cpp * diff --git a/benchmark/syclblas/extension/reduction.cpp b/benchmark/portblas/extension/reduction.cpp similarity index 99% rename from benchmark/syclblas/extension/reduction.cpp rename to benchmark/portblas/extension/reduction.cpp index 78bba15b1..446c29f10 100644 --- a/benchmark/syclblas/extension/reduction.cpp +++ b/benchmark/portblas/extension/reduction.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction.cpp * diff --git a/benchmark/syclblas/main.cpp b/benchmark/portblas/main.cpp similarity index 94% rename from benchmark/syclblas/main.cpp rename to benchmark/portblas/main.cpp index f7ec90fab..a5a2e6813 100644 --- a/benchmark/syclblas/main.cpp +++ b/benchmark/portblas/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * @@ -27,7 +27,7 @@ #include #include -// Create a shared pointer to a sycl blas Handle, so that we don't keep +// Create a shared pointer to a portBLAS Handle, so that we don't keep // reconstructing it each time (which is slow). Although this won't be // cleaned up if RunSpecifiedBenchmarks exits badly, that's okay, as those // are presumably exceptional circumstances. @@ -66,7 +66,7 @@ int main(int argc, char** argv) { utils::print_queue_information(q); - // Create a sycl blas sb_handle from the queue + // Create a portBLAS sb_handle from the queue blas::SB_Handle sb_handle(q); // This will be set to false by a failing benchmark diff --git a/benchmark/syclblas/utils.hpp b/benchmark/portblas/utils.hpp similarity index 96% rename from benchmark/syclblas/utils.hpp rename to benchmark/portblas/utils.hpp index 0c326fefb..b16bc78b6 100644 --- a/benchmark/syclblas/utils.hpp +++ b/benchmark/portblas/utils.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * @@ -30,7 +30,7 @@ #include #include -#include "sycl_blas.h" +#include "portblas.h" #include // Forward declare methods that we use in `benchmark.cpp`, but define in diff --git a/benchmark/rocblas/CMakeLists.txt b/benchmark/rocblas/CMakeLists.txt index e23989801..caa884725 100644 --- a/benchmark/rocblas/CMakeLists.txt +++ b/benchmark/rocblas/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * @@ -83,7 +83,7 @@ foreach(rocblas_benchmark ${sources}) add_executable(bench_rocblas_${rocblas_bench_exec} ${rocblas_benchmark} main.cpp) target_link_libraries(bench_rocblas_${rocblas_bench_exec} PRIVATE benchmark Clara::Clara roc::rocblas bench_info) target_compile_definitions(bench_rocblas_${rocblas_bench_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_BENCHMARK_INDEX_TYPE}) - target_include_directories(bench_rocblas_${rocblas_bench_exec} PRIVATE ${SYCLBLAS_INCLUDE} ${rocblas_INCLUDE_DIRS} ${CBLAS_INCLUDE} ${BLAS_BENCH} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_include_directories(bench_rocblas_${rocblas_bench_exec} PRIVATE ${PORTBLAS_INCLUDE} ${rocblas_INCLUDE_DIRS} ${CBLAS_INCLUDE} ${BLAS_BENCH} ${PORTBLAS_COMMON_INCLUDE_DIR}) # Even though rocblas does not use sycl, the common tools indirectly include sycl headers. add_sycl_to_target( diff --git a/benchmark/rocblas/blas1/asum.cpp b/benchmark/rocblas/blas1/asum.cpp index 550bb7006..d97fbefec 100644 --- a/benchmark/rocblas/blas1/asum.cpp +++ b/benchmark/rocblas/blas1/asum.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename asum.cpp * diff --git a/benchmark/rocblas/blas1/axpy.cpp b/benchmark/rocblas/blas1/axpy.cpp index 5c34fa83d..7eafea1bd 100644 --- a/benchmark/rocblas/blas1/axpy.cpp +++ b/benchmark/rocblas/blas1/axpy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename axpy.cpp * diff --git a/benchmark/rocblas/blas1/dot.cpp b/benchmark/rocblas/blas1/dot.cpp index 9a4c3dbc5..74b4a279a 100644 --- a/benchmark/rocblas/blas1/dot.cpp +++ b/benchmark/rocblas/blas1/dot.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot.cpp * diff --git a/benchmark/rocblas/blas1/iamax.cpp b/benchmark/rocblas/blas1/iamax.cpp index bb99e15e2..8a82ba523 100644 --- a/benchmark/rocblas/blas1/iamax.cpp +++ b/benchmark/rocblas/blas1/iamax.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax.cpp * diff --git a/benchmark/rocblas/blas1/iamin.cpp b/benchmark/rocblas/blas1/iamin.cpp index 83a280b8c..721f327fd 100644 --- a/benchmark/rocblas/blas1/iamin.cpp +++ b/benchmark/rocblas/blas1/iamin.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin.cpp * diff --git a/benchmark/rocblas/blas1/nrm2.cpp b/benchmark/rocblas/blas1/nrm2.cpp index 66a9f2f2b..7a093e762 100644 --- a/benchmark/rocblas/blas1/nrm2.cpp +++ b/benchmark/rocblas/blas1/nrm2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2.cpp * diff --git a/benchmark/rocblas/blas1/rotg.cpp b/benchmark/rocblas/blas1/rotg.cpp index 5cdff1b88..b706f045e 100644 --- a/benchmark/rocblas/blas1/rotg.cpp +++ b/benchmark/rocblas/blas1/rotg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotg.cpp * diff --git a/benchmark/rocblas/blas1/rotm.cpp b/benchmark/rocblas/blas1/rotm.cpp index 4d00ff1ee..7b28ad2d5 100644 --- a/benchmark/rocblas/blas1/rotm.cpp +++ b/benchmark/rocblas/blas1/rotm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotm.cpp * diff --git a/benchmark/rocblas/blas1/rotmg.cpp b/benchmark/rocblas/blas1/rotmg.cpp index 521ad8f36..07a86244b 100644 --- a/benchmark/rocblas/blas1/rotmg.cpp +++ b/benchmark/rocblas/blas1/rotmg.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotmg.cpp * diff --git a/benchmark/rocblas/blas1/scal.cpp b/benchmark/rocblas/blas1/scal.cpp index dcb030003..a31768791 100644 --- a/benchmark/rocblas/blas1/scal.cpp +++ b/benchmark/rocblas/blas1/scal.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename scal.cpp * diff --git a/benchmark/rocblas/blas2/gbmv.cpp b/benchmark/rocblas/blas2/gbmv.cpp index 86a928bee..be21ce9bc 100644 --- a/benchmark/rocblas/blas2/gbmv.cpp +++ b/benchmark/rocblas/blas2/gbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.cpp * diff --git a/benchmark/rocblas/blas2/gemv.cpp b/benchmark/rocblas/blas2/gemv.cpp index 891715dd9..38b80519d 100644 --- a/benchmark/rocblas/blas2/gemv.cpp +++ b/benchmark/rocblas/blas2/gemv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.cpp * diff --git a/benchmark/rocblas/blas2/ger.cpp b/benchmark/rocblas/blas2/ger.cpp index 1e1605db1..819f3ec9a 100644 --- a/benchmark/rocblas/blas2/ger.cpp +++ b/benchmark/rocblas/blas2/ger.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.cpp * diff --git a/benchmark/rocblas/blas2/sbmv.cpp b/benchmark/rocblas/blas2/sbmv.cpp index 1822208bb..ad913dbb5 100644 --- a/benchmark/rocblas/blas2/sbmv.cpp +++ b/benchmark/rocblas/blas2/sbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.cpp * diff --git a/benchmark/rocblas/blas2/spmv.cpp b/benchmark/rocblas/blas2/spmv.cpp index a7b3defc1..700ec321b 100644 --- a/benchmark/rocblas/blas2/spmv.cpp +++ b/benchmark/rocblas/blas2/spmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spmv.cpp * diff --git a/benchmark/rocblas/blas2/spr.cpp b/benchmark/rocblas/blas2/spr.cpp index 886b126d0..6e4e98c33 100644 --- a/benchmark/rocblas/blas2/spr.cpp +++ b/benchmark/rocblas/blas2/spr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.cpp * diff --git a/benchmark/rocblas/blas2/spr2.cpp b/benchmark/rocblas/blas2/spr2.cpp index e68ec2818..7a1178160 100644 --- a/benchmark/rocblas/blas2/spr2.cpp +++ b/benchmark/rocblas/blas2/spr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr2.cpp * diff --git a/benchmark/rocblas/blas2/symv.cpp b/benchmark/rocblas/blas2/symv.cpp index 0625fcbc5..1e95d40c8 100644 --- a/benchmark/rocblas/blas2/symv.cpp +++ b/benchmark/rocblas/blas2/symv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symv.cpp * diff --git a/benchmark/rocblas/blas2/syr.cpp b/benchmark/rocblas/blas2/syr.cpp index ecc51691d..578f16367 100644 --- a/benchmark/rocblas/blas2/syr.cpp +++ b/benchmark/rocblas/blas2/syr.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr.cpp * diff --git a/benchmark/rocblas/blas2/syr2.cpp b/benchmark/rocblas/blas2/syr2.cpp index 731d62f4b..926f1cf4f 100644 --- a/benchmark/rocblas/blas2/syr2.cpp +++ b/benchmark/rocblas/blas2/syr2.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2.cpp * diff --git a/benchmark/rocblas/blas2/tbmv.cpp b/benchmark/rocblas/blas2/tbmv.cpp index d35b930ce..4eb4f91df 100644 --- a/benchmark/rocblas/blas2/tbmv.cpp +++ b/benchmark/rocblas/blas2/tbmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp * diff --git a/benchmark/rocblas/blas2/tbsv.cpp b/benchmark/rocblas/blas2/tbsv.cpp index 4b22c2986..455fe9207 100644 --- a/benchmark/rocblas/blas2/tbsv.cpp +++ b/benchmark/rocblas/blas2/tbsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbsv.cpp * diff --git a/benchmark/rocblas/blas2/tpmv.cpp b/benchmark/rocblas/blas2/tpmv.cpp index 93ca8fb6c..7079fb52c 100644 --- a/benchmark/rocblas/blas2/tpmv.cpp +++ b/benchmark/rocblas/blas2/tpmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpmv.cpp * diff --git a/benchmark/rocblas/blas2/tpsv.cpp b/benchmark/rocblas/blas2/tpsv.cpp index 9203832ef..21fda5e44 100644 --- a/benchmark/rocblas/blas2/tpsv.cpp +++ b/benchmark/rocblas/blas2/tpsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpsv.cpp * diff --git a/benchmark/rocblas/blas2/trmv.cpp b/benchmark/rocblas/blas2/trmv.cpp index a780f46ba..b4f8fb929 100644 --- a/benchmark/rocblas/blas2/trmv.cpp +++ b/benchmark/rocblas/blas2/trmv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmv.cpp * diff --git a/benchmark/rocblas/blas2/trsv.cpp b/benchmark/rocblas/blas2/trsv.cpp index d3a5cf11f..d95f5f524 100644 --- a/benchmark/rocblas/blas2/trsv.cpp +++ b/benchmark/rocblas/blas2/trsv.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp * diff --git a/benchmark/rocblas/blas3/gemm.cpp b/benchmark/rocblas/blas3/gemm.cpp index c254ff377..b403bafec 100644 --- a/benchmark/rocblas/blas3/gemm.cpp +++ b/benchmark/rocblas/blas3/gemm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp * diff --git a/benchmark/rocblas/blas3/gemm_batched.cpp b/benchmark/rocblas/blas3/gemm_batched.cpp index 62b1cbced..4cfb1418d 100644 --- a/benchmark/rocblas/blas3/gemm_batched.cpp +++ b/benchmark/rocblas/blas3/gemm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched.cpp * diff --git a/benchmark/rocblas/blas3/gemm_batched_strided.cpp b/benchmark/rocblas/blas3/gemm_batched_strided.cpp index cdac699ca..15dac9896 100644 --- a/benchmark/rocblas/blas3/gemm_batched_strided.cpp +++ b/benchmark/rocblas/blas3/gemm_batched_strided.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_batched_strided.cpp * diff --git a/benchmark/rocblas/blas3/symm.cpp b/benchmark/rocblas/blas3/symm.cpp index c2fee6d52..ff205cea2 100644 --- a/benchmark/rocblas/blas3/symm.cpp +++ b/benchmark/rocblas/blas3/symm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symm.cpp * diff --git a/benchmark/rocblas/blas3/syr2k.cpp b/benchmark/rocblas/blas3/syr2k.cpp index fb2dee110..12b9fb573 100644 --- a/benchmark/rocblas/blas3/syr2k.cpp +++ b/benchmark/rocblas/blas3/syr2k.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2k.cpp * diff --git a/benchmark/rocblas/blas3/syrk.cpp b/benchmark/rocblas/blas3/syrk.cpp index 4cb909e4b..697a6da66 100644 --- a/benchmark/rocblas/blas3/syrk.cpp +++ b/benchmark/rocblas/blas3/syrk.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syrk.cpp * diff --git a/benchmark/rocblas/blas3/trmm.cpp b/benchmark/rocblas/blas3/trmm.cpp index 09e1d4933..60e05a69c 100644 --- a/benchmark/rocblas/blas3/trmm.cpp +++ b/benchmark/rocblas/blas3/trmm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmm.cpp * diff --git a/benchmark/rocblas/blas3/trsm.cpp b/benchmark/rocblas/blas3/trsm.cpp index dbf72bf3e..da27970e1 100644 --- a/benchmark/rocblas/blas3/trsm.cpp +++ b/benchmark/rocblas/blas3/trsm.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm.cpp * diff --git a/benchmark/rocblas/blas3/trsm_batched.cpp b/benchmark/rocblas/blas3/trsm_batched.cpp index e6d9d89f8..e43894c51 100644 --- a/benchmark/rocblas/blas3/trsm_batched.cpp +++ b/benchmark/rocblas/blas3/trsm_batched.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsm_batched.cpp * diff --git a/benchmark/rocblas/extension/omatcopy.cpp b/benchmark/rocblas/extension/omatcopy.cpp index 5ae607627..5d0e7862d 100644 --- a/benchmark/rocblas/extension/omatcopy.cpp +++ b/benchmark/rocblas/extension/omatcopy.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy.cpp * diff --git a/benchmark/rocblas/main.cpp b/benchmark/rocblas/main.cpp index e3236e306..cd0407543 100644 --- a/benchmark/rocblas/main.cpp +++ b/benchmark/rocblas/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/benchmark/rocblas/utils.hpp b/benchmark/rocblas/utils.hpp index ddf1ecef0..d88832225 100644 --- a/benchmark/rocblas/utils.hpp +++ b/benchmark/rocblas/utils.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * @@ -26,7 +26,7 @@ #ifndef ROCBLAS_UTILS_HPP #define ROCBLAS_UTILS_HPP -#include "sycl_blas.h" +#include "portblas.h" #include #include diff --git a/cmake/CmakeFunctionHelper.cmake b/cmake/CmakeFunctionHelper.cmake index daacdb9d5..f6f7ccae1 100644 --- a/cmake/CmakeFunctionHelper.cmake +++ b/cmake/CmakeFunctionHelper.cmake @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CmakeFUnctionHelper.cmake # * @@ -47,7 +47,7 @@ set(boolean_list "true" "false") function(sanitize_file_name output file_name) string(REGEX REPLACE "(:|\\*|<| |,|>)" "_" file_name ${file_name}) string(REGEX REPLACE "(_____|____|___|__)" "_" file_name ${file_name}) - if (SYCLBLAS_USE_SHORT_NAMES) + if (PORTBLAS_USE_SHORT_NAMES) # Long paths are problematic on Windows and WSL so we hash the filename # to reduce its size string(MD5 file_name ${file_name}) @@ -89,7 +89,7 @@ endfunction() # blas unary function for generating source code function(generate_blas_unary_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach(data ${data_list}) cpp_type(cpp_data ${data}) set(container_list "BufferIterator<${cpp_data}>") @@ -99,19 +99,19 @@ foreach(data ${data_list}) sanitize_file_name(file_name "${func}_${data}_${index}_${container0}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_unary.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_unary.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${index} ${increment} ${container0} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_unary.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_unary.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -122,8 +122,8 @@ foreach(data ${data_list}) endforeach(data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_unary_objects) @@ -131,7 +131,7 @@ endfunction(generate_blas_unary_objects) # blas binary function for generating source code function(generate_blas_binary_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach(data ${data_list}) cpp_type(cpp_data ${data}) set(container_list "BufferIterator<${cpp_data}>") @@ -143,20 +143,20 @@ foreach(data ${data_list}) sanitize_file_name(file_name "${func}_${data}_${index}_${container_names}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_binary.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_binary.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${index} ${increment} ${container0} ${container1} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_binary.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_binary.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -168,8 +168,8 @@ foreach(data ${data_list}) endforeach(data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_binary_objects) @@ -177,7 +177,7 @@ endfunction(generate_blas_binary_objects) # blas binary function for generating source code function(generate_blas_reduction_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") set(operator_list "AddOperator" "MinOperator" "MaxOperator" "ProductOperator" "AbsoluteAddOperator" "MeanOperator") string(FIND ${func} "_const" pos) if(pos) @@ -202,12 +202,12 @@ foreach(data ${data_list}) sanitize_file_name(file_name "${func}_${operator}_${data}_${index}_${container0}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_reduction.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_reduction.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in ${cpp_data} ${index} ${increment} @@ -215,8 +215,8 @@ foreach(data ${data_list}) ${container1} ${operator} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_reduction.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_reduction.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -229,8 +229,8 @@ foreach(data ${data_list}) endforeach(data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_reduction_objects) @@ -238,7 +238,7 @@ endfunction(generate_blas_reduction_objects) # blas special binary function for generating source code function(generate_blas_binary_special_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach(data ${data_list}) cpp_type(cpp_data ${data}) set(container_list_in "BufferIterator<${cpp_data}>") @@ -252,20 +252,20 @@ foreach(data ${data_list}) sanitize_file_name(file_name "${func}_${data}_${index}_${container_names}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_binary_special.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_binary_special.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${index} ${increment} ${container0} ${container1} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_binary_special.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_binary_special.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -277,8 +277,8 @@ foreach(data ${data_list}) endforeach(data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_binary_special_objects) @@ -287,7 +287,7 @@ endfunction(generate_blas_binary_special_objects) # blas ternary function for generating source code function(generate_blas_ternary_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") string(FIND ${func} "_const" const_pos) if(const_pos) string(REPLACE "_const" "" actualfunc ${func}) @@ -311,12 +311,12 @@ foreach(data ${data_list}) sanitize_file_name(file_name "${func}_${data}_${index}_${container_names}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_ternary.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_ternary.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in ${cpp_data} ${index} ${increment} @@ -324,8 +324,8 @@ foreach(data ${data_list}) ${container1} ${container2} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_ternary.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${actualfunc}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_ternary.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -338,8 +338,8 @@ foreach(data ${data_list}) endforeach(data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_ternary_objects) @@ -347,7 +347,7 @@ endfunction(generate_blas_ternary_objects) # blas function for generating source code for the rotg operator (asynchronous version with containers) function(generate_blas_rotg_objects blas_level func) - set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") + set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach (data ${data_list}) cpp_type(cpp_data ${data}) set(container_list_in_out "BufferIterator<${cpp_data}>") @@ -359,20 +359,20 @@ function(generate_blas_rotg_objects blas_level func) sanitize_file_name(file_name "${func}_${data}_${index}_${container_names}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotg.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotg.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${container0} ${container1} ${container2} ${container3} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotg.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotg.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -384,8 +384,8 @@ function(generate_blas_rotg_objects blas_level func) endforeach (data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) - target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) + target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_rotg_objects) @@ -393,23 +393,23 @@ endfunction(generate_blas_rotg_objects) # blas function for generating source code for the rotg operator (synchronous version) function(generate_blas_rotg_return_objects blas_level func) - set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") + set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach (data ${data_list}) cpp_type(cpp_data ${data}) set(container_list "BufferIterator<${cpp_data}>") sanitize_file_name(file_name "${func}_${data}_${index}_${container0}_${increment}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotg_return.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotg_return.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotg_return.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotg_return.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -417,15 +417,15 @@ function(generate_blas_rotg_return_objects blas_level func) endforeach (data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) - target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) + target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_rotg_return_objects) # blas function for generating source code for the rotg operator (asynchronous version with containers) function(generate_blas_rotmg_objects blas_level func) - set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") + set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") foreach (data ${data_list}) cpp_type(cpp_data ${data}) set(container_list_in_out "BufferIterator<${cpp_data}>") @@ -437,12 +437,12 @@ function(generate_blas_rotmg_objects blas_level func) set(container_names "${container0}_${container1}_${container2}_${container3}") sanitize_file_name(file_name "${func}_${data}_${container_names}.cpp") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotmg.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotmg.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${container0} ${container1} @@ -450,8 +450,8 @@ function(generate_blas_rotmg_objects blas_level func) ${container3} ${container4} ${file_name} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_rotmg.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_rotmg.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -464,8 +464,8 @@ function(generate_blas_rotmg_objects blas_level func) endforeach (data) add_library(${func} OBJECT ${FUNC_SRC}) set_target_compile_def(${func}) - target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) + target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${FUNC_SRC}) endfunction(generate_blas_rotmg_objects) @@ -473,7 +473,7 @@ endfunction(generate_blas_rotmg_objects) # blas gemm function for generating source code function(generate_blas_gemm_objects blas_level func) -set(LOCATION "${SYCLBLAS_GENERATED_SRC}/${blas_level}/${func}/") +set(LOCATION "${PORTBLAS_GENERATED_SRC}/${blas_level}/${func}/") set(gemm_sources "") # Generates a file for a new GEMM configuration @@ -532,12 +532,12 @@ function(add_gemm_configuration "${wg_size}_${cache_line_size}.cpp") sanitize_file_name(file_name "${file_name}") add_custom_command(OUTPUT "${LOCATION}/${file_name}" - COMMAND ${PYTHON_EXECUTABLE} ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_gemm_launcher.py + COMMAND ${PYTHON_EXECUTABLE} ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_gemm_launcher.py ${PROJECT_SOURCE_DIR}/external/ - ${SYCLBLAS_SRC_GENERATOR}/gen + ${PORTBLAS_SRC_GENERATOR}/gen ${blas_level} ${func} - ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in ${cpp_data} ${index} ${double_buffer} @@ -572,8 +572,8 @@ function(add_gemm_configuration ${use_joint_matrix} ${symm_a} ${symm_b} - MAIN_DEPENDENCY ${SYCLBLAS_SRC}/interface/${blas_level}/${func}.cpp.in - DEPENDS ${SYCLBLAS_SRC_GENERATOR}/py_gen_blas_gemm_launcher.py + MAIN_DEPENDENCY ${PORTBLAS_SRC}/interface/${blas_level}/${func}.cpp.in + DEPENDS ${PORTBLAS_SRC_GENERATOR}/py_gen_blas_gemm_launcher.py WORKING_DIRECTORY ${PROJECT_BINARY_DIR} VERBATIM ) @@ -778,8 +778,8 @@ endif() add_library(${func} OBJECT ${gemm_sources}) set_target_compile_def(${func}) # The blas library depends on FindComputeCpp -target_include_directories(${func} PRIVATE ${SYCLBLAS_SRC} ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) +target_include_directories(${func} PRIVATE ${PORTBLAS_SRC} ${PORTBLAS_INCLUDE} + ${PORTBLAS_COMMON_INCLUDE_DIR} ${THIRD_PARTIES_INCLUDE}) message(STATUS "Adding SYCL to target ${func}") add_sycl_to_target(TARGET ${func} SOURCES ${gemm_sources}) endfunction(generate_blas_gemm_objects) diff --git a/cmake/Modules/ConfigureSYCLBLAS.cmake b/cmake/Modules/ConfigurePORTBLAS.cmake similarity index 97% rename from cmake/Modules/ConfigureSYCLBLAS.cmake rename to cmake/Modules/ConfigurePORTBLAS.cmake index 7fd297e62..5112e9d02 100644 --- a/cmake/Modules/ConfigureSYCLBLAS.cmake +++ b/cmake/Modules/ConfigurePORTBLAS.cmake @@ -18,9 +18,9 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * -# * @filename ConfigureSYCLBLAS.cmake +# * @filename ConfigurePORTBLAS.cmake # * # **************************************************************************/ diff --git a/cmake/Modules/FindACL.cmake b/cmake/Modules/FindACL.cmake index 63816a047..8bdaee89e 100644 --- a/cmake/Modules/FindACL.cmake +++ b/cmake/Modules/FindACL.cmake @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename FindCLBlast.cmake # * diff --git a/cmake/Modules/FindDPCPP.cmake b/cmake/Modules/FindDPCPP.cmake index e55837105..099444c58 100644 --- a/cmake/Modules/FindDPCPP.cmake +++ b/cmake/Modules/FindDPCPP.cmake @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename FindDPCPP.cmake # * diff --git a/cmake/Modules/FindSB_CLBlast.cmake b/cmake/Modules/FindSB_CLBlast.cmake index 0d41d50b1..be7ed18f2 100644 --- a/cmake/Modules/FindSB_CLBlast.cmake +++ b/cmake/Modules/FindSB_CLBlast.cmake @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename FindCLBlast.cmake # * diff --git a/cmake/Modules/FindSystemBLAS.cmake b/cmake/Modules/FindSystemBLAS.cmake index 2e2ecde6a..aa95b731e 100644 --- a/cmake/Modules/FindSystemBLAS.cmake +++ b/cmake/Modules/FindSystemBLAS.cmake @@ -18,7 +18,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename FindSystemBLAS.cmake # * diff --git a/cmake/Modules/SYCL.cmake b/cmake/Modules/SYCL.cmake index 0d6fbbdb6..42cd90f85 100644 --- a/cmake/Modules/SYCL.cmake +++ b/cmake/Modules/SYCL.cmake @@ -17,13 +17,13 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * # **************************************************************************/ include(CheckCXXCompilerFlag) -include(ConfigureSYCLBLAS) +include(ConfigurePORTBLAS) # find_package(hipSYCL) requires HIPSYCL_TARGETS to be set, so set it to a default value before find_package(hipSYCL) if(SYCL_COMPILER MATCHES "hipsycl" AND NOT HIPSYCL_TARGETS AND NOT ENV{HIPSYCL_TARGETS}) diff --git a/common/include/common/benchmark_identifier.hpp b/common/include/common/benchmark_identifier.hpp index 4f111adf6..d0fa906cf 100644 --- a/common/include/common/benchmark_identifier.hpp +++ b/common/include/common/benchmark_identifier.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename benchmark_identifier.hpp * */ -#ifndef SYCL_BLAS_BENCHMARK_IDENTIFIER_HPP -#define SYCL_BLAS_BENCHMARK_IDENTIFIER_HPP +#ifndef PORTBLAS_BENCHMARK_IDENTIFIER_HPP +#define PORTBLAS_BENCHMARK_IDENTIFIER_HPP namespace blas_benchmark { namespace utils { @@ -179,4 +179,4 @@ std::string get_operator_name() { } // namespace utils } // namespace blas_benchmark -#endif // SYCL_BLAS_BENCHMARK_IDENTIFIER_HPP +#endif // PORTBLAS_BENCHMARK_IDENTIFIER_HPP diff --git a/common/include/common/benchmark_names.hpp b/common/include/common/benchmark_names.hpp index 260f84642..0e3d06c01 100644 --- a/common/include/common/benchmark_names.hpp +++ b/common/include/common/benchmark_names.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename benchmark_names.hpp * */ -#ifndef SYCL_BLAS_BENCHMARK_NAMES_HPP -#define SYCL_BLAS_BENCHMARK_NAMES_HPP +#ifndef PORTBLAS_BENCHMARK_NAMES_HPP +#define PORTBLAS_BENCHMARK_NAMES_HPP #include @@ -251,4 +251,4 @@ get_name(index_t rows, index_t cols, std::string reduction_dim, } // namespace utils } // namespace blas_benchmark -#endif // SYCL_BLAS_BENCHMARK_NAMES_HPP +#endif // PORTBLAS_BENCHMARK_NAMES_HPP diff --git a/common/include/common/blas1_state_counters.hpp b/common/include/common/blas1_state_counters.hpp index 57e4d593f..b1ee96e9b 100644 --- a/common/include/common/blas1_state_counters.hpp +++ b/common/include/common/blas1_state_counters.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_state_counters.hpp * diff --git a/common/include/common/blas2_state_counters.hpp b/common/include/common/blas2_state_counters.hpp index 2dc191010..4f67c7850 100644 --- a/common/include/common/blas2_state_counters.hpp +++ b/common/include/common/blas2_state_counters.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_state_counters.hpp * diff --git a/common/include/common/blas3_state_counters.hpp b/common/include/common/blas3_state_counters.hpp index 5840553cf..443430ae6 100644 --- a/common/include/common/blas3_state_counters.hpp +++ b/common/include/common/blas3_state_counters.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_state_counters.hpp * diff --git a/common/include/common/blas_extension_state_counters.hpp b/common/include/common/blas_extension_state_counters.hpp index 8aaae1de0..3fe58134c 100644 --- a/common/include/common/blas_extension_state_counters.hpp +++ b/common/include/common/blas_extension_state_counters.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_extension_state_counters.hpp * diff --git a/common/include/common/set_benchmark_label.hpp b/common/include/common/set_benchmark_label.hpp index eb8c7f6de..b1d4c3ca7 100644 --- a/common/include/common/set_benchmark_label.hpp +++ b/common/include/common/set_benchmark_label.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename set_benchmark_label.hpp * @@ -130,7 +130,7 @@ namespace computecpp_info { * version of compute++ is tied to the version of ComputeCpp, so the associated * meta-data of compute++ will be the same. * - * SYCL-BLAS benchmarks will include these attributes only if ComputeCpp info is + * portBLAS benchmarks will include these attributes only if ComputeCpp info is * available. Benchmarks from other libraries such as cublas etc. will not * include them. * @@ -200,7 +200,7 @@ inline void add_common_labels( computecpp_info::add_computecpp_version(key_value_map); datatype_info::add_datatype_info(key_value_map); - key_value_map["@library"] = "SYCL-BLAS"; + key_value_map["@library"] = "portBLAS"; key_value_map["git_hash"] = commit_hash; key_value_map["git_hash_date"] = commit_date; } @@ -214,7 +214,7 @@ inline void set_benchmark_label(benchmark::State& state, device_info::add_device_info(dev, key_value_map); internal::add_common_labels(key_value_map); - key_value_map["@backend"] = "sycl-blas"; + key_value_map["@backend"] = "portBLAS"; set_label(state, key_value_map); } diff --git a/common/include/common/system_reference_blas.hpp b/common/include/common/system_reference_blas.hpp index d2581072d..afcb4f5e4 100644 --- a/common/include/common/system_reference_blas.hpp +++ b/common/include/common/system_reference_blas.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename system_reference_blas.hpp * diff --git a/conanfile.py b/conanfile.py index 1a9263955..f03457a2c 100644 --- a/conanfile.py +++ b/conanfile.py @@ -3,12 +3,12 @@ import os -class SyclblasConan(ConanFile): - name = "sycl-blas" +class PortBlasConan(ConanFile): + name = "portBLAS" version = "1.0" settings = "os", "compiler", "build_type", "arch" description = "An implementation of BLAS using the SYCL open standard for acceleration on OpenCL devices" - url = "https://github.com/codeplaysoftware/sycl-blas" + url = "https://github.com/codeplaysoftware/portBLAS" license = "Apache-2.0" author = "Codeplay Software Ltd." topics = ('sycl', 'blas') diff --git a/doc/AddingBlas3Op.md b/doc/AddingBlas3Op.md index 11867fc88..a2ded0e0f 100644 --- a/doc/AddingBlas3Op.md +++ b/doc/AddingBlas3Op.md @@ -1,15 +1,15 @@ # Adding a new BLAS level 3 operation This document is meant to be a guide on how to add a new BLAS Level 3 -operation to SYCL-BLAS. It is mostly based on the work done when adding the -TRSM algorithm to SYCL-BLAS. +operation to portBLAS. It is mostly based on the work done when adding the +TRSM algorithm to portBLAS. The steps described here are just a guideline and may vary when doing the actual implementation. ## Add the operation interface -The folder `include/interface` contains the public interface of SYCL-BLAS. +The folder `include/interface` contains the public interface of portBLAS. The files in this folder contain the functions that users can call to run the available blas operations. @@ -18,14 +18,14 @@ function will be located in `include/interface/blas3_interface.h`. When defining a new level 3 operation, the first step is to define the user-facing function, in this case `blas::_trsm`, and declare the -internal function that will be implemented inside SYCL-BLAS, +internal function that will be implemented inside portBLAS, called `blas::internal::_trsm`, similar to the following: ```c++ namespace blas { namespace internal { -// Internal function that will be implemented in the sycl-blas library +// Internal function that will be implemented in the portBLAS library template typename sb_handle_t::event_t _trsm(sb_handle_t& sb_handle, char Side, @@ -169,7 +169,7 @@ typename sb_handle_t::event_t _trsm( ## Instantiate the new operation The next step is to define the `_trsm` function in a `.cpp` file. In this way, when the client -binary is linked against SYCL-BLAS, the linker will find the definition of the missing symbol. +binary is linked against portBLAS, the linker will find the definition of the missing symbol. To do this, we create the source file that will contain instantiations of the new `_trsm` operation. The file is located at `src/interface/blas3/trsm.cpp.in`. This is not the file that will be @@ -181,11 +181,11 @@ compile `blas::internal::_trsm`, for this particular example, this file looks li ```c++ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "operations/blas_constants.hpp" #include "views/view_sycl.hpp" -#include "sycl_blas_helper.h" +#include "portblas_helper.h" #include "interface/blas1_interface.hpp" #include "interface/trsm_interface.hpp" #include "operations/blas3/trsm.hpp" @@ -222,7 +222,7 @@ can be used in the `.cpp` file). After this, the new object file created must be added to the library and is done by adding a new entry to `cmake/CmakeFunctionHelper.cmake`. At the end of this file there is a list of all object -files that are archived to form the SYCL-BLAS library. A new entry must be added in the function +files that are archived to form the portBLAS library. A new entry must be added in the function `build_library`: ```cmake @@ -290,12 +290,12 @@ The last part is to add this file in the include list in `src/operations/blas3/b ## Adding a call to the system blas version of the new operation -SYCL-BLAS implements the netlib-blas interface, so there will be a system blas version of the operation +portBLAS implements the netlib-blas interface, so there will be a system blas version of the operation being added. In the case of `_trsm`, the [netlib-blas version of the function](http://www.netlib.org/lapack/explore-html/d2/d8b/strsm_8f.html) can be called to verify that our implementation produces the correct results for a certain combination of parameters. -SYCL-BLAS provides a utility header that is used to invoke different system-blas functions for testing and benchmarking purposes. +portBLAS provides a utility header that is used to invoke different system-blas functions for testing and benchmarking purposes. To make the new operation available, add it to the file `include/utils/system_reference_blas.hpp`, like the following: ```c++ @@ -322,5 +322,5 @@ So results from the reference implementation can be used to check if the operati ## Conclusion -By following the steps described in this document, you can add a new operation in SYCL-BLAS alongside +By following the steps described in this document, you can add a new operation in portBLAS alongside the tests required to validate the implementation. diff --git a/doc/Autotuner.md b/doc/Autotuner.md index 1571ab461..1d50e65ea 100644 --- a/doc/Autotuner.md +++ b/doc/Autotuner.md @@ -1,4 +1,4 @@ -# SYCL-BLAS Autotuner Developer Documentation +# portBLAS Autotuner Developer Documentation This documentation aims to cover the inner workings of the autotuner from a developer perspective. It should be useful if you need to work on or extend the autotuner in the future, or just want to better understand how it works. @@ -26,7 +26,7 @@ The fifth binary (`tune_all`) runs all four combinations one after the other. The bulk of the work is done inside `gemm_tuner.hpp` in the `run_tune_gemm()` function. This function is called by each `tune_` binary with different combinations of template parameters for transposition. -Before testing the generated combinations the application first benchmarks the provided `system BLAS`, along with the current `SYCL-BLAS`. +Before testing the generated combinations the application first benchmarks the provided `system BLAS`, along with the current `portBLAS`. The generated combinations are then run with the following macro: ```c++ #define BENCH_PARAMS(MEM, ALG, BATCH, VEC, ...) \ @@ -45,9 +45,9 @@ The file `generated_combinations.def` contains calls to `BENCH_PARAMS` for each and the stored results are sorted from worst to best before being printed. For more information on this file and how it is generated, see [this section](#generated_combinations.def). -There are several wrapper functions which wrap the function `run_tune()`: `tune()` and `tune_syclblas()`. +There are several wrapper functions which wrap the function `run_tune()`: `tune()` and `tune_portblas()`. The `run_tune()` function takes a function object, runs it and measures the execution time. -The `tune()` wrapper function is used in the above macro, and `tune_syclblas()` is used for testing the current SYCL-BLAS. +The `tune()` wrapper function is used in the above macro, and `tune_portblas()` is used for testing the current portBLAS. Calling `run_tune()` happens directly in `run_tune_gemm()` for testing the system BLAS. diff --git a/doc/Doxyfile b/doc/Doxyfile index a8441fb71..0eb27ec19 100644 --- a/doc/Doxyfile +++ b/doc/Doxyfile @@ -32,7 +32,7 @@ DOXYFILE_ENCODING = UTF-8 # title of most generated pages and in a few other places. # The default value is: My Project. -PROJECT_NAME = "SYCL-BLAS" +PROJECT_NAME = "portBLAS" # The PROJECT_NUMBER tag can be used to enter a project or revision number. This # could be handy for archiving the generated documentation or if some version diff --git a/doc/Gemm.md b/doc/Gemm.md index 73c52c9b5..26f9cfdf9 100644 --- a/doc/Gemm.md +++ b/doc/Gemm.md @@ -1,6 +1,6 @@ -# SYCL-BLAS GEMM Developer Documentation +# portBLAS GEMM Developer Documentation -The following is documentation for the `GEMM` kernels and associated areas of code within `SYCL-BLAS`. +The following is documentation for the `GEMM` kernels and associated areas of code within `portBLAS`. This should give you a good understanding of how everything works and where/how to do things such as: - Work on or create a new `GEMM` kernel @@ -11,11 +11,11 @@ Please note that while this document primarily refers to `GEMM` and `Blas3` oper # Contents -- [SYCL-BLAS GEMM Developer Documentation](#sycl-blas-gemm-developer-documentation) +- [portBLAS GEMM Developer Documentation](#portBLAS-gemm-developer-documentation) - [Contents](#contents) - [GEMM](#gemm) - [What is GEMM?](#what-is-gemm) - - [SYCL-BLAS GEMM Kernels](#sycl-blas-gemm-kernels) + - [portBLAS GEMM Kernels](#portBLAS-gemm-kernels) - [Relevant CMake Variables](#relevant-cmake-variables) - [Kernel Structure](#kernel-structure) - [Vectorized Loading/Storing](#vectorized-loadingstoring) @@ -42,9 +42,9 @@ C = alpha * A * B + beta * C where A, B and C are matrices and alpha and beta are scalars. ``` -## SYCL-BLAS GEMM Kernels +## portBLAS GEMM Kernels -`SYCL-BLAS` currently contains the following `GEMM` kernels in : +`portBLAS` currently contains the following `GEMM` kernels in : - `gemm_ref.hpp` - A naive, reference implementation of `GEMM` with no optimizations. @@ -59,7 +59,7 @@ where A, B and C are matrices and alpha and beta are scalars. - `gemm_no_local_full_vec.hpp` - Doesn't use local memory. Supports full vectorization. - `gemm_interleaved.hpp` - An alternative approach to batched `GEMM` calculations where the inputs are interleaved in contiguous memory. This means that the batch axis is the fastest moving dimension. -Uses no local memory and corresponds to HWN data layout (NWH in column major, which is what `SYCL-BLAS` uses). Also, the interleaved batched gemm is not subject to custom striding as it beats its initial purpose. +Uses no local memory and corresponds to HWN data layout (NWH in column major, which is what `portBLAS` uses). Also, the interleaved batched gemm is not subject to custom striding as it beats its initial purpose. ## Relevant CMake Variables @@ -235,7 +235,7 @@ Gemm_Launcher #include @@ -135,16 +135,16 @@ static inline index_t get_power_of_two(index_t wGSize, bool rounUp) { } #ifdef __SYCL_DEVICE_ONLY__ -#define SYCL_BLAS_ALWAYS_INLINE \ +#define PORTBLAS_ALWAYS_INLINE \ __attribute__((flatten)) __attribute__((always_inline)) #else -#define SYCL_BLAS_ALWAYS_INLINE +#define PORTBLAS_ALWAYS_INLINE #endif -#define SYCL_BLAS_INLINE SYCL_BLAS_ALWAYS_INLINE inline +#define PORTBLAS_INLINE PORTBLAS_ALWAYS_INLINE inline template -static SYCL_BLAS_INLINE index_t roundUp(index_t x, index_t y) { +static PORTBLAS_INLINE index_t roundUp(index_t x, index_t y) { return ((x + y - 1) / y) * y; } diff --git a/include/container/sycl_iterator.h b/include/container/sycl_iterator.h index 7e75f0238..aa461e388 100644 --- a/include/container/sycl_iterator.h +++ b/include/container/sycl_iterator.h @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sycl_iterator.h * **************************************************************************/ -#ifndef SYCL_BLAS_BUFFER_ITERATOR_H -#define SYCL_BLAS_BUFFER_ITERATOR_H +#ifndef PORTBLAS_BUFFER_ITERATOR_H +#define PORTBLAS_BUFFER_ITERATOR_H #include "blas_meta.h" #include namespace blas { @@ -113,7 +113,7 @@ class BufferIterator { /*! * @brief Default construct a BufferIterator. * This can be used to provide a placeholder BufferIterator, but it is a user - * error if passed into any of the SYCL-BLAS functions. + * error if passed into any of the portBLAS functions. * * Should be removed once SYCL specifies that buffers are default * constructible. See: @@ -377,4 +377,4 @@ struct RebindType> { } // end namespace blas -#endif // SYCL_BLAS_BUFFER_ITERATOR_H +#endif // PORTBLAS_BUFFER_ITERATOR_H diff --git a/include/interface/blas1_interface.h b/include/interface/blas1_interface.h index ead0e92f8..7f7a888ed 100644 --- a/include/interface/blas1_interface.h +++ b/include/interface/blas1_interface.h @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_interface.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS1_INTERFACE_H -#define SYCL_BLAS_BLAS1_INTERFACE_H +#ifndef PORTBLAS_BLAS1_INTERFACE_H +#define PORTBLAS_BLAS1_INTERFACE_H #include "blas_meta.h" namespace blas { @@ -851,4 +851,4 @@ typename ValueType::type _nrm2(sb_handle_t &sb_handle, index_t _N, } } // end namespace blas -#endif // SYCL_BLAS_BLAS1_INTERFACE +#endif // PORTBLAS_BLAS1_INTERFACE diff --git a/include/interface/blas2_interface.h b/include/interface/blas2_interface.h index 6ca9b3a2f..072b9fb50 100644 --- a/include/interface/blas2_interface.h +++ b/include/interface/blas2_interface.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_interface.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS2_INTERFACE_H -#define SYCL_BLAS_BLAS2_INTERFACE_H +#ifndef PORTBLAS_BLAS2_INTERFACE_H +#define PORTBLAS_BLAS2_INTERFACE_H #include "operations/blas2_trees.h" namespace blas { @@ -1136,4 +1136,4 @@ typename sb_handle_t::event_t _tpsv(sb_handle_t& sb_handle, char _Uplo, } } // namespace blas -#endif // SYCL_BLAS_BLAS2_INTERFACE +#endif // PORTBLAS_BLAS2_INTERFACE diff --git a/include/interface/blas3_interface.h b/include/interface/blas3_interface.h index 2f0151b0f..99e241c77 100644 --- a/include/interface/blas3_interface.h +++ b/include/interface/blas3_interface.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_interface.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_INTERFACE_H -#define SYCL_BLAS_BLAS3_INTERFACE_H +#ifndef PORTBLAS_BLAS3_INTERFACE_H +#define PORTBLAS_BLAS3_INTERFACE_H #include "operations/blas3_trees.h" @@ -148,4 +148,4 @@ typename sb_handle_t::event_t _symm(sb_handle_t& sb_handle, char _side, } } // namespace blas -#endif // SYCL_BLAS_BLAS3_INTERFACE +#endif // PORTBLAS_BLAS3_INTERFACE diff --git a/include/interface/extension_interface.h b/include/interface/extension_interface.h index 32fc7930f..566f7cc79 100644 --- a/include/interface/extension_interface.h +++ b/include/interface/extension_interface.h @@ -17,18 +17,18 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction_interface.h * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_INTERFACE_H -#define SYCL_BLAS_EXTENSION_INTERFACE_H +#ifndef PORTBLAS_EXTENSION_INTERFACE_H +#define PORTBLAS_EXTENSION_INTERFACE_H #include "operations/extension/reduction.h" #include "operations/extension/transpose.h" -#include "sb_handle/sycl_blas_handle.h" +#include "sb_handle/portblas_handle.h" namespace blas { @@ -278,4 +278,4 @@ typename sb_handle_t::event_t _reduction(sb_handle_t& sb_handle, } // namespace blas -#endif // SYCL_BLAS_EXTENSION_INTERFACE_H +#endif // PORTBLAS_EXTENSION_INTERFACE_H diff --git a/include/interface/gemm_launcher.h b/include/interface/gemm_launcher.h index 01dedcde5..ac2239fac 100644 --- a/include/interface/gemm_launcher.h +++ b/include/interface/gemm_launcher.h @@ -17,17 +17,17 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_launcher.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_GEMM_LAUNCHER_H -#define SYCL_BLAS_BLAS3_GEMM_LAUNCHER_H +#ifndef PORTBLAS_BLAS3_GEMM_LAUNCHER_H +#define PORTBLAS_BLAS3_GEMM_LAUNCHER_H #include "operations/blas3_trees.h" -#include "sb_handle/sycl_blas_handle.h" +#include "sb_handle/portblas_handle.h" namespace blas { @@ -53,4 +53,4 @@ struct Gemm_Launcher { } // namespace blas -#endif // SYCL_BLAS_BLAS3_GEMM_LAUNCHER_H +#endif // PORTBLAS_BLAS3_GEMM_LAUNCHER_H diff --git a/include/operations/blas1_trees.h b/include/operations/blas1_trees.h index f3de5f896..b519a62d0 100644 --- a/include/operations/blas1_trees.h +++ b/include/operations/blas1_trees.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_trees.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS1_TREES_H -#define SYCL_BLAS_BLAS1_TREES_H +#ifndef PORTBLAS_BLAS1_TREES_H +#define PORTBLAS_BLAS1_TREES_H #include "operations/blas_constants.h" #include "operations/blas_operators.h" #include diff --git a/include/operations/blas2_trees.h b/include/operations/blas2_trees.h index 3f5bc1b3a..5a324dd23 100644 --- a/include/operations/blas2_trees.h +++ b/include/operations/blas2_trees.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_trees.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS2_TREES_H -#define SYCL_BLAS_BLAS2_TREES_H +#ifndef PORTBLAS_BLAS2_TREES_H +#define PORTBLAS_BLAS2_TREES_H namespace blas { /*! @@ -614,7 +614,7 @@ struct Spr { void adjust_access_displacement(); // Row-Col index calculation for Upper Packed Matrix template - SYCL_BLAS_ALWAYS_INLINE static typename std::enable_if::type + PORTBLAS_ALWAYS_INLINE static typename std::enable_if::type compute_row_col(const int64_t id, const index_t size, index_t &row, index_t &col) { int64_t internal = 1 + 8 * id; @@ -638,7 +638,7 @@ struct Spr { // Row-Col index calculation for Lower Packed Matrix template - SYCL_BLAS_ALWAYS_INLINE static typename std::enable_if::type + PORTBLAS_ALWAYS_INLINE static typename std::enable_if::type compute_row_col(const int64_t id, const index_t size, index_t &row, index_t &col) { index_t temp = 2 * size + 1; diff --git a/include/operations/blas3_trees.h b/include/operations/blas3_trees.h index 3f151f24e..854c992db 100644 --- a/include/operations/blas3_trees.h +++ b/include/operations/blas3_trees.h @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_trees.h * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_TREES_H -#define SYCL_BLAS_BLAS3_TREES_H +#ifndef PORTBLAS_BLAS3_TREES_H +#define PORTBLAS_BLAS3_TREES_H #include @@ -334,4 +334,4 @@ make_diag_blocks_inverter(matrix_t& A, matrix_t& invA) { } // namespace blas -#endif // SYCL_BLAS_BLAS3_TREES_H +#endif // PORTBLAS_BLAS3_TREES_H diff --git a/include/operations/blas_constants.h b/include/operations/blas_constants.h index 68c02fb6d..103c78152 100644 --- a/include/operations/blas_constants.h +++ b/include/operations/blas_constants.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_constants.h * **************************************************************************/ -#ifndef SYCL_BLAS_CONSTANTS_H -#define SYCL_BLAS_CONSTANTS_H +#ifndef PORTBLAS_CONSTANTS_H +#define PORTBLAS_CONSTANTS_H #include "blas_meta.h" @@ -46,13 +46,13 @@ template struct GetTupleValue { using return_t = value_t; - SYCL_BLAS_INLINE static return_t get(const value_t val) { return val; } + PORTBLAS_INLINE static return_t get(const value_t val) { return val; } }; template struct GetTupleValue> { using return_t = value_t; - SYCL_BLAS_INLINE static return_t get( + PORTBLAS_INLINE static return_t get( const IndexValueTuple val) { return val.get_value(); } @@ -76,8 +76,8 @@ struct IndexValueTuple { constexpr explicit IndexValueTuple(index_t _ind, value_t _val) : ind(_ind), val(_val){}; - SYCL_BLAS_INLINE index_t get_index() const { return ind; } - SYCL_BLAS_INLINE typename GetTupleValue::return_t get_value() const { + PORTBLAS_INLINE index_t get_index() const { return ind; } + PORTBLAS_INLINE typename GetTupleValue::return_t get_value() const { return GetTupleValue::get(val); } // This operator is required due to a ComputeCPP bug @@ -116,42 +116,42 @@ variable of the type value_t initialized to the specified constant. */ template struct constant { - constexpr static SYCL_BLAS_INLINE value_t value() { + constexpr static PORTBLAS_INLINE value_t value() { return static_cast(Indicator); } }; template struct constant { - constexpr static SYCL_BLAS_INLINE value_t value() { + constexpr static PORTBLAS_INLINE value_t value() { return std::numeric_limits::max(); } }; template struct constant { - constexpr static SYCL_BLAS_INLINE value_t value() { + constexpr static PORTBLAS_INLINE value_t value() { return std::numeric_limits::lowest(); } }; template struct constant { - constexpr static SYCL_BLAS_INLINE value_t value() { + constexpr static PORTBLAS_INLINE value_t value() { return std::numeric_limits::max(); } }; template struct constant { - constexpr static SYCL_BLAS_INLINE value_t value() { + constexpr static PORTBLAS_INLINE value_t value() { return static_cast(0); } }; template struct constant, const_val::abs_max> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), std::numeric_limits::max()); @@ -160,7 +160,7 @@ struct constant, const_val::abs_max> { template struct constant, const_val::abs_min> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), 0); } @@ -168,7 +168,7 @@ struct constant, const_val::abs_min> { template struct constant, const_val::max> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), std::numeric_limits::max()); @@ -177,7 +177,7 @@ struct constant, const_val::max> { template struct constant, const_val::min> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), std::numeric_limits::min()); @@ -186,7 +186,7 @@ struct constant, const_val::min> { template struct constant, Indicator> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), constant::value()); @@ -195,7 +195,7 @@ struct constant, Indicator> { template struct constant, const_val::collapse> { - constexpr static SYCL_BLAS_INLINE IndexValueTuple value() { + constexpr static PORTBLAS_INLINE IndexValueTuple value() { return IndexValueTuple( std::numeric_limits::max(), std::numeric_limits::max()); @@ -204,7 +204,7 @@ struct constant, const_val::collapse> { template struct constant, Indicator> { - constexpr static SYCL_BLAS_INLINE std::complex value() { + constexpr static PORTBLAS_INLINE std::complex value() { return std::complex(constant::value(), constant::value()); } @@ -254,7 +254,7 @@ struct constant template struct constant_pair { - constexpr static SYCL_BLAS_INLINE iv_type value() { + constexpr static PORTBLAS_INLINE iv_type value() { return iv_type( constant::value(), constant::value()); diff --git a/include/operations/blas_operators.h b/include/operations/blas_operators.h index 974265c1c..a41414ade 100644 --- a/include/operations/blas_operators.h +++ b/include/operations/blas_operators.h @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_operators.h * @@ -25,8 +25,8 @@ // NO H for this one as this one is internal. but all the macro will be // generated by cmake in cpp file -#ifndef SYCL_BLAS_OPERATORS_H -#define SYCL_BLAS_OPERATORS_H +#ifndef PORTBLAS_OPERATORS_H +#define PORTBLAS_OPERATORS_H namespace blas { struct Operators; diff --git a/include/operations/extension/reduction.h b/include/operations/extension/reduction.h index 51ac619a7..cc9563964 100644 --- a/include/operations/extension/reduction.h +++ b/include/operations/extension/reduction.h @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction.h * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_REDUCTION_H -#define SYCL_BLAS_EXTENSION_REDUCTION_H +#ifndef PORTBLAS_EXTENSION_REDUCTION_H +#define PORTBLAS_EXTENSION_REDUCTION_H #include @@ -202,4 +202,4 @@ inline Reduction make_reduction( } // namespace blas -#endif // SYCL_BLAS_EXTENSION_REDUCTION_H +#endif // PORTBLAS_EXTENSION_REDUCTION_H diff --git a/include/operations/extension/transpose.h b/include/operations/extension/transpose.h index 4236052d9..471eb6814 100644 --- a/include/operations/extension/transpose.h +++ b/include/operations/extension/transpose.h @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename transpose.h * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_TRANSPOSE_H -#define SYCL_BLAS_EXTENSION_TRANSPOSE_H +#ifndef PORTBLAS_EXTENSION_TRANSPOSE_H +#define PORTBLAS_EXTENSION_TRANSPOSE_H #include @@ -248,4 +248,4 @@ make_transpose_add(in1_t &A, in2_t &B, out_t &C, element_t &alpha, } // namespace blas -#endif // SYCL_BLAS_EXTENSION_TRANSPOSE_H +#endif // PORTBLAS_EXTENSION_TRANSPOSE_H diff --git a/include/sycl_blas.h b/include/portblas.h similarity index 91% rename from include/sycl_blas.h rename to include/portblas.h index 0770e7b97..e289c55ce 100644 --- a/include/sycl_blas.h +++ b/include/portblas.h @@ -17,9 +17,9 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * - * @filename sycl_blas.h + * @filename portblas.h * **************************************************************************/ @@ -29,7 +29,7 @@ #include "container/sycl_iterator.h" -#include "sb_handle/sycl_blas_handle.h" +#include "sb_handle/portblas_handle.h" #include "sb_handle/kernel_constructor.h" @@ -57,6 +57,6 @@ #include "operations/blas_operators.h" -#include "sycl_blas_helper.h" +#include "portblas_helper.h" #include "views/view.h" diff --git a/include/sycl_blas_helper.h b/include/portblas_helper.h similarity index 95% rename from include/sycl_blas_helper.h rename to include/portblas_helper.h index 043ada44b..4683da6e9 100644 --- a/include/sycl_blas_helper.h +++ b/include/portblas_helper.h @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * - * @filename sycl_blas_helper.h + * @filename portblas_helper.h * **************************************************************************/ -#ifndef SYCL_BLAS_HELPER_H -#define SYCL_BLAS_HELPER_H +#ifndef PORTBLAS_HELPER_H +#define PORTBLAS_HELPER_H #include "blas_meta.h" #include "container/sycl_iterator.h" @@ -99,4 +99,4 @@ inline cl::sycl::event fill(cl::sycl::queue q, BufferIterator buff, } } // end namespace helper } // end namespace blas -#endif // SYCL_BLAS_HELPER_H +#endif // PORTBLAS_HELPER_H diff --git a/include/sb_handle/kernel_constructor.h b/include/sb_handle/kernel_constructor.h index 2bdb8c040..fd7481d63 100644 --- a/include/sb_handle/kernel_constructor.h +++ b/include/sb_handle/kernel_constructor.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename kernel_constructor.h * **************************************************************************/ -#ifndef SYCL_BLAS_KERNEL_CONSTRUCTOR_H -#define SYCL_BLAS_KERNEL_CONSTRUCTOR_H +#ifndef PORTBLAS_KERNEL_CONSTRUCTOR_H +#define PORTBLAS_KERNEL_CONSTRUCTOR_H #include @@ -124,4 +124,4 @@ static cl::sycl::event execute_tree(queue_t q, expression_tree_t t, } // namespace blas -#endif // SYCL_BLAS_KERNEL_CONSTRUCTOR_H +#endif // PORTBLAS_KERNEL_CONSTRUCTOR_H diff --git a/include/sb_handle/sycl_blas_handle.h b/include/sb_handle/portblas_handle.h similarity index 96% rename from include/sb_handle/sycl_blas_handle.h rename to include/sb_handle/portblas_handle.h index a8c1e388e..db84498d2 100644 --- a/include/sb_handle/sycl_blas_handle.h +++ b/include/sb_handle/portblas_handle.h @@ -17,20 +17,20 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * - * @filename sycl_blas_handle.h + * @filename portblas_handle.h * **************************************************************************/ -#ifndef SYCL_BLAS_HANDLE_H -#define SYCL_BLAS_HANDLE_H +#ifndef PORTBLAS_HANDLE_H +#define PORTBLAS_HANDLE_H #include "blas_meta.h" #include "operations/blas1_trees.h" #include "operations/blas2_trees.h" #include "operations/blas3_trees.h" #include "operations/extension/reduction.h" -#include "sycl_blas_helper.h" +#include "portblas_helper.h" namespace blas { /** SB_Handle. @@ -147,4 +147,4 @@ class SB_Handle { } // namespace blas -#endif // SYCL_BLAS_HANDLE_H +#endif // PORTBLAS_HANDLE_H diff --git a/include/views/view.h b/include/views/view.h index 129fe7800..1b36cf67b 100644 --- a/include/views/view.h +++ b/include/views/view.h @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename view.h * **************************************************************************/ -#ifndef SYCL_BLAS_VIEW_H -#define SYCL_BLAS_VIEW_H +#ifndef PORTBLAS_VIEW_H +#define PORTBLAS_VIEW_H #include "blas_meta.h" #include diff --git a/python_generator/py_gen_blas_binary.py b/python_generator/py_gen_blas_binary.py index 0fbb383b8..bc7f90df2 100644 --- a/python_generator/py_gen_blas_binary.py +++ b/python_generator/py_gen_blas_binary.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_binary.py # * diff --git a/python_generator/py_gen_blas_binary_special.py b/python_generator/py_gen_blas_binary_special.py index 8d32cf7e9..ae35c19de 100644 --- a/python_generator/py_gen_blas_binary_special.py +++ b/python_generator/py_gen_blas_binary_special.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_binary_special.py # * diff --git a/python_generator/py_gen_blas_gemm_launcher.py b/python_generator/py_gen_blas_gemm_launcher.py index 9bac81c4d..c34ab3849 100644 --- a/python_generator/py_gen_blas_gemm_launcher.py +++ b/python_generator/py_gen_blas_gemm_launcher.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_gemm_launcher.py # * diff --git a/python_generator/py_gen_blas_reduction.py b/python_generator/py_gen_blas_reduction.py index d9f43d23a..27c141fa8 100644 --- a/python_generator/py_gen_blas_reduction.py +++ b/python_generator/py_gen_blas_reduction.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_reduction.py # * diff --git a/python_generator/py_gen_blas_rotg.py b/python_generator/py_gen_blas_rotg.py index d441c7709..fdc178a9e 100644 --- a/python_generator/py_gen_blas_rotg.py +++ b/python_generator/py_gen_blas_rotg.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_rotg.py # * diff --git a/python_generator/py_gen_blas_rotg_return.py b/python_generator/py_gen_blas_rotg_return.py index 7f945c516..e6a28eb23 100644 --- a/python_generator/py_gen_blas_rotg_return.py +++ b/python_generator/py_gen_blas_rotg_return.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_rotg_return.py # * diff --git a/python_generator/py_gen_blas_rotmg.py b/python_generator/py_gen_blas_rotmg.py index 01bfa16dc..f204d7046 100644 --- a/python_generator/py_gen_blas_rotmg.py +++ b/python_generator/py_gen_blas_rotmg.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_rotmg.py # * diff --git a/python_generator/py_gen_blas_ternary.py b/python_generator/py_gen_blas_ternary.py index 7c44e980a..39478fa0a 100644 --- a/python_generator/py_gen_blas_ternary.py +++ b/python_generator/py_gen_blas_ternary.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_ternary.py # * diff --git a/python_generator/py_gen_blas_unary.py b/python_generator/py_gen_blas_unary.py index deca93835..96443a488 100644 --- a/python_generator/py_gen_blas_unary.py +++ b/python_generator/py_gen_blas_unary.py @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename py_gen_blas_unary.py # * diff --git a/run_docker.sh b/run_docker.sh index fb81fa5c8..0c9270a66 100755 --- a/run_docker.sh +++ b/run_docker.sh @@ -1,23 +1,23 @@ #!/bin/bash -# This script runs the SYCL-BLAS tests using the provided Dockerfile. +# This script runs the portBLAS tests using the provided Dockerfile. # The intention is to provide a seamless alternative to .travis.yml, so that # developers can locally test changes in a (somewhat) platform-agnostic manner # without the usual delay that travis testing entails. # -# By default, this script will compile the SYCL-BLAS with g++-7. Other compilers +# By default, this script will compile the portBLAS with g++-7. Other compilers # can be enabled by changing the `CXX_COMPILER` and `CC_COMPILER` environment # variables, e.g.: # export CXX_COMPILER=clang++-6.0 # export CC_COMPILER=clang-6.0 # Targets and git "slug" are also equally configurable. By default, the target -# is OpenCL, and the git repository cloned is codeplay's sycl blas master. +# is OpenCL, and the git repository cloned is codeplay's portBLAS master. export IMPL=COMPUTECPP export CXX_COMPILER=g++-7 export CC_COMPILER=gcc-7 export TARGET=opencl -export GIT_SLUG="codeplaysoftware/sycl-blas" +export GIT_SLUG="codeplaysoftware/portBLAS" export GIT_BRANCH="master" @@ -27,6 +27,6 @@ docker build --build-arg c_compiler=${CC_COMPILER} \ --build-arg git_slug=${GIT_SLUG} \ --build-arg impl=${IMPL} \ --build-arg target=${TARGET} \ - -t sycl-blas . + -t portBLAS . -docker run sycl-blas +docker run portBLAS diff --git a/samples/CMakeLists.txt b/samples/CMakeLists.txt index 4e3849e1f..61ddd8727 100644 --- a/samples/CMakeLists.txt +++ b/samples/CMakeLists.txt @@ -1,12 +1,12 @@ -set(SyclBLAS_INCLUDE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../include) -set(SyclBLAS_SRC_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../src) +set(PORTBLAS_INCLUDE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../include) +set(PORTBLAS_SRC_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../src) list(APPEND CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}) list(APPEND CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/../cmake/Modules) list(APPEND CMAKE_PREFIX_PATH ${CMAKE_CURRENT_SOURCE_DIR}/..) -include(ConfigureSYCLBLAS) +include(ConfigurePORTBLAS) include(SYCL) -find_package(SyclBLAS REQUIRED) +find_package(PORTBLAS REQUIRED) set(SAMPLES_LIST gemv.cpp @@ -30,10 +30,10 @@ foreach(src_file ${SAMPLES_LIST}) TARGET ${sample_exec} SOURCES ${src_file} ) - target_link_libraries(${sample_exec} PRIVATE SyclBLAS::SyclBLAS) + target_link_libraries(${sample_exec} PRIVATE PORTBLAS::PORTBLAS) if(IMGDNN_DIR) target_link_libraries(${sample_exec} PRIVATE IMGDNN::IMGDNN) endif() install(TARGETS ${sample_exec} RUNTIME DESTINATION bin) endforeach() -include_directories(${SYCLBLAS_INCLUDE} ${SYCL_INCLUDE_DIRS} ${THIRD_PARTIES_INCLUDE}) +include_directories(${PORTBLAS_INCLUDE} ${SYCL_INCLUDE_DIRS} ${THIRD_PARTIES_INCLUDE}) diff --git a/samples/FindPORTBLAS.cmake b/samples/FindPORTBLAS.cmake new file mode 100644 index 000000000..161cba8b9 --- /dev/null +++ b/samples/FindPORTBLAS.cmake @@ -0,0 +1,49 @@ +# Try to find the PORTBLAS library. +# +# If the library is found then the `PORTBLAS::PORTBLAS` target will be exported +# with the required include directories. +# +# Sets the following variables: +# PORTBLAS_FOUND - whether the system has PORTBLAS +# PORTBLAS_INCLUDE_DIRS - the PORTBLAS include directory + +find_path(PORTBLAS_INCLUDE_DIR + NAMES portblas.h + PATH_SUFFIXES include + HINTS ${PORTBLAS_DIR} + DOC "The PORTBLAS include directory" +) + +find_path(PORTBLAS_SRC_DIR + NAMES portblas.hpp + PATH_SUFFIXES src + HINTS ${PORTBLAS_DIR} + DOC "The PORTBLAS source directory" +) + + +include(FindPackageHandleStandardArgs) +find_package_handle_standard_args(PORTBLAS + FOUND_VAR PORTBLAS_FOUND + REQUIRED_VARS PORTBLAS_INCLUDE_DIR + PORTBLAS_SRC_DIR +) + +mark_as_advanced(PORTBLAS_FOUND + PORTBLAS_SRC_DIR + PORTBLAS_INCLUDE_DIR +) + +if(PORTBLAS_FOUND) + set(PORTBLAS_INCLUDE_DIRS + ${PORTBLAS_INCLUDE_DIR} + ${PORTBLAS_SRC_DIR} + ) +endif() + +if(PORTBLAS_FOUND AND NOT TARGET PORTBLAS::PORTBLAS) + add_library(PORTBLAS::PORTBLAS INTERFACE IMPORTED) + set_target_properties(PORTBLAS::PORTBLAS PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${PORTBLAS_INCLUDE_DIRS}" + ) +endif() diff --git a/samples/FindSyclBLAS.cmake b/samples/FindSyclBLAS.cmake deleted file mode 100644 index df07179a9..000000000 --- a/samples/FindSyclBLAS.cmake +++ /dev/null @@ -1,49 +0,0 @@ -# Try to find the SyclBLAS library. -# -# If the library is found then the `SyclBLAS::SyclBLAS` target will be exported -# with the required include directories. -# -# Sets the following variables: -# SyclBLAS_FOUND - whether the system has SyclBLAS -# SyclBLAS_INCLUDE_DIRS - the SyclBLAS include directory - -find_path(SyclBLAS_INCLUDE_DIR - NAMES sycl_blas.h - PATH_SUFFIXES include - HINTS ${SyclBLAS_DIR} - DOC "The SyclBLAS include directory" -) - -find_path(SyclBLAS_SRC_DIR - NAMES sycl_blas.hpp - PATH_SUFFIXES src - HINTS ${SyclBLAS_DIR} - DOC "The SyclBLAS source directory" -) - - -include(FindPackageHandleStandardArgs) -find_package_handle_standard_args(SyclBLAS - FOUND_VAR SyclBLAS_FOUND - REQUIRED_VARS SyclBLAS_INCLUDE_DIR - SyclBLAS_SRC_DIR -) - -mark_as_advanced(SyclBLAS_FOUND - SyclBLAS_SRC_DIR - SyclBLAS_INCLUDE_DIR -) - -if(SyclBLAS_FOUND) - set(SyclBLAS_INCLUDE_DIRS - ${SyclBLAS_INCLUDE_DIR} - ${SyclBLAS_SRC_DIR} - ) -endif() - -if(SyclBLAS_FOUND AND NOT TARGET SyclBLAS::SyclBLAS) - add_library(SyclBLAS::SyclBLAS INTERFACE IMPORTED) - set_target_properties(SyclBLAS::SyclBLAS PROPERTIES - INTERFACE_INCLUDE_DIRECTORIES "${SyclBLAS_INCLUDE_DIRS}" - ) -endif() diff --git a/samples/README.md b/samples/README.md index 6ec098f0f..ec5717f84 100644 --- a/samples/README.md +++ b/samples/README.md @@ -1,31 +1,31 @@ -SYCL-BLAS samples +portBLAS samples === ## How to compile the samples -At the moment any project using SYCL-BLAS requires: +At the moment any project using portBLAS requires: * OpenCL * [ComputeCpp](http://www.computecpp.com) -* SYCL-BLAS, either: - * as a library (install the library and include `sycl_blas.h` in an application) - * as a header-only framework (include `sycl_blas.hpp` in an application) +* portBLAS, either: + * as a library (install the library and include `portblas.h` in an application) + * as a header-only framework (include `portblas.hpp` in an application) ### With CMake This folder contains a basic CMake configuration file and a module to find -SYCL-BLAS (which will be used as a header-only framework). It also uses a module +portBLAS (which will be used as a header-only framework). It also uses a module to find ComputeCpp that is located in the folder `cmake/Modules`. Usage: * set `ComputeCpp_DIR` to your ComputeCpp root path -* set `SyclBLAS_DIR` to your SYCL-BLAS root path +* set `PORTBLAS_DIR` to your portBLAS root path ```bash mkdir build cd build cmake .. -GNinja -DComputeCpp_DIR=/path/to/computecpp \ - -DSyclBLAS_DIR=~/path/to/syclblas + -DPORTBLAS_DIR=~/path/to/portblas ninja ``` diff --git a/samples/gemm.cpp b/samples/gemm.cpp index 5c11a5490..c39b9462b 100644 --- a/samples/gemm.cpp +++ b/samples/gemm.cpp @@ -1,4 +1,4 @@ -#include "sycl_blas.hpp" +#include "portblas.hpp" #include #include "util.hpp" @@ -7,12 +7,12 @@ int main(int argc, char** argv) { /* Create a SYCL queue with the default device selector */ cl::sycl::queue q = cl::sycl::queue(cl::sycl::default_selector()); - /* Create a SYCL-BLAS sb_handle and get the policy handler */ + /* Create a portBLAS sb_handle and get the policy handler */ blas::SB_Handle sb_handle(q); /* Arguments of the Gemm operation. * Note: these matrix dimensions are too small to get a performance gain by - * using SYCL-BLAS, but they are convenient for this sample */ + * using portBLAS, but they are convenient for this sample */ const size_t m = 7; const size_t k = 9; const size_t n = 5; diff --git a/samples/gemv.cpp b/samples/gemv.cpp index d4667e2c3..634176aa5 100644 --- a/samples/gemv.cpp +++ b/samples/gemv.cpp @@ -1,4 +1,4 @@ -#include "sycl_blas.hpp" +#include "portblas.hpp" #include #include "util.hpp" @@ -7,12 +7,12 @@ int main(int argc, char** argv) { /* Create a SYCL queue with the default device selector */ cl::sycl::queue q = cl::sycl::queue(cl::sycl::default_selector()); - /* Create a SYCL-BLAS sb_handle and get the policy handler */ + /* Create a portBLAS sb_handle and get the policy handler */ blas::SB_Handle sb_handle(q); /* Arguments of the Gemm operation. * Note: these matrix dimensions are too small to get a performance gain by - * using SYCL-BLAS, but they are convenient for this sample */ + * using portBLAS, but they are convenient for this sample */ const size_t m = 7; const size_t n = 7; const size_t lda = 12; diff --git a/samples/symm.cpp b/samples/symm.cpp index 7b68dc1c0..1bb4a9a53 100644 --- a/samples/symm.cpp +++ b/samples/symm.cpp @@ -1,6 +1,6 @@ -#include "sycl_blas.hpp" +#include "portblas.hpp" #include #include "util.hpp" @@ -9,12 +9,12 @@ int main(int argc, char** argv) { /* Create a SYCL queue with the default device selector */ cl::sycl::queue q = cl::sycl::queue(cl::sycl::default_selector()); - /* Create a SYCL-BLAS sb_handle and get the policy handler */ + /* Create a portBLAS sb_handle and get the policy handler */ blas::SB_Handle sb_handle(q); /* Arguments of the SYMM operation. * Note: these matrix dimensions are too small to get a performance gain by - * using SYCL-BLAS, but they are convenient for this sample */ + * using portBLAS, but they are convenient for this sample */ const size_t m = 5; const size_t n = 7; const size_t lda = 32; diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index edda45e44..7f31c3563 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * diff --git a/src/container/sycl_iterator.hpp b/src/container/sycl_iterator.hpp index d9c1d89cb..0cdbcbb57 100644 --- a/src/container/sycl_iterator.hpp +++ b/src/container/sycl_iterator.hpp @@ -17,12 +17,12 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sycl_iterator.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BUFFER_ITERATOR_HPP -#define SYCL_BLAS_BUFFER_ITERATOR_HPP +#ifndef PORTBLAS_BUFFER_ITERATOR_HPP +#define PORTBLAS_BUFFER_ITERATOR_HPP #include "container/sycl_iterator.h" #endif // BLAS_SYCL_ITERATOR_HPP diff --git a/src/interface/CMakeLists.txt b/src/interface/CMakeLists.txt index cd50dd9e0..43831202a 100644 --- a/src/interface/CMakeLists.txt +++ b/src/interface/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * diff --git a/src/interface/blas1/CMakeLists.txt b/src/interface/blas1/CMakeLists.txt index 1ed769075..abce86626 100644 --- a/src/interface/blas1/CMakeLists.txt +++ b/src/interface/blas1/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeList.txt # * diff --git a/src/interface/blas1/asum.cpp.in b/src/interface/blas1/asum.cpp.in index 304ac3f09..a9c3ff0b5 100644 --- a/src/interface/blas1/asum.cpp.in +++ b/src/interface/blas1/asum.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename _asum.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/asum_return.cpp.in b/src/interface/blas1/asum_return.cpp.in index d1f182d7e..740ed0db1 100644 --- a/src/interface/blas1/asum_return.cpp.in +++ b/src/interface/blas1/asum_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename _asum_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/axpy.cpp.in b/src/interface/blas1/axpy.cpp.in index 9d2b64a9f..5d24b1e84 100644 --- a/src/interface/blas1/axpy.cpp.in +++ b/src/interface/blas1/axpy.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename _axpy.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/copy.cpp.in b/src/interface/blas1/copy.cpp.in index bff01e630..474ee582e 100644 --- a/src/interface/blas1/copy.cpp.in +++ b/src/interface/blas1/copy.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename _copy.cpp * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/dot.cpp.in b/src/interface/blas1/dot.cpp.in index f3a4539b6..5aa560b00 100644 --- a/src/interface/blas1/dot.cpp.in +++ b/src/interface/blas1/dot.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/dot_return.cpp.in b/src/interface/blas1/dot_return.cpp.in index 6f57aae41..ac79665e1 100644 --- a/src/interface/blas1/dot_return.cpp.in +++ b/src/interface/blas1/dot_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename dot_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/iamax.cpp.in b/src/interface/blas1/iamax.cpp.in index b8ab27b83..90bbf0d6f 100644 --- a/src/interface/blas1/iamax.cpp.in +++ b/src/interface/blas1/iamax.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/iamax_return.cpp.in b/src/interface/blas1/iamax_return.cpp.in index 9a6e445e6..5b50678f2 100644 --- a/src/interface/blas1/iamax_return.cpp.in +++ b/src/interface/blas1/iamax_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamax_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/iamin.cpp.in b/src/interface/blas1/iamin.cpp.in index e93043958..a7c827bd6 100644 --- a/src/interface/blas1/iamin.cpp.in +++ b/src/interface/blas1/iamin.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/iamin_return.cpp.in b/src/interface/blas1/iamin_return.cpp.in index c442baed4..94bbfaa8b 100644 --- a/src/interface/blas1/iamin_return.cpp.in +++ b/src/interface/blas1/iamin_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename iamin_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/nrm2.cpp.in b/src/interface/blas1/nrm2.cpp.in index 654e78e39..df3787e27 100644 --- a/src/interface/blas1/nrm2.cpp.in +++ b/src/interface/blas1/nrm2.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/nrm2_return.cpp.in b/src/interface/blas1/nrm2_return.cpp.in index ea5f274ca..b9e63bcb3 100644 --- a/src/interface/blas1/nrm2_return.cpp.in +++ b/src/interface/blas1/nrm2_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nrm2_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/rot.cpp.in b/src/interface/blas1/rot.cpp.in index ee2a32701..fb16046bc 100644 --- a/src/interface/blas1/rot.cpp.in +++ b/src/interface/blas1/rot.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rot.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/rotg.cpp.in b/src/interface/blas1/rotg.cpp.in index 20789789c..40f7d0463 100644 --- a/src/interface/blas1/rotg.cpp.in +++ b/src/interface/blas1/rotg.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotg.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/rotg_return.cpp.in b/src/interface/blas1/rotg_return.cpp.in index 9e42fa73d..54a22852b 100644 --- a/src/interface/blas1/rotg_return.cpp.in +++ b/src/interface/blas1/rotg_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotg_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/rotm.cpp.in b/src/interface/blas1/rotm.cpp.in index cde9b39c5..ee36440cb 100644 --- a/src/interface/blas1/rotm.cpp.in +++ b/src/interface/blas1/rotm.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename rotm.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/rotmg.cpp.in b/src/interface/blas1/rotmg.cpp.in index 24a2e0e81..1ef9e3c0e 100644 --- a/src/interface/blas1/rotmg.cpp.in +++ b/src/interface/blas1/rotmg.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * -* SYCL-BLAS: BLAS implementation using SYCL +* portBLAS: BLAS implementation using SYCL * * @filename rotmg.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/scal.cpp.in b/src/interface/blas1/scal.cpp.in index fe1e80958..83f9ed1e9 100644 --- a/src/interface/blas1/scal.cpp.in +++ b/src/interface/blas1/scal.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename scal.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/sdsdot.cpp.in b/src/interface/blas1/sdsdot.cpp.in index deadbb74f..c219c2007 100644 --- a/src/interface/blas1/sdsdot.cpp.in +++ b/src/interface/blas1/sdsdot.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sdsdot.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/sdsdot_return.cpp.in b/src/interface/blas1/sdsdot_return.cpp.in index bdfd60feb..a34b18ab7 100644 --- a/src/interface/blas1/sdsdot_return.cpp.in +++ b/src/interface/blas1/sdsdot_return.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sdsdot_return.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1/swap.cpp.in b/src/interface/blas1/swap.cpp.in index fff79479a..550a43d49 100644 --- a/src/interface/blas1/swap.cpp.in +++ b/src/interface/blas1/swap.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename swap.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas1_interface.hpp b/src/interface/blas1_interface.hpp index 309969e06..f6a01ab1c 100644 --- a/src/interface/blas1_interface.hpp +++ b/src/interface/blas1_interface.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_interface.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS1_INTERFACE_HPP -#define SYCL_BLAS_BLAS1_INTERFACE_HPP +#ifndef PORTBLAS_BLAS1_INTERFACE_HPP +#define PORTBLAS_BLAS1_INTERFACE_HPP #include #include @@ -37,7 +37,7 @@ #include "operations/blas1_trees.h" #include "operations/blas_constants.h" #include "operations/blas_operators.hpp" -#include "sb_handle/sycl_blas_handle.h" +#include "sb_handle/portblas_handle.h" namespace blas { namespace internal { diff --git a/src/interface/blas2/CMakeLists.txt b/src/interface/blas2/CMakeLists.txt index 43a1a2fdb..a3e46d0f4 100644 --- a/src/interface/blas2/CMakeLists.txt +++ b/src/interface/blas2/CMakeLists.txt @@ -18,7 +18,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * diff --git a/src/interface/blas2/backend/amd_gpu.hpp b/src/interface/blas2/backend/amd_gpu.hpp index 156d1a4eb..1c7557e8a 100644 --- a/src/interface/blas2/backend/amd_gpu.hpp +++ b/src/interface/blas2/backend/amd_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename amd_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMV_AMD_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMV_AMD_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMV_AMD_GPU_BACKEND_HPP +#define PORTBLAS_GEMV_AMD_GPU_BACKEND_HPP #include "interface/blas2_interface.h" namespace blas { diff --git a/src/interface/blas2/backend/backend.hpp b/src/interface/blas2/backend/backend.hpp index e48b75b98..2a88c8624 100644 --- a/src/interface/blas2/backend/backend.hpp +++ b/src/interface/blas2/backend/backend.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename backend.hpp * diff --git a/src/interface/blas2/backend/default_cpu.hpp b/src/interface/blas2/backend/default_cpu.hpp index a6de9a435..1972688b0 100644 --- a/src/interface/blas2/backend/default_cpu.hpp +++ b/src/interface/blas2/backend/default_cpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename default_cpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMV_DEFAULT_CPU_BACKEND_HPP -#define SYCL_BLAS_GEMV_DEFAULT_CPU_BACKEND_HPP +#ifndef PORTBLAS_GEMV_DEFAULT_CPU_BACKEND_HPP +#define PORTBLAS_GEMV_DEFAULT_CPU_BACKEND_HPP #include "interface/blas2_interface.h" namespace blas { diff --git a/src/interface/blas2/backend/intel_gpu.hpp b/src/interface/blas2/backend/intel_gpu.hpp index 3571e37cb..4895a6ce6 100644 --- a/src/interface/blas2/backend/intel_gpu.hpp +++ b/src/interface/blas2/backend/intel_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename intel_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMV_INTEL_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMV_INTEL_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMV_INTEL_GPU_BACKEND_HPP +#define PORTBLAS_GEMV_INTEL_GPU_BACKEND_HPP #include "interface/blas2_interface.h" namespace blas { diff --git a/src/interface/blas2/backend/nvidia_gpu.hpp b/src/interface/blas2/backend/nvidia_gpu.hpp index df968fedc..24922008a 100644 --- a/src/interface/blas2/backend/nvidia_gpu.hpp +++ b/src/interface/blas2/backend/nvidia_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nvidia_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMV_NVIDIA_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMV_NVIDIA_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMV_NVIDIA_GPU_BACKEND_HPP +#define PORTBLAS_GEMV_NVIDIA_GPU_BACKEND_HPP #include "interface/blas2_interface.h" namespace blas { diff --git a/src/interface/blas2/backend/power_vr.hpp b/src/interface/blas2/backend/power_vr.hpp index e52c802d7..4c9776299 100644 --- a/src/interface/blas2/backend/power_vr.hpp +++ b/src/interface/blas2/backend/power_vr.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename power_vr.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMV_POWER_VR_BACKEND_HPP -#define SYCL_BLAS_GEMV_POWER_VR_BACKEND_HPP +#ifndef PORTBLAS_GEMV_POWER_VR_BACKEND_HPP +#define PORTBLAS_GEMV_POWER_VR_BACKEND_HPP #include "interface/blas2_interface.h" namespace blas { diff --git a/src/interface/blas2/gbmv.cpp.in b/src/interface/blas2/gbmv.cpp.in index 54455c49e..3b9bf41e9 100644 --- a/src/interface/blas2/gbmv.cpp.in +++ b/src/interface/blas2/gbmv.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/gemv.cpp.in b/src/interface/blas2/gemv.cpp.in index d19fe0a60..5696b8a07 100644 --- a/src/interface/blas2/gemv.cpp.in +++ b/src/interface/blas2/gemv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/ger.cpp.in b/src/interface/blas2/ger.cpp.in index 70b46cd5b..8514cc66b 100644 --- a/src/interface/blas2/ger.cpp.in +++ b/src/interface/blas2/ger.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/sbmv.cpp.in b/src/interface/blas2/sbmv.cpp.in index 82cfe93cc..213f67fb7 100644 --- a/src/interface/blas2/sbmv.cpp.in +++ b/src/interface/blas2/sbmv.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/spmv.cpp.in b/src/interface/blas2/spmv.cpp.in index c47cf594c..2c1137bfe 100644 --- a/src/interface/blas2/spmv.cpp.in +++ b/src/interface/blas2/spmv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spmv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/spr.cpp.in b/src/interface/blas2/spr.cpp.in index bae1d711d..0ddab611d 100644 --- a/src/interface/blas2/spr.cpp.in +++ b/src/interface/blas2/spr.cpp.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.cpp.in * @@ -26,7 +26,7 @@ #include "interface/blas2_interface.hpp" #include "operations/blas2_trees.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "views/view_sycl.hpp" namespace blas { diff --git a/src/interface/blas2/spr2.cpp.in b/src/interface/blas2/spr2.cpp.in index 6167c75dd..de3096247 100644 --- a/src/interface/blas2/spr2.cpp.in +++ b/src/interface/blas2/spr2.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr2.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/symv.cpp.in b/src/interface/blas2/symv.cpp.in index 963632ece..e4ef84634 100644 --- a/src/interface/blas2/symv.cpp.in +++ b/src/interface/blas2/symv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/syr.cpp.in b/src/interface/blas2/syr.cpp.in index dfa828675..5740f074c 100644 --- a/src/interface/blas2/syr.cpp.in +++ b/src/interface/blas2/syr.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/syr2.cpp.in b/src/interface/blas2/syr2.cpp.in index c4f7db49f..8eaed7376 100644 --- a/src/interface/blas2/syr2.cpp.in +++ b/src/interface/blas2/syr2.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename syr2.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/tbmv.cpp.in b/src/interface/blas2/tbmv.cpp.in index 5d97a5bc1..9d40211ef 100644 --- a/src/interface/blas2/tbmv.cpp.in +++ b/src/interface/blas2/tbmv.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" #include "operations/blas1_trees.hpp" diff --git a/src/interface/blas2/tbsv.cpp.in b/src/interface/blas2/tbsv.cpp.in index 2f3eb26bf..e19af0ba1 100644 --- a/src/interface/blas2/tbsv.cpp.in +++ b/src/interface/blas2/tbsv.cpp.in @@ -17,12 +17,12 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbsv.cpp.in * **************************************************************************/ -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" diff --git a/src/interface/blas2/tpmv.cpp.in b/src/interface/blas2/tpmv.cpp.in index e55ad2298..f7e51976b 100644 --- a/src/interface/blas2/tpmv.cpp.in +++ b/src/interface/blas2/tpmv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { template typename SB_Handle::event_t _tpmv( diff --git a/src/interface/blas2/tpsv.cpp.in b/src/interface/blas2/tpsv.cpp.in index 495666ef7..3569072ba 100644 --- a/src/interface/blas2/tpsv.cpp.in +++ b/src/interface/blas2/tpsv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tpsv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/trmv.cpp.in b/src/interface/blas2/trmv.cpp.in index 62cda27fc..252342073 100644 --- a/src/interface/blas2/trmv.cpp.in +++ b/src/interface/blas2/trmv.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trmv.cpp.in * **************************************************************************/ #include "interface/blas2_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" namespace blas { namespace internal { diff --git a/src/interface/blas2/trsv.cpp.in b/src/interface/blas2/trsv.cpp.in index f0d81a333..26161ab67 100644 --- a/src/interface/blas2/trsv.cpp.in +++ b/src/interface/blas2/trsv.cpp.in @@ -17,12 +17,12 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename trsv.cpp.in * **************************************************************************/ -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas2_interface.hpp" diff --git a/src/interface/blas2_interface.hpp b/src/interface/blas2_interface.hpp index f0aefb84a..b23cee47e 100644 --- a/src/interface/blas2_interface.hpp +++ b/src/interface/blas2_interface.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_interface.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS2_INTERFACE_HPP -#define SYCL_BLAS_BLAS2_INTERFACE_HPP +#ifndef PORTBLAS_BLAS2_INTERFACE_HPP +#define PORTBLAS_BLAS2_INTERFACE_HPP #include "blas_meta.h" #include "container/sycl_iterator.h" @@ -33,7 +33,7 @@ #include "operations/blas2_trees.h" #include "operations/blas_constants.h" #include "operations/blas_operators.hpp" -#include "sb_handle/sycl_blas_handle.h" +#include "sb_handle/portblas_handle.h" #include "views/view.h" #include #include diff --git a/src/interface/blas3/CMakeLists.txt b/src/interface/blas3/CMakeLists.txt index 4852becee..144215a03 100644 --- a/src/interface/blas3/CMakeLists.txt +++ b/src/interface/blas3/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * diff --git a/src/interface/blas3/backend/amd_gpu.hpp b/src/interface/blas3/backend/amd_gpu.hpp index b2afcd849..b53bdc35b 100644 --- a/src/interface/blas3/backend/amd_gpu.hpp +++ b/src/interface/blas3/backend/amd_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename amd_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMM_AMD_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMM_AMD_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMM_AMD_GPU_BACKEND_HPP +#define PORTBLAS_GEMM_AMD_GPU_BACKEND_HPP #include "interface/gemm_launcher.h" namespace blas { diff --git a/src/interface/blas3/backend/backend.hpp b/src/interface/blas3/backend/backend.hpp index 88ea9d474..b108af0c4 100644 --- a/src/interface/blas3/backend/backend.hpp +++ b/src/interface/blas3/backend/backend.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename backend.hpp * diff --git a/src/interface/blas3/backend/default_cpu.hpp b/src/interface/blas3/backend/default_cpu.hpp index 6fd5c8ca5..fcf148c69 100644 --- a/src/interface/blas3/backend/default_cpu.hpp +++ b/src/interface/blas3/backend/default_cpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename default_cpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMM_DEFAULT_CPU_BACKEND_HPP -#define SYCL_BLAS_GEMM_DEFAULT_CPU_BACKEND_HPP +#ifndef PORTBLAS_GEMM_DEFAULT_CPU_BACKEND_HPP +#define PORTBLAS_GEMM_DEFAULT_CPU_BACKEND_HPP #include "interface/gemm_launcher.h" namespace blas { diff --git a/src/interface/blas3/backend/intel_gpu.hpp b/src/interface/blas3/backend/intel_gpu.hpp index a42615d4a..c35ba8893 100644 --- a/src/interface/blas3/backend/intel_gpu.hpp +++ b/src/interface/blas3/backend/intel_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename intel_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMM_INTEL_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMM_INTEL_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMM_INTEL_GPU_BACKEND_HPP +#define PORTBLAS_GEMM_INTEL_GPU_BACKEND_HPP #include "interface/gemm_launcher.h" namespace blas { diff --git a/src/interface/blas3/backend/nvidia_gpu.hpp b/src/interface/blas3/backend/nvidia_gpu.hpp index c84f7964f..66c9d8349 100644 --- a/src/interface/blas3/backend/nvidia_gpu.hpp +++ b/src/interface/blas3/backend/nvidia_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename NVIDIA_GPU.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMM_NVIDIA_GPU_BACKEND_HPP -#define SYCL_BLAS_GEMM_NVIDIA_GPU_BACKEND_HPP +#ifndef PORTBLAS_GEMM_NVIDIA_GPU_BACKEND_HPP +#define PORTBLAS_GEMM_NVIDIA_GPU_BACKEND_HPP #include "interface/gemm_launcher.h" namespace blas { diff --git a/src/interface/blas3/backend/power_vr.hpp b/src/interface/blas3/backend/power_vr.hpp index bec1ed5d8..734d30084 100644 --- a/src/interface/blas3/backend/power_vr.hpp +++ b/src/interface/blas3/backend/power_vr.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename power_vr.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_GEMM_POWERVR_BACKEND_HPP -#define SYCL_BLAS_GEMM_POWERVR_BACKEND_HPP +#ifndef PORTBLAS_GEMM_POWERVR_BACKEND_HPP +#define PORTBLAS_GEMM_POWERVR_BACKEND_HPP #include "interface/gemm_launcher.h" #ifdef IMGDNN_LIBRARY diff --git a/src/interface/blas3/gemm.cpp.in b/src/interface/blas3/gemm.cpp.in index b6108a8ef..7eebdb6a8 100644 --- a/src/interface/blas3/gemm.cpp.in +++ b/src/interface/blas3/gemm.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "interface/gemm_interface.hpp" #include "operations/blas_constants.hpp" #include "views/view_sycl.hpp" diff --git a/src/interface/blas3/gemm_launcher.cpp.in b/src/interface/blas3/gemm_launcher.cpp.in index 971f605d4..6f0b99667 100644 --- a/src/interface/blas3/gemm_launcher.cpp.in +++ b/src/interface/blas3/gemm_launcher.cpp.in @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_launcher.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/gemm_launcher.hpp" #include "operations/blas3_trees.hpp" diff --git a/src/interface/blas3/symm.cpp.in b/src/interface/blas3/symm.cpp.in index d3f0fcd98..1ad29ebad 100644 --- a/src/interface/blas3/symm.cpp.in +++ b/src/interface/blas3/symm.cpp.in @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename symm.cpp.in * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "interface/gemm_interface.hpp" #include "interface/symm_interface.hpp" #include "operations/blas_constants.hpp" diff --git a/src/interface/blas3/trsm.cpp.in b/src/interface/blas3/trsm.cpp.in index 0ffeabd39..fd9fab28a 100644 --- a/src/interface/blas3/trsm.cpp.in +++ b/src/interface/blas3/trsm.cpp.in @@ -17,12 +17,12 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * **************************************************************************/ #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "interface/trsm_interface.hpp" diff --git a/src/interface/blas3_interface.hpp b/src/interface/blas3_interface.hpp index 1697fed62..ba7ab9e15 100644 --- a/src/interface/blas3_interface.hpp +++ b/src/interface/blas3_interface.hpp @@ -17,18 +17,18 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_interface.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_INTERFACE_HPP -#define SYCL_BLAS_BLAS3_INTERFACE_HPP +#ifndef PORTBLAS_BLAS3_INTERFACE_HPP +#define PORTBLAS_BLAS3_INTERFACE_HPP #include "interface/gemm_interface.hpp" #include "interface/gemm_launcher.hpp" #include "interface/symm_interface.hpp" #include "interface/trsm_interface.hpp" -#endif // SYCL_BLAS_BLAS3_INTERFACE_HPP +#endif // PORTBLAS_BLAS3_INTERFACE_HPP diff --git a/src/interface/extension/CMakeLists.txt b/src/interface/extension/CMakeLists.txt index 91328e897..65ba0617a 100644 --- a/src/interface/extension/CMakeLists.txt +++ b/src/interface/extension/CMakeLists.txt @@ -17,7 +17,7 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * diff --git a/src/interface/extension/backend/amd_gpu.hpp b/src/interface/extension/backend/amd_gpu.hpp index 51ceba109..13d2036a7 100644 --- a/src/interface/extension/backend/amd_gpu.hpp +++ b/src/interface/extension/backend/amd_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename amd_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_TRANSPOSE_AMD_GPU_BACKEND_HPP -#define SYCL_BLAS_TRANSPOSE_AMD_GPU_BACKEND_HPP +#ifndef PORTBLAS_TRANSPOSE_AMD_GPU_BACKEND_HPP +#define PORTBLAS_TRANSPOSE_AMD_GPU_BACKEND_HPP #include "interface/extension_interface.h" namespace blas { diff --git a/src/interface/extension/backend/backend.hpp b/src/interface/extension/backend/backend.hpp index b29746211..6de002ad0 100644 --- a/src/interface/extension/backend/backend.hpp +++ b/src/interface/extension/backend/backend.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename backend.hpp * diff --git a/src/interface/extension/backend/default_cpu.hpp b/src/interface/extension/backend/default_cpu.hpp index bf717a0dd..0484c4ad0 100644 --- a/src/interface/extension/backend/default_cpu.hpp +++ b/src/interface/extension/backend/default_cpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename default_cpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_TRANSPOSE_DEFAULT_CPU_BACKEND_HPP -#define SYCL_BLAS_TRANSPOSE_DEFAULT_CPU_BACKEND_HPP +#ifndef PORTBLAS_TRANSPOSE_DEFAULT_CPU_BACKEND_HPP +#define PORTBLAS_TRANSPOSE_DEFAULT_CPU_BACKEND_HPP #include "interface/extension_interface.h" namespace blas { diff --git a/src/interface/extension/backend/intel_gpu.hpp b/src/interface/extension/backend/intel_gpu.hpp index cd2734840..0df0c6d86 100644 --- a/src/interface/extension/backend/intel_gpu.hpp +++ b/src/interface/extension/backend/intel_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename intel_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_TRANSPOSE_INTEL_GPU_BACKEND_HPP -#define SYCL_BLAS_TRANSPOSE_INTEL_GPU_BACKEND_HPP +#ifndef PORTBLAS_TRANSPOSE_INTEL_GPU_BACKEND_HPP +#define PORTBLAS_TRANSPOSE_INTEL_GPU_BACKEND_HPP #include "interface/extension_interface.h" namespace blas { diff --git a/src/interface/extension/backend/nvidia_gpu.hpp b/src/interface/extension/backend/nvidia_gpu.hpp index a25206a2b..05de307c8 100644 --- a/src/interface/extension/backend/nvidia_gpu.hpp +++ b/src/interface/extension/backend/nvidia_gpu.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename nvidia_gpu.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_TRANSPOSE_NVIDIA_GPU_BACKEND_HPP -#define SYCL_BLAS_TRANSPOSE_NVIDIA_GPU_BACKEND_HPP +#ifndef PORTBLAS_TRANSPOSE_NVIDIA_GPU_BACKEND_HPP +#define PORTBLAS_TRANSPOSE_NVIDIA_GPU_BACKEND_HPP #include "interface/extension_interface.h" namespace blas { diff --git a/src/interface/extension/matcopy.cpp.in b/src/interface/extension/matcopy.cpp.in index 8f90649f5..eb3ab7b19 100644 --- a/src/interface/extension/matcopy.cpp.in +++ b/src/interface/extension/matcopy.cpp.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename matcopy.cpp.in * @@ -25,7 +25,7 @@ #include "interface/extension_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "operations/extension/transpose.hpp" namespace blas { diff --git a/src/interface/extension/reduction.cpp.in b/src/interface/extension/reduction.cpp.in index 59603fc0a..860066710 100644 --- a/src/interface/extension/reduction.cpp.in +++ b/src/interface/extension/reduction.cpp.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction.cpp.in * @@ -25,7 +25,7 @@ #include "operations/extension/reduction.hpp" #include "container/sycl_iterator.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "sb_handle/kernel_constructor.hpp" #include "interface/blas1_interface.hpp" #include "interface/extension_interface.hpp" diff --git a/src/interface/extension/transpose.cpp.in b/src/interface/extension/transpose.cpp.in index 5a8ac8cc2..875cdd8f7 100644 --- a/src/interface/extension/transpose.cpp.in +++ b/src/interface/extension/transpose.cpp.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename transpose.cpp.in * @@ -25,7 +25,7 @@ #include "interface/extension_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "operations/extension/transpose.hpp" namespace blas { diff --git a/src/interface/extension_interface.hpp b/src/interface/extension_interface.hpp index f05f6716b..d44ecfb71 100644 --- a/src/interface/extension_interface.hpp +++ b/src/interface/extension_interface.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction_interface.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_INTERFACE_HPP -#define SYCL_BLAS_EXTENSION_INTERFACE_HPP +#ifndef PORTBLAS_EXTENSION_INTERFACE_HPP +#define PORTBLAS_EXTENSION_INTERFACE_HPP #include "blas_meta.h" #include "interface/extension/backend/backend.hpp" @@ -33,8 +33,8 @@ #include "operations/blas_operators.hpp" #include "operations/extension/reduction.h" #include "operations/extension/transpose.h" -#include "sb_handle/sycl_blas_handle.h" -#include "sycl_blas_helper.h" +#include "sb_handle/portblas_handle.h" +#include "portblas_helper.h" #include "views/view.h" namespace blas { @@ -405,4 +405,4 @@ typename sb_handle_t::event_t _reduction(sb_handle_t& sb_handle, } // namespace internal } // namespace blas -#endif // SYCL_BLAS_EXTENSION_INTERFACE_HPP +#endif // PORTBLAS_EXTENSION_INTERFACE_HPP diff --git a/src/interface/gemm_interface.hpp b/src/interface/gemm_interface.hpp index 1bd56f3f3..e7bdb79fa 100644 --- a/src/interface/gemm_interface.hpp +++ b/src/interface/gemm_interface.hpp @@ -17,20 +17,20 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_GEMM_INTERFACE_HPP -#define SYCL_BLAS_BLAS3_GEMM_INTERFACE_HPP +#ifndef PORTBLAS_BLAS3_GEMM_INTERFACE_HPP +#define PORTBLAS_BLAS3_GEMM_INTERFACE_HPP #include "blas_meta.h" #include "interface/blas1_interface.h" #include "interface/blas3/backend/backend.hpp" #include "interface/blas3_interface.h" #include "operations/blas3_trees.h" -#include "sb_handle/sycl_blas_handle.h" -#include "sycl_blas_helper.h" +#include "sb_handle/portblas_handle.h" +#include "portblas_helper.h" #include #include @@ -213,4 +213,4 @@ typename sb_handle_t::event_t _gemm_strided_batched( } // namespace internal } // namespace blas -#endif // SYCL_BLAS_BLAS3_GEMM_INTERFACE_HPP +#endif // PORTBLAS_BLAS3_GEMM_INTERFACE_HPP diff --git a/src/interface/gemm_launcher.hpp b/src/interface/gemm_launcher.hpp index 0ef58c077..531a87b7d 100644 --- a/src/interface/gemm_launcher.hpp +++ b/src/interface/gemm_launcher.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_launcher.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_LAUNCHER_HPP -#define SYCL_BLAS_BLAS3_LAUNCHER_HPP +#ifndef PORTBLAS_BLAS3_LAUNCHER_HPP +#define PORTBLAS_BLAS3_LAUNCHER_HPP #include "interface/gemm_launcher.h" #include "views/view.h" @@ -69,4 +69,4 @@ Gemm_Launcher Join::Join(lhs_t &_l, rhs_t _r) : lhs_(_l), rhs_(_r) {} template -SYCL_BLAS_INLINE typename Join::index_t +PORTBLAS_INLINE typename Join::index_t Join::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool Join::valid_thread( +PORTBLAS_INLINE bool Join::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < Join::get_size())); } template -SYCL_BLAS_INLINE typename Join::value_t Join::eval( +PORTBLAS_INLINE typename Join::value_t Join::eval( typename Join::index_t i) { lhs_.eval(i); return rhs_.eval(i); } template -SYCL_BLAS_INLINE typename Join::value_t Join::eval( +PORTBLAS_INLINE typename Join::value_t Join::eval( cl::sycl::nd_item<1> ndItem) { return Join::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void Join::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void Join::bind(cl::sycl::handler &h) { lhs_.bind(h); rhs_.bind(h); } template -SYCL_BLAS_INLINE void Join::adjust_access_displacement() { +PORTBLAS_INLINE void Join::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_.adjust_access_displacement(); } @@ -162,13 +162,13 @@ template Assign::Assign(lhs_t &_l, rhs_t _r) : lhs_(_l), rhs_(_r){}; template -SYCL_BLAS_INLINE typename Assign::index_t +PORTBLAS_INLINE typename Assign::index_t Assign::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool Assign::valid_thread( +PORTBLAS_INLINE bool Assign::valid_thread( cl::sycl::nd_item<1> ndItem) const { using index_t = typename Assign::index_t; return (static_cast(ndItem.get_global_id(0)) < @@ -176,25 +176,25 @@ SYCL_BLAS_INLINE bool Assign::valid_thread( } template -SYCL_BLAS_INLINE typename Assign::value_t +PORTBLAS_INLINE typename Assign::value_t Assign::eval(typename Assign::index_t i) { auto val = lhs_.eval(i) = rhs_.eval(i); return val; } template -SYCL_BLAS_INLINE typename Assign::value_t +PORTBLAS_INLINE typename Assign::value_t Assign::eval(cl::sycl::nd_item<1> ndItem) { return Assign::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void Assign::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void Assign::bind(cl::sycl::handler &h) { lhs_.bind(h); rhs_.bind(h); } template -SYCL_BLAS_INLINE void Assign::adjust_access_displacement() { +PORTBLAS_INLINE void Assign::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_.adjust_access_displacement(); } @@ -203,13 +203,13 @@ SYCL_BLAS_INLINE void Assign::adjust_access_displacement() { */ template -SYCL_BLAS_INLINE DoubleAssign::DoubleAssign( +PORTBLAS_INLINE DoubleAssign::DoubleAssign( lhs_1_t &_l1, lhs_2_t &_l2, rhs_1_t _r1, rhs_2_t _r2) : lhs_1_(_l1), lhs_2_(_l2), rhs_1_(_r1), rhs_2_(_r2){}; template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename DoubleAssign::index_t DoubleAssign::get_size() const { return rhs_2_.get_size(); @@ -217,7 +217,7 @@ SYCL_BLAS_INLINE template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool DoubleAssign::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < get_size())); @@ -225,7 +225,7 @@ DoubleAssign::valid_thread( template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename DoubleAssign::value_t DoubleAssign::eval( typename DoubleAssign::index_t i) { @@ -238,7 +238,7 @@ SYCL_BLAS_INLINE template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename DoubleAssign::value_t DoubleAssign::eval( cl::sycl::nd_item<1> ndItem) { @@ -247,7 +247,7 @@ SYCL_BLAS_INLINE } template -SYCL_BLAS_INLINE void DoubleAssign::bind( +PORTBLAS_INLINE void DoubleAssign::bind( cl::sycl::handler &h) { lhs_1_.bind(h); rhs_1_.bind(h); @@ -257,7 +257,7 @@ SYCL_BLAS_INLINE void DoubleAssign::bind( template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void DoubleAssign::adjust_access_displacement() { lhs_1_.adjust_access_displacement(); rhs_1_.adjust_access_displacement(); @@ -274,35 +274,35 @@ ScalarOp::ScalarOp(scalar_t _scl, rhs_t &_r) : scalar_(_scl), rhs_(_r) {} template -SYCL_BLAS_INLINE typename ScalarOp::index_t +PORTBLAS_INLINE typename ScalarOp::index_t ScalarOp::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool ScalarOp::valid_thread( +PORTBLAS_INLINE bool ScalarOp::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < ScalarOp::get_size())); } template -SYCL_BLAS_INLINE typename ScalarOp::value_t +PORTBLAS_INLINE typename ScalarOp::value_t ScalarOp::eval( typename ScalarOp::index_t i) { return operator_t::eval(internal::get_scalar(scalar_), rhs_.eval(i)); } template -SYCL_BLAS_INLINE typename ScalarOp::value_t +PORTBLAS_INLINE typename ScalarOp::value_t ScalarOp::eval(cl::sycl::nd_item<1> ndItem) { return ScalarOp::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void ScalarOp::bind( +PORTBLAS_INLINE void ScalarOp::bind( cl::sycl::handler &h) { rhs_.bind(h); } template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void ScalarOp::adjust_access_displacement() { rhs_.adjust_access_displacement(); } @@ -313,35 +313,35 @@ template UnaryOp::UnaryOp(rhs_t &_r) : rhs_(_r) {} template -SYCL_BLAS_INLINE typename UnaryOp::index_t +PORTBLAS_INLINE typename UnaryOp::index_t UnaryOp::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool UnaryOp::valid_thread( +PORTBLAS_INLINE bool UnaryOp::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < UnaryOp::get_size())); } template -SYCL_BLAS_INLINE typename UnaryOp::value_t +PORTBLAS_INLINE typename UnaryOp::value_t UnaryOp::eval( typename UnaryOp::index_t i) { return operator_t::eval(rhs_.eval(i)); } template -SYCL_BLAS_INLINE typename UnaryOp::value_t +PORTBLAS_INLINE typename UnaryOp::value_t UnaryOp::eval(cl::sycl::nd_item<1> ndItem) { return UnaryOp::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void UnaryOp::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void UnaryOp::bind(cl::sycl::handler &h) { rhs_.bind(h); } template -SYCL_BLAS_INLINE void UnaryOp::adjust_access_displacement() { +PORTBLAS_INLINE void UnaryOp::adjust_access_displacement() { rhs_.adjust_access_displacement(); } @@ -353,36 +353,36 @@ BinaryOp::BinaryOp(lhs_t &_l, rhs_t &_r) : lhs_(_l), rhs_(_r){}; template -SYCL_BLAS_INLINE typename BinaryOp::index_t +PORTBLAS_INLINE typename BinaryOp::index_t BinaryOp::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool BinaryOp::valid_thread( +PORTBLAS_INLINE bool BinaryOp::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < get_size())); } template -SYCL_BLAS_INLINE typename BinaryOp::value_t +PORTBLAS_INLINE typename BinaryOp::value_t BinaryOp::eval( typename BinaryOp::index_t i) { return operator_t::eval(lhs_.eval(i), rhs_.eval(i)); } template -SYCL_BLAS_INLINE typename BinaryOp::value_t +PORTBLAS_INLINE typename BinaryOp::value_t BinaryOp::eval(cl::sycl::nd_item<1> ndItem) { return BinaryOp::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void BinaryOp::bind( +PORTBLAS_INLINE void BinaryOp::bind( cl::sycl::handler &h) { lhs_.bind(h); rhs_.bind(h); } template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void BinaryOp::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_.adjust_access_displacement(); @@ -395,33 +395,33 @@ template TupleOp::TupleOp(rhs_t &_r) : rhs_(_r) {} template -SYCL_BLAS_INLINE typename TupleOp::index_t TupleOp::get_size() +PORTBLAS_INLINE typename TupleOp::index_t TupleOp::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool TupleOp::valid_thread( +PORTBLAS_INLINE bool TupleOp::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < get_size())); } template -SYCL_BLAS_INLINE typename TupleOp::value_t TupleOp::eval( +PORTBLAS_INLINE typename TupleOp::value_t TupleOp::eval( typename TupleOp::index_t i) { return TupleOp::value_t(i, rhs_.eval(i)); } template -SYCL_BLAS_INLINE typename TupleOp::value_t TupleOp::eval( +PORTBLAS_INLINE typename TupleOp::value_t TupleOp::eval( cl::sycl::nd_item<1> ndItem) { return TupleOp::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void TupleOp::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void TupleOp::bind(cl::sycl::handler &h) { rhs_.bind(h); } template -SYCL_BLAS_INLINE void TupleOp::adjust_access_displacement() { +PORTBLAS_INLINE void TupleOp::adjust_access_displacement() { rhs_.adjust_access_displacement(); } @@ -436,18 +436,18 @@ AssignReduction::AssignReduction(lhs_t &_l, rhs_t &_r, : lhs_(_l), rhs_(_r), local_num_thread_(_blqS), global_num_thread_(_grdS){}; template -SYCL_BLAS_INLINE typename AssignReduction::index_t +PORTBLAS_INLINE typename AssignReduction::index_t AssignReduction::get_size() const { return rhs_.get_size(); } template -SYCL_BLAS_INLINE bool AssignReduction::valid_thread( +PORTBLAS_INLINE bool AssignReduction::valid_thread( cl::sycl::nd_item<1> ndItem) const { return true; } template -SYCL_BLAS_INLINE typename AssignReduction::value_t +PORTBLAS_INLINE typename AssignReduction::value_t AssignReduction::eval( typename AssignReduction::index_t i) { index_t vecS = rhs_.get_size(); @@ -477,14 +477,14 @@ AssignReduction::eval( return val; } template -SYCL_BLAS_INLINE typename AssignReduction::value_t +PORTBLAS_INLINE typename AssignReduction::value_t AssignReduction::eval(cl::sycl::nd_item<1> ndItem) { return AssignReduction::eval( ndItem.get_global_id(0)); } template template -SYCL_BLAS_INLINE typename AssignReduction::value_t +PORTBLAS_INLINE typename AssignReduction::value_t AssignReduction::eval(sharedT scratch, cl::sycl::nd_item<1> ndItem) { index_t localid = ndItem.get_local_id(0); @@ -525,14 +525,14 @@ AssignReduction::eval(sharedT scratch, } template -SYCL_BLAS_INLINE void AssignReduction::bind( +PORTBLAS_INLINE void AssignReduction::bind( cl::sycl::handler &h) { lhs_.bind(h); rhs_.bind(h); } template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void AssignReduction::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_.adjust_access_displacement(); @@ -544,13 +544,13 @@ Rotg::Rotg(operand_t &_a, operand_t &_b, operand_t &_c, : a_{_a}, b_{_b}, c_{_c}, s_{_s} {} template -SYCL_BLAS_INLINE typename Rotg::index_t Rotg::get_size() +PORTBLAS_INLINE typename Rotg::index_t Rotg::get_size() const { return static_cast::index_t>(1); } template -SYCL_BLAS_INLINE typename Rotg::value_t Rotg::eval( +PORTBLAS_INLINE typename Rotg::value_t Rotg::eval( typename Rotg::index_t i) { using zero = constant; using one = constant; @@ -592,19 +592,19 @@ SYCL_BLAS_INLINE typename Rotg::value_t Rotg::eval( } template -SYCL_BLAS_INLINE typename Rotg::value_t Rotg::eval( +PORTBLAS_INLINE typename Rotg::value_t Rotg::eval( cl::sycl::nd_item<1> ndItem) { return Rotg::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE bool Rotg::valid_thread( +PORTBLAS_INLINE bool Rotg::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < Rotg::get_size())); } template -SYCL_BLAS_INLINE void Rotg::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void Rotg::bind(cl::sycl::handler &h) { a_.bind(h); b_.bind(h); c_.bind(h); @@ -612,7 +612,7 @@ SYCL_BLAS_INLINE void Rotg::bind(cl::sycl::handler &h) { } template -SYCL_BLAS_INLINE void Rotg::adjust_access_displacement() { +PORTBLAS_INLINE void Rotg::adjust_access_displacement() { a_.adjust_access_displacement(); b_.adjust_access_displacement(); c_.adjust_access_displacement(); @@ -625,7 +625,7 @@ Rotmg::Rotmg(operand_t &_d1, operand_t &_d2, operand_t &_x1, : d1_{_d1}, d2_{_d2}, x1_{_x1}, y1_{_y1}, param_{_param} {} template -SYCL_BLAS_INLINE typename Rotmg::index_t Rotmg::get_size() +PORTBLAS_INLINE typename Rotmg::index_t Rotmg::get_size() const { return static_cast::index_t>(1); } @@ -643,7 +643,7 @@ SYCL_BLAS_INLINE typename Rotmg::index_t Rotmg::get_size() * 1979. */ template -SYCL_BLAS_INLINE typename Rotmg::value_t Rotmg::eval( +PORTBLAS_INLINE typename Rotmg::value_t Rotmg::eval( typename Rotmg::index_t i) { using zero = constant; using one = constant; @@ -862,19 +862,19 @@ SYCL_BLAS_INLINE typename Rotmg::value_t Rotmg::eval( } template -SYCL_BLAS_INLINE typename Rotmg::value_t Rotmg::eval( +PORTBLAS_INLINE typename Rotmg::value_t Rotmg::eval( cl::sycl::nd_item<1> ndItem) { return Rotmg::eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE bool Rotmg::valid_thread( +PORTBLAS_INLINE bool Rotmg::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < Rotmg::get_size())); } template -SYCL_BLAS_INLINE void Rotmg::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void Rotmg::bind(cl::sycl::handler &h) { d1_.bind(h); d2_.bind(h); x1_.bind(h); @@ -883,7 +883,7 @@ SYCL_BLAS_INLINE void Rotmg::bind(cl::sycl::handler &h) { } template -SYCL_BLAS_INLINE void Rotmg::adjust_access_displacement() { +PORTBLAS_INLINE void Rotmg::adjust_access_displacement() { d1_.adjust_access_displacement(); d2_.adjust_access_displacement(); x1_.adjust_access_displacement(); diff --git a/src/operations/blas2/gbmv.hpp b/src/operations/blas2/gbmv.hpp index 19b950fd2..409a103ec 100644 --- a/src/operations/blas2/gbmv.hpp +++ b/src/operations/blas2/gbmv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gbmv.hpp * @@ -38,7 +38,7 @@ namespace blas { */ template -SYCL_BLAS_INLINE +PORTBLAS_INLINE Gbmv::Gbmv( lhs_t &_l, matrix_t &_matrix, typename Gbmv::Gbmv( template -SYCL_BLAS_INLINE typename Gbmv::index_t Gbmv::get_size() const { return matrix_.get_size(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Gbmv::valid_thread( cl::sycl::nd_item<1> ndItem) const { // Valid threads are established by ::eval. @@ -76,7 +76,7 @@ Gbmv::valid_thread( template -SYCL_BLAS_INLINE typename Gbmv::value_t Gbmv::eval( cl::sycl::nd_item<1> ndItem) { @@ -108,7 +108,7 @@ Gbmv::eval( template -SYCL_BLAS_INLINE void Gbmv::bind(cl::sycl::handler &h) { lhs_.bind(h); matrix_.bind(h); @@ -116,7 +116,7 @@ SYCL_BLAS_INLINE void Gbmv -SYCL_BLAS_INLINE void Gbmv::adjust_access_displacement() { lhs_.adjust_access_displacement(); matrix_.adjust_access_displacement(); diff --git a/src/operations/blas2/gemv.hpp b/src/operations/blas2/gemv.hpp index df32bfe60..34aef5174 100644 --- a/src/operations/blas2/gemv.hpp +++ b/src/operations/blas2/gemv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemv.hpp * @@ -40,19 +40,19 @@ template SumMatrixColumns::SumMatrixColumns(rhs_t &_r) : rhs_(_r) {} template -SYCL_BLAS_INLINE typename SumMatrixColumns::index_t +PORTBLAS_INLINE typename SumMatrixColumns::index_t SumMatrixColumns::get_size() const { return rhs_.get_size_row(); } template -SYCL_BLAS_INLINE bool SumMatrixColumns::valid_thread( +PORTBLAS_INLINE bool SumMatrixColumns::valid_thread( cl::sycl::nd_item<1> ndItem) const { return ((ndItem.get_global_id(0) < get_size())); } template -SYCL_BLAS_INLINE typename SumMatrixColumns::value_t +PORTBLAS_INLINE typename SumMatrixColumns::value_t SumMatrixColumns::eval(typename SumMatrixColumns::index_t i) { auto dimR = rhs_.get_size_row(); auto dimC = rhs_.get_size_col(); @@ -67,18 +67,18 @@ SumMatrixColumns::eval(typename SumMatrixColumns::index_t i) { } template -SYCL_BLAS_INLINE typename SumMatrixColumns::value_t +PORTBLAS_INLINE typename SumMatrixColumns::value_t SumMatrixColumns::eval(cl::sycl::nd_item<1> ndItem) { return eval(ndItem.get_global_id(0)); } template -SYCL_BLAS_INLINE void SumMatrixColumns::bind(cl::sycl::handler &h) { +PORTBLAS_INLINE void SumMatrixColumns::bind(cl::sycl::handler &h) { rhs_.bind(h); } template -SYCL_BLAS_INLINE void SumMatrixColumns::adjust_access_displacement() { +PORTBLAS_INLINE void SumMatrixColumns::adjust_access_displacement() { rhs_.adjust_access_displacement(); } @@ -89,7 +89,7 @@ SYCL_BLAS_INLINE void SumMatrixColumns::adjust_access_displacement() { template -SYCL_BLAS_INLINE +PORTBLAS_INLINE Gemv::Gemv(lhs_t &_l, matrix_t &_matrix_a, vector_t &_vector_x, typename vector_t::index_t &_wgs_per_nc, @@ -109,7 +109,7 @@ Gemv -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Gemv::valid_thread(cl::sycl::nd_item<1>) const { return true; @@ -125,7 +125,7 @@ Gemv -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Gemv::value_t Gemv template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Gemv::value_t Gemv template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemv::extract_input_block(ScratchPointerType matrix_scratch, const index_t &local_id, @@ -350,7 +350,7 @@ Gemv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemv::bind(cl::sycl::handler &h) { lhs_.bind(h); @@ -361,7 +361,7 @@ Gemv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemv::adjust_access_displacement() { lhs_.adjust_access_displacement(); @@ -377,7 +377,7 @@ Gemv -SYCL_BLAS_INLINE +PORTBLAS_INLINE GemvRow:: GemvRow(lhs_t &_l, matrix_t &_matrix, vector_t &_vector, typename GemvRow:: template -SYCL_BLAS_INLINE typename GemvRow::index_t GemvRow::get_size() const { @@ -403,7 +403,7 @@ GemvRow -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool GemvRow::valid_thread(cl::sycl::nd_item<1> ndItem) const { return true; @@ -413,7 +413,7 @@ GemvRow -SYCL_BLAS_INLINE typename GemvRow::value_t GemvRow::eval( typename GemvRow::eval( template -SYCL_BLAS_INLINE typename GemvRow::value_t GemvRow::eval( cl::sycl::nd_item<1> ndItem) { @@ -554,7 +554,7 @@ GemvRow::eval( template template -SYCL_BLAS_INLINE typename GemvRow::value_t GemvRow::eval( local_memory_t shrMem, cl::sycl::nd_item<1> ndItem) { @@ -683,7 +683,7 @@ GemvRow::eval( } template -SYCL_BLAS_INLINE void GemvRow::bind(cl::sycl::handler &h) { lhs_.bind(h); matrix_.bind(h); @@ -691,7 +691,7 @@ SYCL_BLAS_INLINE void GemvRow -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void GemvRow::adjust_access_displacement() { lhs_.adjust_access_displacement(); @@ -708,7 +708,7 @@ GemvRow -SYCL_BLAS_INLINE +PORTBLAS_INLINE GemvCol::GemvCol( lhs_t &_l, matrix_t &_matrix, vector_t &_vector, typename GemvCol::GemvCol( local_memory_size_(_shrMemSize){}; template -SYCL_BLAS_INLINE typename GemvCol::index_t GemvCol::get_size() const { return matrix_.get_size_row(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool GemvCol::valid_thread( cl::sycl::nd_item<1> ndItem) const { return true; @@ -740,7 +740,7 @@ GemvCol::valid_thread( template -SYCL_BLAS_INLINE typename GemvCol::value_t GemvCol::eval(index_t i) { auto dim = vector_.get_size(); @@ -755,7 +755,7 @@ GemvCol::eval(index_t i) { template -SYCL_BLAS_INLINE typename GemvCol::value_t GemvCol::eval( cl::sycl::nd_item<1> ndItem) { @@ -820,7 +820,7 @@ GemvCol::eval( template template -SYCL_BLAS_INLINE typename GemvCol::value_t GemvCol::eval( local_memory_t shrMem, cl::sycl::nd_item<1> ndItem) { @@ -905,7 +905,7 @@ GemvCol::eval( } template -SYCL_BLAS_INLINE void GemvCol::bind(cl::sycl::handler &h) { lhs_.bind(h); matrix_.bind(h); @@ -914,7 +914,7 @@ SYCL_BLAS_INLINE void GemvCol -SYCL_BLAS_INLINE void GemvCol::adjust_access_displacement() { lhs_.adjust_access_displacement(); matrix_.adjust_access_displacement(); diff --git a/src/operations/blas2/ger.hpp b/src/operations/blas2/ger.hpp index 99ff18db4..65300ee8a 100644 --- a/src/operations/blas2/ger.hpp +++ b/src/operations/blas2/ger.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename ger.hpp * @@ -38,7 +38,7 @@ namespace blas { // template template -SYCL_BLAS_INLINE GerRow::GerRow(lhs_t &_l, value_t _scl, rhs_1_t &_r1, rhs_2_t &_r2, index_t &_nWG_row, index_t &_nWG_col, @@ -53,14 +53,14 @@ SYCL_BLAS_INLINE GerRow -SYCL_BLAS_INLINE typename GerRow::index_t GerRow::get_size() const { return rhs_1_.get_size(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool GerRow::valid_thread( cl::sycl::nd_item<1> ndItem) const { return true; @@ -68,7 +68,7 @@ GerRow::valid_thread( template -SYCL_BLAS_INLINE typename GerRow::value_t GerRow::eval( typename GerRow::eval( template -SYCL_BLAS_INLINE typename GerRow::value_t GerRow::eval( cl::sycl::nd_item<1> ndItem) { @@ -158,7 +158,7 @@ GerRow::eval( template template -SYCL_BLAS_INLINE typename GerRow::value_t GerRow::eval( sharedT shrMem, cl::sycl::nd_item<1> ndItem) { @@ -259,7 +259,7 @@ GerRow::eval( } template -SYCL_BLAS_INLINE void GerRow::bind(cl::sycl::handler &h) { lhs_.bind(h); rhs_1_.bind(h); @@ -267,7 +267,7 @@ SYCL_BLAS_INLINE void GerRow -SYCL_BLAS_INLINE void GerRow::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_1_.adjust_access_displacement(); @@ -278,7 +278,7 @@ SYCL_BLAS_INLINE void GerRow template -SYCL_BLAS_INLINE GerCol::GerCol(lhs_t &_l, value_t _scl, rhs_1_t &_r1, rhs_2_t &_r2, index_t &_nWG_row, index_t &_nWG_col, @@ -293,21 +293,21 @@ SYCL_BLAS_INLINE GerCol -SYCL_BLAS_INLINE typename GerCol::index_t GerCol::get_size() const { return rhs_1_.get_size(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool GerCol::valid_thread( cl::sycl::nd_item<1> ndItem) const { return true; } template -SYCL_BLAS_INLINE typename GerCol::value_t GerCol::eval( typename GerCol::eval( template -SYCL_BLAS_INLINE typename GerCol::value_t GerCol::eval( cl::sycl::nd_item<1> ndItem) { @@ -386,7 +386,7 @@ GerCol::eval( template template -SYCL_BLAS_INLINE typename GerCol::value_t GerCol::eval( sharedT shrMem, cl::sycl::nd_item<1> ndItem) { @@ -490,7 +490,7 @@ GerCol::eval( } template -SYCL_BLAS_INLINE void GerCol::bind(cl::sycl::handler &h) { lhs_.bind(h); rhs_1_.bind(h); @@ -499,7 +499,7 @@ SYCL_BLAS_INLINE void GerCol -SYCL_BLAS_INLINE void GerCol::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_1_.adjust_access_displacement(); diff --git a/src/operations/blas2/sbmv.hpp b/src/operations/blas2/sbmv.hpp index 6341873e9..53dc76a25 100644 --- a/src/operations/blas2/sbmv.hpp +++ b/src/operations/blas2/sbmv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sbmv.hpp * @@ -39,7 +39,7 @@ namespace blas { */ template -SYCL_BLAS_INLINE Sbmv::Sbmv( +PORTBLAS_INLINE Sbmv::Sbmv( lhs_t &_l, matrix_t &_matrix, typename Sbmv::index_t &_k, @@ -57,14 +57,14 @@ SYCL_BLAS_INLINE Sbmv::Sbmv( template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Sbmv::index_t Sbmv::get_size() const { return matrix_.get_size(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Sbmv::valid_thread( cl::sycl::nd_item<1> ndItem) const { // Valid threads are established by ::eval. @@ -73,7 +73,7 @@ Sbmv::valid_thread( template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Sbmv::value_t Sbmv::eval( cl::sycl::nd_item<1> ndItem) { @@ -108,7 +108,7 @@ SYCL_BLAS_INLINE template -SYCL_BLAS_INLINE void Sbmv::bind(cl::sycl::handler &h) { lhs_.bind(h); matrix_.bind(h); @@ -116,7 +116,7 @@ SYCL_BLAS_INLINE void Sbmv -SYCL_BLAS_INLINE void Sbmv::adjust_access_displacement() { lhs_.adjust_access_displacement(); matrix_.adjust_access_displacement(); diff --git a/src/operations/blas2/spr.hpp b/src/operations/blas2/spr.hpp index ac584aa25..f8b5a533f 100644 --- a/src/operations/blas2/spr.hpp +++ b/src/operations/blas2/spr.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spr.hpp * @@ -34,7 +34,7 @@ namespace blas { template -SYCL_BLAS_INLINE Spr::Spr( +PORTBLAS_INLINE Spr::Spr( lhs_t& _l, typename rhs_1_t::index_t _N, value_t _alpha, rhs_1_t& _r1, typename rhs_1_t::index_t _incX_1, rhs_2_t& _r2, typename rhs_1_t::index_t _incX_2) @@ -81,7 +81,7 @@ typename rhs_1_t::value_t Spr::eval( } template -SYCL_BLAS_INLINE void Spr::bind( +PORTBLAS_INLINE void Spr::bind( cl::sycl::handler& h) { lhs_.bind(h); rhs_1_.bind(h); @@ -90,7 +90,7 @@ SYCL_BLAS_INLINE void Spr::bind( template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Spr::adjust_access_displacement() { lhs_.adjust_access_displacement(); rhs_1_.adjust_access_displacement(); @@ -99,14 +99,14 @@ Spr::adjust_access_displacement() { template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Spr::index_t Spr::get_size() const { return rhs_1_.get_size(); } template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Spr::valid_thread( cl::sycl::nd_item<1> ndItem) const { return true; diff --git a/src/operations/blas2/tbmv.hpp b/src/operations/blas2/tbmv.hpp index b6e964985..f0b768e14 100644 --- a/src/operations/blas2/tbmv.hpp +++ b/src/operations/blas2/tbmv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tbmv.hpp * @@ -40,7 +40,7 @@ namespace blas { template -SYCL_BLAS_INLINE +PORTBLAS_INLINE Tbmv::Tbmv(lhs_t &_l, matrix_t &_matrix, typename Tbmv -SYCL_BLAS_INLINE typename Tbmv::index_t Tbmv::get_size() const { @@ -61,7 +61,7 @@ Tbmv -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Tbmv::valid_thread(cl::sycl::nd_item<1> ndItem) const { // Valid threads are established by ::eval. @@ -71,7 +71,7 @@ Tbmv -SYCL_BLAS_INLINE typename Tbmv::value_t Tbmv::eval(cl::sycl::nd_item<1> ndItem) { @@ -110,7 +110,7 @@ Tbmv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Tbmv::bind(cl::sycl::handler &h) { lhs_.bind(h); @@ -120,7 +120,7 @@ Tbmv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Tbmv::adjust_access_displacement() { lhs_.adjust_access_displacement(); diff --git a/src/operations/blas2/txsv.hpp b/src/operations/blas2/txsv.hpp index 8e6b64187..2548b919d 100644 --- a/src/operations/blas2/txsv.hpp +++ b/src/operations/blas2/txsv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename txsv.hpp * @@ -36,7 +36,7 @@ template -SYCL_BLAS_INLINE Txsv:: Txsv(vector_t &_l, matrix_t &_matrix, typename Txsv -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Txsv::value_t Txsv template -SYCL_BLAS_INLINE typename Txsv::value_t Txsv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Txsv::bind(cl::sycl::handler &h) { lhs_.bind(h); @@ -312,7 +312,7 @@ template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Txsv::adjust_access_displacement() { lhs_.adjust_access_displacement(); diff --git a/src/operations/blas2/xpmv.hpp b/src/operations/blas2/xpmv.hpp index b54758e1c..dd8e7f2ea 100644 --- a/src/operations/blas2/xpmv.hpp +++ b/src/operations/blas2/xpmv.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename spmv.hpp * @@ -38,7 +38,7 @@ namespace blas { template -SYCL_BLAS_INLINE Xpmv:: Xpmv(lhs_t &_l, matrix_t &_matrix, vector_t &_vector, typename Xpmv -SYCL_BLAS_INLINE bool Xpmv::valid_thread(cl::sycl::nd_item<1> ndItem) const { @@ -68,7 +68,7 @@ template template -SYCL_BLAS_INLINE +PORTBLAS_INLINE typename Xpmv::value_t Xpmv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Xpmv::bind(cl::sycl::handler &h) { lhs_.bind(h); @@ -276,7 +276,7 @@ Xpmv -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Xpmv::adjust_access_displacement() { lhs_.adjust_access_displacement(); diff --git a/src/operations/blas2_trees.hpp b/src/operations/blas2_trees.hpp index a96628cf1..b689fd161 100644 --- a/src/operations/blas2_trees.hpp +++ b/src/operations/blas2_trees.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_trees.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS2_TREES_HPP -#define SYCL_BLAS_BLAS2_TREES_HPP +#ifndef PORTBLAS_BLAS2_TREES_HPP +#define PORTBLAS_BLAS2_TREES_HPP #include "blas2/gbmv.hpp" #include "blas2/gemv.hpp" diff --git a/src/operations/blas3/gemm_common.hpp b/src/operations/blas3/gemm_common.hpp index 6610bbb70..4966b9f13 100644 --- a/src/operations/blas3/gemm_common.hpp +++ b/src/operations/blas3/gemm_common.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_common.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_GEMM_COMMON_HPP -#define SYCL_BLAS_BLAS3_GEMM_COMMON_HPP +#ifndef PORTBLAS_BLAS3_GEMM_COMMON_HPP +#define PORTBLAS_BLAS3_GEMM_COMMON_HPP #include "operations/blas3_trees.h" #include "views/view.h" @@ -60,7 +60,7 @@ ENABLE_TYPE_STRING(double) template -SYCL_BLAS_INLINE std::string +PORTBLAS_INLINE std::string Tile::get_type_string() noexcept { std::ostringstream str{}; @@ -80,14 +80,14 @@ Tile -SYCL_BLAS_INLINE bool do_check(bool cond) { +PORTBLAS_INLINE bool do_check(bool cond) { return cond; } template <> -SYCL_BLAS_INLINE bool do_check(bool) { +PORTBLAS_INLINE bool do_check(bool) { return true; } } // namespace blas -#endif // SYCL_BLAS_BLAS3_GEMM_COMMON_HPP +#endif // PORTBLAS_BLAS3_GEMM_COMMON_HPP diff --git a/src/operations/blas3/gemm_interleaved.hpp b/src/operations/blas3/gemm_interleaved.hpp index 553694053..551bb465a 100644 --- a/src/operations/blas3/gemm_interleaved.hpp +++ b/src/operations/blas3/gemm_interleaved.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_interleaved.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_GEMM_INTERLEAVED_HPP -#define SYCL_BLAS_BLAS3_GEMM_INTERLEAVED_HPP +#ifndef PORTBLAS_BLAS3_GEMM_INTERLEAVED_HPP +#define PORTBLAS_BLAS3_GEMM_INTERLEAVED_HPP #include "gemm_common.hpp" @@ -47,7 +47,7 @@ using address_t = cl::sycl::access::address_space; * @brief Load a packet of size 1. */ template -SYCL_BLAS_INLINE void load(T &packet, PtrT ptr) { +PORTBLAS_INLINE void load(T &packet, PtrT ptr) { packet = *ptr; } @@ -55,7 +55,7 @@ SYCL_BLAS_INLINE void load(T &packet, PtrT ptr) { * @brief Store a packet of size 1. */ template -SYCL_BLAS_INLINE void store(T packet, PtrT ptr) { +PORTBLAS_INLINE void store(T packet, PtrT ptr) { *ptr = packet; } @@ -64,7 +64,7 @@ SYCL_BLAS_INLINE void store(T packet, PtrT ptr) { */ template -SYCL_BLAS_INLINE void load(cl::sycl::vec &packet, PtrT ptr) { +PORTBLAS_INLINE void load(cl::sycl::vec &packet, PtrT ptr) { packet.template load
(0, cl::sycl::multi_ptr(ptr)); } @@ -73,7 +73,7 @@ SYCL_BLAS_INLINE void load(cl::sycl::vec &packet, PtrT ptr) { */ template -SYCL_BLAS_INLINE void store(const cl::sycl::vec &packet, PtrT ptr) { +PORTBLAS_INLINE void store(const cl::sycl::vec &packet, PtrT ptr) { packet.template store
(0, cl::sycl::multi_ptr(ptr)); } @@ -158,7 +158,7 @@ class Gemm get_nd_range(index_t) const noexcept { + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range(index_t) const noexcept { const index_t number_of_block_per_row = ((m_ - 1) / block_rows) + 1; const index_t number_of_block_per_cols = ((n_ - 1) / block_cols) + 1; @@ -205,11 +205,11 @@ class Gemm(nwg * wgs, wgs); } - SYCL_BLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { + PORTBLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { return true; } - SYCL_BLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { + PORTBLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { auto A = a_.get_pointer(); auto B = b_.get_pointer(); auto C = c_.get_pointer(); @@ -266,7 +266,7 @@ class Gemm(boundary_check, m_stride, n_stride, mb_start, @@ -279,7 +279,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_panel(check_t boundary_check, index_t m_stride, + PORTBLAS_INLINE void compute_panel(check_t boundary_check, index_t m_stride, index_t n_stride, index_t mb_start, index_t m_start, index_t n_start, in_ptr_t A, in_ptr_t B, out_ptr_t C) { @@ -306,7 +306,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load( + PORTBLAS_INLINE typename std::enable_if::type load( check_t, index_t, index_t, index_t, packet_type *reg_res, ptr_t input, index_t stride) { #pragma unroll @@ -322,7 +322,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load( + PORTBLAS_INLINE typename std::enable_if::type load( check_t boundary_check, index_t index_start, index_t mb_start, index_t dim_size, packet_type *reg_res, ptr_t input, index_t stride) { #pragma unroll @@ -361,7 +361,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type store( + PORTBLAS_INLINE typename std::enable_if::type store( check_t boundary_check, index_t m_start, index_t n_start, index_t mb_start, packet_type *reg_res, ptr_t C) { #pragma unroll @@ -403,7 +403,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type store( + PORTBLAS_INLINE typename std::enable_if::type store( check_t, index_t, index_t, index_t, packet_type *reg_res, ptr_t C) { #pragma unroll for (int i = 0; i < item_cols; ++i) { @@ -424,7 +424,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( check_t boundary_check, index_t m_start, index_t n_start, index_t mb_start, packet_type *reg_res, ptr_t C) { #pragma unroll @@ -465,7 +465,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( check_t, index_t, index_t, index_t, packet_type *reg_res, ptr_t) { #pragma unroll for (int i = 0; i < item_rows * item_cols * (item_batchs / VectorSize); @@ -481,7 +481,7 @@ class Gemm; static constexpr int packet_size = vector_size; template - SYCL_BLAS_INLINE static constexpr bool check_size() { + PORTBLAS_INLINE static constexpr bool check_size() { return packet_size == 1 || dimension == packet_size; } #else @@ -50,7 +50,7 @@ struct Packetize { using PacketType = cl::sycl::vec; static constexpr int packet_size = 1; template - SYCL_BLAS_INLINE static constexpr bool check_size() { + PORTBLAS_INLINE static constexpr bool check_size() { return true; } #endif @@ -65,7 +65,7 @@ struct Packetize { template - static SYCL_BLAS_INLINE typename std::enable_if::type load( + static PORTBLAS_INLINE typename std::enable_if::type load( const bool in_range, SrcPointerType src, DestPointerType dest, EdgePredicate) { #ifdef SB_ENABLE_JOINT_MATRIX @@ -100,7 +100,7 @@ struct Packetize { * @tparam ld The leading dimension of the destination memory. */ template - static SYCL_BLAS_INLINE typename std::enable_if::type load( + static PORTBLAS_INLINE typename std::enable_if::type load( const bool in_range, SrcPointerType src, DestPointerType dest, EdgePredicate edge_in_range) { PacketType packet{}; @@ -124,7 +124,7 @@ struct Packetize { * @tparam trans Whether the source matrix is transposed or not. * @tparam ld The leading dimension of the destination memory.*/ template - static SYCL_BLAS_INLINE typename std::enable_if::type store( + static PORTBLAS_INLINE typename std::enable_if::type store( PacketType &packet, DestPointerType dest) { #ifdef SB_ENABLE_JOINT_MATRIX using address_t = cl::sycl::access::address_space; @@ -160,7 +160,7 @@ struct Packetize { * @tparam trans Whether the source matrix is transposed or not. * @tparam ld The leading dimension of the destination memory.*/ template - static SYCL_BLAS_INLINE typename std::enable_if::type store( + static PORTBLAS_INLINE typename std::enable_if::type store( PacketType &packet, DestPointerType dest) { using address_t = cl::sycl::access::address_space; #ifdef SB_ENABLE_JOINT_MATRIX @@ -187,4 +187,4 @@ struct Packetize { }; } // namespace blas -#endif // SYCL_BLAS_BLAS3_GEMM_LOAD_STORE_HPP +#endif // PORTBLAS_BLAS3_GEMM_LOAD_STORE_HPP diff --git a/src/operations/blas3/gemm_local.hpp b/src/operations/blas3/gemm_local.hpp index 0837e07c7..c2307e914 100644 --- a/src/operations/blas3/gemm_local.hpp +++ b/src/operations/blas3/gemm_local.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_local.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_LOCAL_GEMM_HPP -#define SYCL_BLAS_BLAS3_LOCAL_GEMM_HPP +#ifndef PORTBLAS_BLAS3_LOCAL_GEMM_HPP +#define PORTBLAS_BLAS3_LOCAL_GEMM_HPP #include "gemm_common.hpp" #include "gemm_load_store.hpp" @@ -165,7 +165,7 @@ class Gemm get_nd_range( + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range( index_t compute_units) const noexcept { const cl::sycl::range<1> nwg(get_workgroup_cluster() * get_num_workgroup_cluster(compute_units)); @@ -240,7 +240,7 @@ class Gemm(nwg * wgs, wgs); } - SYCL_BLAS_INLINE index_t get_size() const { + PORTBLAS_INLINE index_t get_size() const { return a_.get_size_row() * b_.get_size_col(); } @@ -251,7 +251,7 @@ class Gemm - SYCL_BLAS_INLINE void eval(local_memory_t scratch_acc, + PORTBLAS_INLINE void eval(local_memory_t scratch_acc, const cl::sycl::nd_item<1> &id) noexcept { index_t m = a_.get_size_row(); index_t n = b_.get_size_col(); @@ -358,7 +358,7 @@ class Gemm &ndItem) const { + PORTBLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &ndItem) const { return true; } @@ -368,7 +368,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType C, const index_t &mc, const index_t &nc, const index_t &ldc, const bool out_of_range) { if (out_of_range) { @@ -397,7 +397,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType, const index_t &, const index_t &, const index_t &, const bool) { #pragma unroll @@ -422,7 +422,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_panel_gemm( + PORTBLAS_INLINE void compute_panel_gemm( const cl::sycl::nd_item<1> &id, const index_t &item_id, const index_t &row_a, const index_t &col_a, const index_t &row_b, const index_t &col_b, const index_t &m, const index_t &n, @@ -517,14 +517,14 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type store_packet( + PORTBLAS_INLINE typename std::enable_if::type store_packet( element_t *reg, OutputPointerType out_ptr) { *out_ptr = alpha_ * (*reg); } template - SYCL_BLAS_INLINE typename std::enable_if::type store_packet( + PORTBLAS_INLINE typename std::enable_if::type store_packet( element_t *reg, OutputPointerType out_ptr) { vector_t out_vec{}; @@ -554,7 +554,7 @@ class Gemm - SYCL_BLAS_INLINE void store_output_block(index_t, index_t mc, index_t nc, + PORTBLAS_INLINE void store_output_block(index_t, index_t mc, index_t nc, OutputPointerType C, index_t ldc, element_t *reg_res, const bool out_of_range) noexcept { @@ -589,7 +589,7 @@ class Gemm - SYCL_BLAS_INLINE void extract_input_blocks( + PORTBLAS_INLINE void extract_input_blocks( index_t item_id, index_t m, index_t n, index_t k, index_t row_a, index_t col_a, index_t row_b, index_t col_b, InputPointerType A, index_t lda, InputPointerType B, index_t ldb, ScratchPointerType sB, @@ -602,16 +602,16 @@ class Gemm( item_id, row_a, col_a, A, lda, sA, - [&](index_t, index_t cr) SYCL_BLAS_ALWAYS_INLINE { return cr < m; }, + [&](index_t, index_t cr) PORTBLAS_ALWAYS_INLINE { return cr < m; }, [&](index_t ic, index_t cc) - SYCL_BLAS_ALWAYS_INLINE { return cc < k - ic; }); + PORTBLAS_ALWAYS_INLINE { return cc < k - ic; }); extract_block( item_id, row_b, col_b, B, ldb, sB, [&](index_t ir, index_t cr) - SYCL_BLAS_ALWAYS_INLINE { return cr < k - ir; }, - [&](index_t, index_t cc) SYCL_BLAS_ALWAYS_INLINE { return cc < n; }); + PORTBLAS_ALWAYS_INLINE { return cr < k - ir; }, + [&](index_t, index_t cc) PORTBLAS_ALWAYS_INLINE { return cc < n; }); } /*! @@ -652,7 +652,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type extract_block( + PORTBLAS_INLINE typename std::enable_if::type extract_block( index_t item_id, index_t row, index_t col, InputPointerType ptr, index_t ld, ScratchPointerType scratch, RowPredicate in_row, ColPredicate in_col) { @@ -689,7 +689,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type extract_block( + PORTBLAS_INLINE typename std::enable_if::type extract_block( index_t item_id, index_t row, index_t col, InputPointerType ptr, index_t ld, ScratchPointerType scratch, RowPredicate in_row, ColPredicate in_col) { @@ -706,7 +706,7 @@ class Gemm(in_col( (item_id * multiplier) % cols, multiplier - 1)); - auto edge_in_range = [&](const index_t &ofs) SYCL_BLAS_ALWAYS_INLINE { + auto edge_in_range = [&](const index_t &ofs) PORTBLAS_ALWAYS_INLINE { return in_col((item_id * multiplier) % cols, ofs) && in_row((item_id * multiplier) / cols, row_ofs); }; @@ -736,7 +736,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_block_gemm(index_t, InputPointerType B, + PORTBLAS_INLINE void compute_block_gemm(index_t, InputPointerType B, InputPointerType A, element_t *reg_a, element_t ®_b, element_t *reg_res) noexcept { @@ -787,7 +787,7 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &id, index_t &ofs_sign, P &s, Ps &... ss) noexcept { s += ofs_sign * o; @@ -795,13 +795,13 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &, index_t &ofs_sign) noexcept { ofs_sign = -ofs_sign; } template - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &id, index_t &, Ps &...) noexcept { id.barrier(cl::sycl::access::fence_space::local_space); } @@ -812,7 +812,7 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type load_symm( + static PORTBLAS_INLINE typename std::enable_if::type load_symm( index_t row, index_t col, index_t col_ofs, index_t ld, const bool in_range, SrcPointerType src, DestPointerType dest, EdgePredicate edge_in_range) { @@ -841,7 +841,7 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type load_symm( + static PORTBLAS_INLINE typename std::enable_if::type load_symm( index_t row, index_t col, index_t col_ofs, index_t ld, const bool in_range, SrcPointerType src, DestPointerType dest, EdgePredicate edge_in_range) { @@ -855,7 +855,7 @@ class Gemm - static SYCL_BLAS_INLINE bool in_triangle(index_t row, index_t col) { + static PORTBLAS_INLINE bool in_triangle(index_t row, index_t col) { // If the matrix is symmetric, the valid values are expected on the lower // triangle unless it is transposed, in which case valid values will // be on the upper side. @@ -875,4 +875,4 @@ class Gemm get_nd_range(index_t) const noexcept { + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range(index_t) const noexcept { size_t x_groups = static_cast((get_wg_x_cluster() - 1) / jm_row_frags + 1); size_t y_groups = @@ -249,7 +249,7 @@ class Gemm - SYCL_BLAS_INLINE void eval(local_memory_t scratch_acc, + PORTBLAS_INLINE void eval(local_memory_t scratch_acc, const cl::sycl::nd_item<1> &id) noexcept { index_t m = a_.get_size_row(); index_t n = b_.get_size_col(); @@ -405,7 +405,7 @@ class Gemm &ndItem) const { + PORTBLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &ndItem) const { return true; } @@ -426,7 +426,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_panel_gemm( + PORTBLAS_INLINE void compute_panel_gemm( const cl::sycl::nd_item<1> &id, const index_t &item_id, const index_t &m, const index_t &n, const index_t &orig_k, const index_t &mc, const index_t &nc, const index_t &a_size, const index_t &b_size, @@ -515,7 +515,7 @@ class Gemm - SYCL_BLAS_INLINE void store_output_block(cl::sycl::nd_item<1> id, index_t mc, + PORTBLAS_INLINE void store_output_block(cl::sycl::nd_item<1> id, index_t mc, index_t nc, OutputPointerType C, ScratchPointerType scratch, index_t ldc, @@ -664,7 +664,7 @@ class Gemm - SYCL_BLAS_INLINE void extract_input_blocks( + PORTBLAS_INLINE void extract_input_blocks( index_t item_id, index_t m, index_t n, index_t k, InputPointerType A, index_t lda, InputPointerType B, index_t ldb, ScratchPointerType sB, ScratchPointerType sA, const bool out_of_range) noexcept { @@ -675,15 +675,15 @@ class Gemm( item_id, A, lda, sA, - [&](index_t, index_t cr) SYCL_BLAS_ALWAYS_INLINE { return cr < m; }, + [&](index_t, index_t cr) PORTBLAS_ALWAYS_INLINE { return cr < m; }, [&](index_t ic, index_t cc) - SYCL_BLAS_ALWAYS_INLINE { return cc < k - ic; }); + PORTBLAS_ALWAYS_INLINE { return cc < k - ic; }); extract_block( item_id, B, ldb, sB, [&](index_t ir, index_t cr) - SYCL_BLAS_ALWAYS_INLINE { return cr < k - ir; }, - [&](index_t, index_t cc) SYCL_BLAS_ALWAYS_INLINE { return cc < n; }); + PORTBLAS_ALWAYS_INLINE { return cr < k - ir; }, + [&](index_t, index_t cc) PORTBLAS_ALWAYS_INLINE { return cc < n; }); } /*! @@ -722,7 +722,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type extract_block( + PORTBLAS_INLINE typename std::enable_if::type extract_block( index_t item_id, InputPointerType ptr, index_t ld, ScratchPointerType scratch, RowPredicate in_row, ColPredicate in_col) { constexpr index_t bs = rows * cols; @@ -751,7 +751,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type extract_block( + PORTBLAS_INLINE typename std::enable_if::type extract_block( index_t item_id, InputPointerType ptr, index_t ld, ScratchPointerType scratch, RowPredicate in_row, ColPredicate in_col) { constexpr index_t bs = rows * cols; @@ -775,7 +775,7 @@ class Gemm( in_range, ptr + row_ofs * ld, scratch + local_row_ofs, - [&](const index_t &ofs) SYCL_BLAS_ALWAYS_INLINE { + [&](const index_t &ofs) PORTBLAS_ALWAYS_INLINE { return in_col((item_id * multiplier) % cols, ofs) && in_row((item_id * multiplier) / cols, row_ofs); }); @@ -794,7 +794,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_block_gemm( + PORTBLAS_INLINE void compute_block_gemm( const cl::sycl::nd_item<1> &id, InputPointerType s2, InputPointerType s4, CType (®_res)[frags_per_sg]) noexcept { using namespace cl::sycl::ext::oneapi::experimental::matrix; @@ -849,7 +849,7 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &id, index_t &ofs_sign, P &s, Ps &... ss) noexcept { s += ofs_sign * o; @@ -857,13 +857,13 @@ class Gemm - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &, index_t &ofs_sign) noexcept { ofs_sign = -ofs_sign; } template - static SYCL_BLAS_INLINE typename std::enable_if::type sync_smem( + static PORTBLAS_INLINE typename std::enable_if::type sync_smem( const cl::sycl::nd_item<1> &id, index_t &, Ps &...) noexcept { id.barrier(cl::sycl::access::fence_space::local_space); } @@ -874,4 +874,4 @@ class Gemm get_nd_range( + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range( index_t compute_units) const noexcept { const cl::sycl::range<1> nwg(get_workgroup_cluster() * get_num_workgroup_cluster(compute_units)); @@ -172,15 +172,15 @@ class Gemm(nwg * wgs, wgs); } - SYCL_BLAS_INLINE index_t get_size() const { + PORTBLAS_INLINE index_t get_size() const { return a_.get_size_row() * b_.get_size_col(); } - SYCL_BLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { + PORTBLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { return true; } - SYCL_BLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { + PORTBLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { index_t m = a_.get_size_row(); index_t n = b_.get_size_col(); const index_t original_m = m; @@ -267,14 +267,14 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType C, const index_t &ldc, const index_t &dim_m_c_start, const index_t &dim_n_c_start, CheckBoundaryType check_boundary, bool out_of_range) { @@ -337,7 +337,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType, const index_t &, const index_t &, const index_t &, CheckBoundaryType, bool) { #pragma unroll @@ -349,7 +349,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_gemm_no_shared_pannel( + PORTBLAS_INLINE void compute_gemm_no_shared_pannel( A_t orig_A, B_t orig_B, C_t orig_C, const index_t &a_size, const index_t &b_size, const index_t &c_size, index_t orig_k, index_t k, const index_t &dim_m_a_start, const index_t &dim_n_b_start, @@ -451,7 +451,7 @@ class Gemm - SYCL_BLAS_INLINE void load_and_compute_block( + PORTBLAS_INLINE void load_and_compute_block( PointerType A, PointerType B, BoundaryCheckOriginalM boundary_check_original_m, BoundaryCheckN boundary_check_n, const index_t &A_ptr_index, @@ -464,7 +464,7 @@ class Gemm( A, reg_a, A_ptr_index, lda, boundary_check_original_m, - [=](const index_t &idx) SYCL_BLAS_ALWAYS_INLINE { return idx < k; }, + [=](const index_t &idx) PORTBLAS_ALWAYS_INLINE { return idx < k; }, out_of_range); #pragma unroll @@ -478,7 +478,7 @@ class Gemm( B + ofs, reg_b, j, col_ofs, - [=](const index_t &idx) SYCL_BLAS_ALWAYS_INLINE { return idx < k; }, + [=](const index_t &idx) PORTBLAS_ALWAYS_INLINE { return idx < k; }, boundary_check_n, out_of_range); /* @@ -530,7 +530,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load_block_a( + PORTBLAS_INLINE typename std::enable_if::type load_block_a( PointerType ptr, element_t *reg, const index_t &ptr_next, const index_t &ld, const RowCheckType &is_valid_row, const ColCheckType &is_valid_col, const bool out_of_range) noexcept { @@ -614,7 +614,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load_block_a( + PORTBLAS_INLINE typename std::enable_if::type load_block_a( PointerType ptr, element_t *reg, const index_t &, const index_t &ld, const RowCheckType &is_valid_row, const ColCheckType &is_valid_col, const bool out_of_range) noexcept { @@ -693,7 +693,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load_single_b( + PORTBLAS_INLINE typename std::enable_if::type load_single_b( PointerType ptr, element_t *reg, const index_t &, const index_t &col_ofs, const RowCheckType &is_valid_row, const ColCheckType &is_valid_col, const bool out_of_range) noexcept { @@ -756,7 +756,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type load_single_b( + PORTBLAS_INLINE typename std::enable_if::type load_single_b( PointerType ptr, element_t *reg, const index_t &row_ofs, const index_t &, const RowCheckType &is_valid_row, const ColCheckType &is_valid_col, const bool out_of_range) noexcept { @@ -800,7 +800,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type + PORTBLAS_INLINE typename std::enable_if::type compute_block_gemm_no_shared(index_t iteration, element_t *reg_a, element_t *reg_b, element_t *reg_res) noexcept { reg_res += iteration * item_rows; @@ -828,7 +828,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if<(packet_size != 1 && trans)>::type + PORTBLAS_INLINE typename std::enable_if<(packet_size != 1 && trans)>::type compute_block_gemm_no_shared(index_t iteration, element_t *reg_a, element_t *reg_b, element_t *reg_res) noexcept { reg_a += iteration * item_rows; @@ -854,7 +854,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if<(packet_size == 1 && trans)>::type + PORTBLAS_INLINE typename std::enable_if<(packet_size == 1 && trans)>::type compute_block_gemm_no_shared(index_t iteration, element_t *reg_a, element_t *reg_b, element_t *reg_res) noexcept { reg_res += iteration * item_rows; @@ -886,7 +886,7 @@ class Gemm - SYCL_BLAS_INLINE void store(PointerType C, element_t *reg_res, + PORTBLAS_INLINE void store(PointerType C, element_t *reg_res, const index_t &dim_m_c_start, const index_t &dim_n_c_start, const check_boundary &chk_boundary, @@ -920,4 +920,4 @@ class Gemm get_nd_range( + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range( index_t compute_units) const noexcept { const cl::sycl::range<1> nwg(get_workgroup_cluster() * get_num_workgroup_cluster(compute_units)); @@ -168,15 +168,15 @@ class Gemm(nwg * wgs, wgs); } - SYCL_BLAS_INLINE index_t get_size() const { + PORTBLAS_INLINE index_t get_size() const { return a_.get_size_row() * b_.get_size_col(); } - SYCL_BLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { + PORTBLAS_INLINE bool valid_thread(const cl::sycl::nd_item<1> &) const { return true; } - SYCL_BLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { + PORTBLAS_INLINE void eval(cl::sycl::nd_item<1> id) noexcept { index_t m = a_.get_size_row(); index_t n = b_.get_size_col(); const index_t k = a_.get_size_col(); @@ -258,13 +258,13 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType C, const index_t &ldc, const index_t &dim_m_c_start, const index_t &dim_n_c_start, CheckBoundaryType check_boundary, bool out_of_range) { @@ -327,7 +327,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type scaling_c( + PORTBLAS_INLINE typename std::enable_if::type scaling_c( element_t *reg_res, InputPointerType, const index_t &, const index_t &, const index_t &, CheckBoundaryType, bool) { #pragma unroll @@ -340,7 +340,7 @@ class Gemm - SYCL_BLAS_INLINE void compute_gemm_no_shared_pannel( + PORTBLAS_INLINE void compute_gemm_no_shared_pannel( A_t orig_A, B_t orig_B, C_t orig_C, const index_t &a_size, const index_t &b_size, const index_t &c_size, index_t orig_k, index_t k, const index_t &dim_m_a_start, const index_t &dim_n_b_start, @@ -444,7 +444,7 @@ class Gemm - SYCL_BLAS_INLINE void load(PointerType ptr, element_t *reg, const index_t &ld, + PORTBLAS_INLINE void load(PointerType ptr, element_t *reg, const index_t &ld, index_t index, const check_boundary &chk_boundary, const bool out_of_range) noexcept { if (out_of_range) { @@ -481,7 +481,7 @@ class Gemm - SYCL_BLAS_INLINE typename std::enable_if::type store_packet( + PORTBLAS_INLINE typename std::enable_if::type store_packet( element_t *reg, OutputPointerType out_ptr) { *out_ptr = alpha_ * (*reg); } template - SYCL_BLAS_INLINE typename std::enable_if::type store_packet( + PORTBLAS_INLINE typename std::enable_if::type store_packet( element_t *reg, OutputPointerType out_ptr) { cl::sycl::vec out_vec{0}; @@ -530,7 +530,7 @@ class Gemm - SYCL_BLAS_INLINE void store(PointerType C, element_t *reg_res, + PORTBLAS_INLINE void store(PointerType C, element_t *reg_res, const index_t &dim_m_c_start, const index_t &dim_n_c_start, const check_boundary &chk_boundary, @@ -566,4 +566,4 @@ class Gemm get_nd_range( + PORTBLAS_INLINE cl::sycl::nd_range<1> get_nd_range( index_t compute_units) noexcept { const cl::sycl::range<1> nwg(get_workgroup_cluster(compute_units)); const cl::sycl::range<1> wgs(local_thread_size); @@ -220,7 +220,7 @@ class GemmPartial - SYCL_BLAS_INLINE void eval(local_memory_t scratch, + PORTBLAS_INLINE void eval(local_memory_t scratch, cl::sycl::nd_item<1> id) noexcept { /* Pointers to the scratch memory (lhs and rhs) */ value_t* scratch_ptr = scratch.localAcc.get_pointer(); @@ -356,7 +356,7 @@ class GemmPartial - SYCL_BLAS_INLINE void extract_input_blocks( + PORTBLAS_INLINE void extract_input_blocks( const index_t& lhs_row, const index_t& lhs_col, const index_t& rhs_row, const index_t& rhs_col, const index_t& tile_idx, local_ptr_t scratch_ptr, local_ptr_t rhs_scratch_ptr, const index_t& global_m_offset, @@ -378,7 +378,7 @@ class GemmPartial - SYCL_BLAS_INLINE void load_blocks( + PORTBLAS_INLINE void load_blocks( const index_t& lhs_row, const index_t& lhs_col, const index_t& rhs_row, const index_t& rhs_col, const index_t& tile_idx, local_ptr_t scratch_ptr, local_ptr_t rhs_scratch_ptr, const index_t& global_m_offset, @@ -404,7 +404,7 @@ class GemmPartial - static SYCL_BLAS_INLINE void load_block( + static PORTBLAS_INLINE void load_block( const index_t& local_row, const index_t& local_col, const index_t& tile_idx, const input_t& in_view, const index_t& leading_dim, local_ptr_t local_ptr, @@ -458,4 +458,4 @@ class GemmPartial -SYCL_BLAS_INLINE +PORTBLAS_INLINE Gemm:: @@ -76,7 +76,7 @@ template -SYCL_BLAS_INLINE std::string +PORTBLAS_INLINE std::string Gemm -SYCL_BLAS_INLINE typename Gemm -SYCL_BLAS_INLINE typename Gemm -SYCL_BLAS_INLINE cl::sycl::nd_range<1> +PORTBLAS_INLINE cl::sycl::nd_range<1> Gemm -SYCL_BLAS_INLINE typename Gemm -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Gemm -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemm -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemm -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Gemm -SYCL_BLAS_INLINE DiagonalBlocksInverter:: +PORTBLAS_INLINE DiagonalBlocksInverter:: DiagonalBlocksInverter(matrix_t& A, matrix_t& invA) : A_(A), invA_(invA), N_(A_.get_size_col()), lda_(A_.getSizeL()) {} template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool DiagonalBlocksInverter::valid_thread( cl::sycl::nd_item<1> id) const { return true; } template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void DiagonalBlocksInverter::bind( cl::sycl::handler& cgh) { A_.bind(cgh); @@ -51,7 +51,7 @@ DiagonalBlocksInverter::bind( } template -SYCL_BLAS_INLINE void DiagonalBlocksInverter< +PORTBLAS_INLINE void DiagonalBlocksInverter< UnitDiag, Upper, BlockSize, matrix_t>::adjust_access_displacement() { A_.adjust_access_displacement(); invA_.adjust_access_displacement(); @@ -59,7 +59,7 @@ SYCL_BLAS_INLINE void DiagonalBlocksInverter< template template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void DiagonalBlocksInverter::eval( local_memory_t localMem, cl::sycl::nd_item<1> item) noexcept { auto A = A_.get_data().get_pointer() + A_.get_access_displacement(); @@ -150,4 +150,4 @@ DiagonalBlocksInverter::eval( } // namespace blas -#endif // SYCL_BLAS_BLAS3_TRSM_HPP +#endif // PORTBLAS_BLAS3_TRSM_HPP diff --git a/src/operations/blas3_trees.hpp b/src/operations/blas3_trees.hpp index a188d73f8..4eb8eb18a 100644 --- a/src/operations/blas3_trees.hpp +++ b/src/operations/blas3_trees.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_trees.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_BLAS3_TREES_HPP -#define SYCL_BLAS_BLAS3_TREES_HPP +#ifndef PORTBLAS_BLAS3_TREES_HPP +#define PORTBLAS_BLAS3_TREES_HPP #include "blas3/gemm_interleaved.hpp" #include "blas3/gemm_local.hpp" @@ -33,4 +33,4 @@ #include "blas3/gemm_partial_local.hpp" #include "blas3/gemm_ref.hpp" #include "blas3/trsm.hpp" -#endif // SYCL_BLAS_BLAS3_TREES_HPP +#endif // PORTBLAS_BLAS3_TREES_HPP diff --git a/src/operations/blas_constants.hpp b/src/operations/blas_constants.hpp index cb2ac59fc..fb65a441e 100644 --- a/src/operations/blas_constants.hpp +++ b/src/operations/blas_constants.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_constants.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_CONSTANTS_HPP -#define SYCL_BLAS_CONSTANTS_HPP +#ifndef PORTBLAS_CONSTANTS_HPP +#define PORTBLAS_CONSTANTS_HPP #include "operations/blas_constants.h" #endif // BLAS_CONSTANTS_HPP diff --git a/src/operations/blas_operators.hpp b/src/operations/blas_operators.hpp index eb9ec9741..a2960f300 100644 --- a/src/operations/blas_operators.hpp +++ b/src/operations/blas_operators.hpp @@ -17,15 +17,15 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_operators.hpp * **************************************************************************/ // NO H for this one as this one is internal. but all the macro will be // generated by cmake in cpp file -#ifndef SYCL_BLAS_OPERATORS_HPP -#define SYCL_BLAS_OPERATORS_HPP +#ifndef PORTBLAS_OPERATORS_HPP +#define PORTBLAS_OPERATORS_HPP #include #include @@ -121,14 +121,14 @@ struct AbsoluteValue { #endif // BLAS_DATA_TYPE_HALF template - static SYCL_BLAS_INLINE value_t eval( + static PORTBLAS_INLINE value_t eval( const value_t &val, typename std::enable_if::value>::type * = 0) { return cl::sycl::abs(val); } template - static SYCL_BLAS_INLINE value_t + static PORTBLAS_INLINE value_t eval(const value_t &val, typename std::enable_if::value>::type * = 0) { return cl::sycl::fabs(val); @@ -157,63 +157,63 @@ Definitions of unary operators */ struct AdditionIdentity : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t) { return constant::value(); } }; struct ProductIdentity : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (constant::value()); } }; struct IdentityOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (r); } }; struct SignOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return cl::sycl::sign(r); } }; struct NegationOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (-r); } }; struct SqrtOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (cl::sycl::sqrt(r)); } }; struct HypotenuseOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const lhs_t l, const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const lhs_t l, const rhs_t r) { return (cl::sycl::hypot(l, r)); } }; struct DoubleOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (r + r); } }; struct SquareOperator : public Operators { template - static SYCL_BLAS_INLINE rhs_t eval(const rhs_t r) { + static PORTBLAS_INLINE rhs_t eval(const rhs_t r) { return (r * r); } }; @@ -224,18 +224,18 @@ struct SquareOperator : public Operators { struct AddOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return (l + r); } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -243,18 +243,18 @@ struct AddOperator : public Operators { struct ProductOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return (l * r); } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -262,18 +262,18 @@ struct ProductOperator : public Operators { struct DivisionOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return (l / r); } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -281,18 +281,18 @@ struct DivisionOperator : public Operators { struct MeanOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval( + static PORTBLAS_INLINE typename StripASP::type eval( const element_t &accumulator, const element_t &val) { return accumulator + val; } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &r) { return (l / static_cast(r)); } @@ -300,18 +300,18 @@ struct MeanOperator : public Operators { struct MaxOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return ((l > r) ? l : r); } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -319,18 +319,18 @@ struct MaxOperator : public Operators { struct MinOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return ((l < r) ? l : r); } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -338,18 +338,18 @@ struct MinOperator : public Operators { struct AbsoluteAddOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { return AbsoluteValue::eval(l) + AbsoluteValue::eval(r); } // namespace blas template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant::value(); } template - static SYCL_BLAS_INLINE typename StripASP::type get_final_value( + static PORTBLAS_INLINE typename StripASP::type get_final_value( const element_t &l, const index_t &) { return l; } @@ -357,7 +357,7 @@ struct AbsoluteAddOperator : public Operators { struct IMaxOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { if (AbsoluteValue::eval( static_cast::type>(l).get_value()) < @@ -375,7 +375,7 @@ struct IMaxOperator : public Operators { } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant_pair::value(); } @@ -383,7 +383,7 @@ struct IMaxOperator : public Operators { struct IMinOperator : public Operators { template - static SYCL_BLAS_INLINE typename StripASP::type eval(const lhs_t &l, + static PORTBLAS_INLINE typename StripASP::type eval(const lhs_t &l, const rhs_t &r) { if (AbsoluteValue::eval( static_cast::type>(l).get_value()) > @@ -401,7 +401,7 @@ struct IMinOperator : public Operators { } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant_pair::value(); } @@ -409,7 +409,7 @@ struct IMinOperator : public Operators { struct CollapseIndexTupleOperator : public Operators { template - static SYCL_BLAS_INLINE + static PORTBLAS_INLINE typename ResolveReturnType::type>::type eval(const lhs_t &l, const rhs_t &r) { @@ -420,7 +420,7 @@ struct CollapseIndexTupleOperator : public Operators { } template - constexpr static SYCL_BLAS_INLINE typename rhs_t::value_t init() { + constexpr static PORTBLAS_INLINE typename rhs_t::value_t init() { return constant_pair::value()>::value(); diff --git a/src/operations/extension/reduction.hpp b/src/operations/extension/reduction.hpp index cf3d6d34a..b807d2b04 100644 --- a/src/operations/extension/reduction.hpp +++ b/src/operations/extension/reduction.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_REDUCTION_HPP -#define SYCL_BLAS_EXTENSION_REDUCTION_HPP +#ifndef PORTBLAS_EXTENSION_REDUCTION_HPP +#define PORTBLAS_EXTENSION_REDUCTION_HPP #include "blas_meta.h" #include "operations/extension/reduction.h" @@ -34,7 +34,7 @@ namespace blas { template -SYCL_BLAS_INLINE Reduction::Reduction( +PORTBLAS_INLINE Reduction::Reduction( input_t in, output_t out) : in_(in), out_(out), @@ -64,7 +64,7 @@ SYCL_BLAS_INLINE Reduction::Reduction( */ template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Reduction::valid_thread( cl::sycl::nd_item<1> id) const { return true; @@ -72,7 +72,7 @@ Reduction::valid_thread( template -SYCL_BLAS_INLINE void Reduction::bind( +PORTBLAS_INLINE void Reduction::bind( cl::sycl::handler& h) { in_.bind(h); out_.bind(h); @@ -80,7 +80,7 @@ SYCL_BLAS_INLINE void Reduction::bind( template -SYCL_BLAS_INLINE void Reduction::adjust_access_displacement() { in_.adjust_access_displacement(); out_.adjust_access_displacement(); @@ -92,7 +92,7 @@ SYCL_BLAS_INLINE void Reduction -SYCL_BLAS_INLINE cl::sycl::nd_range<1> +PORTBLAS_INLINE cl::sycl::nd_range<1> Reduction::get_nd_range( index_t compute_units) noexcept { constexpr index_t local_range = params_t::get_local_thread_size_preserve() * @@ -116,7 +116,7 @@ Reduction::get_nd_range( */ template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Reduction::reduce( index_t global_reduce_id, index_t global_preserve_id, element_t& accumulator) noexcept { @@ -152,7 +152,7 @@ Reduction::reduce( template template -SYCL_BLAS_INLINE void Reduction::eval( +PORTBLAS_INLINE void Reduction::eval( local_memory_t scratch, cl::sycl::nd_item<1> id) noexcept { const index_t local_id = id.get_local_id(0); const index_t group_id = id.get_group(0); @@ -237,4 +237,4 @@ const typename params_t::element_t } // namespace blas -#endif // SYCL_BLAS_EXTENSION_REDUCTION_HPP +#endif // PORTBLAS_EXTENSION_REDUCTION_HPP diff --git a/src/operations/extension/transpose.hpp b/src/operations/extension/transpose.hpp index 3242cfaba..1e4e2d536 100644 --- a/src/operations/extension/transpose.hpp +++ b/src/operations/extension/transpose.hpp @@ -16,14 +16,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename transpose.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_TRANSPOSE_HPP -#define SYCL_BLAS_EXTENSION_TRANSPOSE_HPP +#ifndef PORTBLAS_EXTENSION_TRANSPOSE_HPP +#define PORTBLAS_EXTENSION_TRANSPOSE_HPP #include "operations/extension/transpose.h" @@ -32,7 +32,7 @@ namespace blas { // Transpose template -SYCL_BLAS_INLINE bool +PORTBLAS_INLINE bool Transpose::valid_thread(cl::sycl::nd_item<1> item) const { // Valid threads are established by ::eval() @@ -42,7 +42,7 @@ Transpose -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::bind(cl::sycl::handler &cgh) { A_.bind(cgh); @@ -51,7 +51,7 @@ Transpose -SYCL_BLAS_INLINE typename in_t::index_t +PORTBLAS_INLINE typename in_t::index_t Transpose::get_size() const { // Smallest TileSize square-multiple containing input/output matrices @@ -60,7 +60,7 @@ Transpose -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::adjust_access_displacement() { A_.adjust_access_displacement(); @@ -79,7 +79,7 @@ Transpose -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::get_indices(cl::sycl::nd_item<1> id, index_t &in_idx, index_t &out_idx, index_t &i, index_t &j) { @@ -107,7 +107,7 @@ Transpose -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::eval(cl::sycl::nd_item<1> id) { index_t idx = id.get_global_linear_id(); @@ -145,7 +145,7 @@ Transpose -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::get_indices(cl::sycl::nd_item<1> id, index_t &in_idx, index_t &in_local_idx, index_t &out_idx, @@ -181,7 +181,7 @@ Transpose template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void Transpose::eval(local_memory_t local_mem, cl::sycl::nd_item<1> id) { value_t *local = local_mem.localAcc.get_pointer(); @@ -486,4 +486,4 @@ TransposeAdd #include @@ -44,7 +44,7 @@ struct LocalMemory { @param size Size in elements of the local accessor. @param cgh SYCL command group handler. */ - SYCL_BLAS_INLINE LocalMemory(size_t size, cl::sycl::handler &cgh) + PORTBLAS_INLINE LocalMemory(size_t size, cl::sycl::handler &cgh) : localAcc(cl::sycl::range<1>(size), cgh) {} /*! @@ -53,7 +53,7 @@ struct LocalMemory { @param id SYCL id. @return Reference to an element of the local accessor. */ - SYCL_BLAS_INLINE value_t &operator[](cl::sycl::id<1> id) { + PORTBLAS_INLINE value_t &operator[](cl::sycl::id<1> id) { return localAcc[id]; } @@ -78,7 +78,7 @@ struct LocalMemory { @param size Size in elements of the local accessor. @param cgh SYCL command group handler. */ - SYCL_BLAS_INLINE LocalMemory(size_t, cl::sycl::handler &) {} + PORTBLAS_INLINE LocalMemory(size_t, cl::sycl::handler &) {} }; /*! @brief A struct for containing a local accessor if shared memory is enabled. @@ -95,7 +95,7 @@ struct LocalMemory { @param size Size in elements of the local accessor. @param cgh SYCL command group handler. */ - SYCL_BLAS_INLINE LocalMemory(size_t size, cl::sycl::handler &cgh) + PORTBLAS_INLINE LocalMemory(size_t size, cl::sycl::handler &cgh) : subgroupAcc(cl::sycl::range<1>(size), cgh) {} /*! @@ -104,7 +104,7 @@ struct LocalMemory { @param id SYCL id. @return Reference to an element of the subgroup accessor. */ - SYCL_BLAS_INLINE value_t &operator[](cl::sycl::id<1> id) { + PORTBLAS_INLINE value_t &operator[](cl::sycl::id<1> id) { return subgroupAcc[id]; } @@ -136,7 +136,7 @@ struct ExpressionTreeEvaluator { @param scratch Shared memory object. @param index SYCL nd_item. */ - static SYCL_BLAS_INLINE void eval( + static PORTBLAS_INLINE void eval( expression_tree_t &tree, LocalMemory scratch, cl::sycl::nd_item<1> index) { @@ -161,7 +161,7 @@ struct ExpressionTreeEvaluator, cl::sycl::nd_item<1> index) { @@ -189,7 +189,7 @@ index. @param scratch subgroup memory object. @param index SYCL nd_item. */ - static SYCL_BLAS_INLINE void eval( + static PORTBLAS_INLINE void eval( expression_tree_t &tree, LocalMemory scratch, cl::sycl::nd_item<1> index) { @@ -211,10 +211,10 @@ template i) const { + PORTBLAS_INLINE void operator()(cl::sycl::nd_item<1> i) const { expression_tree_t &non_const_t = *const_cast(&t_); non_const_t.adjust_access_displacement(); ExpressionTreeEvaluator -static SYCL_BLAS_INLINE cl::sycl::event execute_tree(queue_t q_, +static PORTBLAS_INLINE cl::sycl::event execute_tree(queue_t q_, expression_tree_t t, size_t _localSize, size_t _globalSize, diff --git a/src/sb_handle/sycl_blas_handle.hpp b/src/sb_handle/portblas_handle.hpp similarity index 98% rename from src/sb_handle/sycl_blas_handle.hpp rename to src/sb_handle/portblas_handle.hpp index f8662bec8..1e4967919 100644 --- a/src/sb_handle/sycl_blas_handle.hpp +++ b/src/sb_handle/portblas_handle.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * - * @filename sycl_blas_handle.hpp + * @filename portblas_handle.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_HANDLE_HPP -#define SYCL_BLAS_HANDLE_HPP +#ifndef PORTBLAS_HANDLE_HPP +#define PORTBLAS_HANDLE_HPP #include @@ -33,8 +33,8 @@ #include "operations/blas2_trees.hpp" #include "operations/blas_operators.hpp" #include "sb_handle/kernel_constructor.h" -#include "sb_handle/sycl_blas_handle.h" -#include "sycl_blas_helper.h" +#include "sb_handle/portblas_handle.h" +#include "portblas_helper.h" #include "views/view.h" namespace blas { @@ -351,4 +351,4 @@ inline typename SB_Handle::event_t SB_Handle::execute( } // namespace blas -#endif // SYCL_BLAS_HANDLE_HPP +#endif // PORTBLAS_HANDLE_HPP diff --git a/src/views/view.hpp b/src/views/view.hpp index b00c85d83..3ce9becd6 100644 --- a/src/views/view.hpp +++ b/src/views/view.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename view.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_VIEW_HPP -#define SYCL_BLAS_VIEW_HPP +#ifndef PORTBLAS_VIEW_HPP +#define PORTBLAS_VIEW_HPP #include "views/view.h" #include @@ -412,4 +412,4 @@ _value_t &MatrixView<_value_t, _container_t, _IndexType, layout>::eval( } // namespace blas -#endif // VIEW_HPP +#endif // PORTBLAS_VIEW_HPP diff --git a/src/views/view_sycl.hpp b/src/views/view_sycl.hpp index 266b0649f..1a9278eeb 100644 --- a/src/views/view_sycl.hpp +++ b/src/views/view_sycl.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename view_sycl.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_VIEW_SYCL_HPP -#define SYCL_BLAS_VIEW_SYCL_HPP +#ifndef PORTBLAS_VIEW_SYCL_HPP +#define PORTBLAS_VIEW_SYCL_HPP #include #include @@ -77,14 +77,14 @@ struct VectorView< cl::sycl::global_ptr ptr_; // Round up the ration num / den, i.e. compute ceil(num / den) - static SYCL_BLAS_INLINE index_t round_up_ratio(index_t num, index_t den) { + static PORTBLAS_INLINE index_t round_up_ratio(index_t num, index_t den) { return (num + den - 1) / den; } // Compute the number of elements to read from data. This is useful when a // VectorView is created without an explicit size, so that only the necessary // number of threads are launched. - static SYCL_BLAS_INLINE index_t calculate_input_data_size( + static PORTBLAS_INLINE index_t calculate_input_data_size( container_t &data, index_t, increment_t stride, index_t size) noexcept { increment_t const positive_stride = stride < 0 ? -stride : stride; index_t const calc_size = round_up_ratio(data.get_count(), positive_stride); @@ -94,7 +94,7 @@ struct VectorView< /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE VectorView(container_t data, index_t disp, increment_t strd, + PORTBLAS_INLINE VectorView(container_t data, index_t disp, increment_t strd, index_t size) : data_{data}, size_(calculate_input_data_size(data, disp, strd, size)), @@ -104,86 +104,86 @@ struct VectorView< /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE VectorView(container_t data) + PORTBLAS_INLINE VectorView(container_t data) : VectorView(data, 0, 1, data_.get_size()) {} /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE VectorView(container_t data, index_t disp) + PORTBLAS_INLINE VectorView(container_t data, index_t disp) : VectorView(data, disp, 1, data_.get_size()) {} /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE VectorView(self_t &opV, index_t disp, increment_t strd, + PORTBLAS_INLINE VectorView(self_t &opV, index_t disp, increment_t strd, index_t size) : VectorView(opV.get_data(), disp, strd, size) {} /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE container_t &get_data() { return data_; } + PORTBLAS_INLINE container_t &get_data() { return data_; } /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE scalar_t *get_pointer() const { return ptr_; } + PORTBLAS_INLINE scalar_t *get_pointer() const { return ptr_; } /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE index_t get_data_size() const { return data_.get_size(); } + PORTBLAS_INLINE index_t get_data_size() const { return data_.get_size(); } /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE index_t get_size() const { return size_; } + PORTBLAS_INLINE index_t get_size() const { return size_; } /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE index_t get_access_displacement() const { return disp_; } + PORTBLAS_INLINE index_t get_access_displacement() const { return disp_; } /*! * @brief See VectorView. */ - SYCL_BLAS_INLINE increment_t get_stride() const { return stride_; } + PORTBLAS_INLINE increment_t get_stride() const { return stride_; } /**** EVALUATING ****/ template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t i) { return (stride_ == 1) ? *(ptr_ + i) : *(ptr_ + i * stride_); } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t i) const { return (stride_ == 1) ? *(ptr_ + i) : *(ptr_ + i * stride_); } - SYCL_BLAS_INLINE scalar_t &eval(cl::sycl::nd_item<1> ndItem) { + PORTBLAS_INLINE scalar_t &eval(cl::sycl::nd_item<1> ndItem) { return eval(ndItem.get_global_id(0)); } - SYCL_BLAS_INLINE const scalar_t eval(cl::sycl::nd_item<1> ndItem) const { + PORTBLAS_INLINE const scalar_t eval(cl::sycl::nd_item<1> ndItem) const { return eval(ndItem.get_global_id(0)); } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) { return *(ptr_ + indx); } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) const noexcept { return *(ptr_ + indx); } - SYCL_BLAS_INLINE void bind(cl::sycl::handler &h) { h.require(data_); } - SYCL_BLAS_INLINE void adjust_access_displacement() { + PORTBLAS_INLINE void bind(cl::sycl::handler &h) { h.require(data_); } + PORTBLAS_INLINE void adjust_access_displacement() { ptr_ = data_.get_pointer() + disp_; } }; @@ -228,7 +228,7 @@ struct MatrixView< ptr_; // global pointer access inside the kernel /**** CONSTRUCTORS ****/ - SYCL_BLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC, + PORTBLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC, index_t sizeL, index_t disp) : data_{data}, sizeR_(sizeR), @@ -237,15 +237,15 @@ struct MatrixView< inc_(1), disp_(disp) {} - SYCL_BLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC) + PORTBLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC) : MatrixView(data, sizeR, sizeC, (layout::is_col_major() ? sizeR_ : sizeC_), 0) {} - SYCL_BLAS_INLINE MatrixView(self_t opM, index_t sizeR, index_t sizeC, + PORTBLAS_INLINE MatrixView(self_t opM, index_t sizeR, index_t sizeC, index_t sizeL, index_t disp) : MatrixView(opM.data_, sizeR, sizeC, sizeL, disp) {} - SYCL_BLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC, + PORTBLAS_INLINE MatrixView(container_t data, index_t sizeR, index_t sizeC, index_t sizeL, index_t inc, index_t disp) : data_{data}, sizeR_(sizeR), @@ -255,25 +255,25 @@ struct MatrixView< disp_(disp) {} /**** RETRIEVING DATA ****/ - SYCL_BLAS_INLINE container_t &get_data() { return data_; } + PORTBLAS_INLINE container_t &get_data() { return data_; } - SYCL_BLAS_INLINE const index_t get_size() const { return sizeR_ * sizeC_; } + PORTBLAS_INLINE const index_t get_size() const { return sizeR_ * sizeC_; } - SYCL_BLAS_INLINE index_t get_data_size() const { return data_.get_size(); } + PORTBLAS_INLINE index_t get_data_size() const { return data_.get_size(); } - SYCL_BLAS_INLINE const index_t getSizeL() const { return sizeL_; } + PORTBLAS_INLINE const index_t getSizeL() const { return sizeL_; } - SYCL_BLAS_INLINE const index_t get_size_row() const { return sizeR_; } + PORTBLAS_INLINE const index_t get_size_row() const { return sizeR_; } - SYCL_BLAS_INLINE const index_t get_size_col() const { return sizeC_; } + PORTBLAS_INLINE const index_t get_size_col() const { return sizeC_; } - SYCL_BLAS_INLINE index_t get_access_displacement() const { return disp_; } + PORTBLAS_INLINE index_t get_access_displacement() const { return disp_; } - SYCL_BLAS_INLINE scalar_t *get_pointer() const { return ptr_; } + PORTBLAS_INLINE scalar_t *get_pointer() const { return ptr_; } /**** EVALUATING ***/ - SYCL_BLAS_INLINE scalar_t &eval(index_t i, index_t j) { + PORTBLAS_INLINE scalar_t &eval(index_t i, index_t j) { if constexpr (is_inc) { return ((layout::is_col_major()) ? *(ptr_ + i * inc_ + sizeL_ * j) : *(ptr_ + j * inc_ + sizeL_ * i)); @@ -283,7 +283,7 @@ struct MatrixView< } } - SYCL_BLAS_INLINE scalar_t &eval(index_t i, index_t j) const noexcept { + PORTBLAS_INLINE scalar_t &eval(index_t i, index_t j) const noexcept { if constexpr (is_inc) { return ((layout::is_col_major()) ? *(ptr_ + i * inc_ + sizeL_ * j) : *(ptr_ + j * inc_ + sizeL_ * i)); @@ -294,7 +294,7 @@ struct MatrixView< } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) { const index_t j = indx / sizeR_; const index_t i = indx - sizeR_ * j; @@ -302,36 +302,36 @@ struct MatrixView< } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) const noexcept { const index_t j = indx / sizeR_; const index_t i = indx - sizeR_ * j; return eval(i, j); } - SYCL_BLAS_INLINE scalar_t &eval(cl::sycl::nd_item<1> ndItem) { + PORTBLAS_INLINE scalar_t &eval(cl::sycl::nd_item<1> ndItem) { return eval(ndItem.get_global_id(0)); } - SYCL_BLAS_INLINE scalar_t eval(cl::sycl::nd_item<1> ndItem) const noexcept { + PORTBLAS_INLINE scalar_t eval(cl::sycl::nd_item<1> ndItem) const noexcept { return eval(ndItem.get_global_id(0)); } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) { return *(ptr_ + indx); } template - SYCL_BLAS_INLINE typename std::enable_if::type eval( + PORTBLAS_INLINE typename std::enable_if::type eval( index_t indx) const noexcept { return *(ptr_ + indx); } - SYCL_BLAS_INLINE void bind(cl::sycl::handler &h) { h.require(data_); } + PORTBLAS_INLINE void bind(cl::sycl::handler &h) { h.require(data_); } - SYCL_BLAS_INLINE void adjust_access_displacement() { + PORTBLAS_INLINE void adjust_access_displacement() { ptr_ = data_.get_pointer() + disp_; } }; diff --git a/test/CMakeLists.txt b/test/CMakeLists.txt index 7daafb57a..4c78e25f9 100644 --- a/test/CMakeLists.txt +++ b/test/CMakeLists.txt @@ -4,7 +4,7 @@ find_package(Clara REQUIRED) include(ExternalProject) find_package(SystemBLAS REQUIRED) -set(SYCLBLAS_TEST ${CMAKE_CURRENT_SOURCE_DIR}) +set(PORTBLAS_TEST ${CMAKE_CURRENT_SOURCE_DIR}) set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-deprecated-declarations -Wno-shorten-64-to-32 -Wno-cast-align") set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-deprecated-copy-with-user-provided-copy -Wno-unused-variable") @@ -36,7 +36,7 @@ add_subdirectory(${CMAKE_CURRENT_BINARY_DIR}/googletest-src ${CMAKE_CURRENT_BINARY_DIR}/googletest-build EXCLUDE_FROM_ALL) -# Set SYCL-BLAS options +# Set portBLAS options if(CMAKE_CROSSCOMPILING) set(cmake_toolchain "-DCMAKE_TOOLCHAIN_FILE=${CMAKE_TOOLCHAIN_FILE}" diff --git a/test/blas_test.hpp b/test/blas_test.hpp index fdb43b31d..2c292e1b7 100644 --- a/test/blas_test.hpp +++ b/test/blas_test.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_test.hpp * @@ -38,7 +38,7 @@ #include -#include +#include #include #include @@ -55,7 +55,7 @@ extern Args args; using namespace blas; -// The sycl blas handle type used in tests +// The portBLAS handle type used in tests using index_t = BLAS_INDEX_T; diff --git a/test/blas_test_macros.hpp b/test/blas_test_macros.hpp index 844512ff5..5b4cf979c 100644 --- a/test/blas_test_macros.hpp +++ b/test/blas_test_macros.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas_test_macros.hpp * diff --git a/test/exprtest/CMakeLists.txt b/test/exprtest/CMakeLists.txt index 5d494c719..20cf0cc5e 100644 --- a/test/exprtest/CMakeLists.txt +++ b/test/exprtest/CMakeLists.txt @@ -17,23 +17,23 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * # **************************************************************************/ -set(SYCLBLAS_EXPRTEST ${CMAKE_CURRENT_SOURCE_DIR}) +set(PORTBLAS_EXPRTEST ${CMAKE_CURRENT_SOURCE_DIR}) # compiling tests set(SYCL_EXPRTEST_SRCS - ${SYCLBLAS_EXPRTEST}/blas1_scal_asum_test.cpp - ${SYCLBLAS_EXPRTEST}/blas1_axpy_copy_test.cpp + ${PORTBLAS_EXPRTEST}/blas1_scal_asum_test.cpp + ${PORTBLAS_EXPRTEST}/blas1_axpy_copy_test.cpp ) # Temporary disabling the following tests for Intel DPC++ as currently Intel compiler crashes while running the following tests # https://github.com/intel/llvm/issues/7075 if(is_computecpp) - list(APPEND SYCL_EXPRTEST_SRCS "${SYCLBLAS_EXPRTEST}/collapse_nested_tuple.cpp") + list(APPEND SYCL_EXPRTEST_SRCS "${PORTBLAS_EXPRTEST}/collapse_nested_tuple.cpp") endif() foreach(blas_test ${SYCL_EXPRTEST_SRCS}) @@ -42,10 +42,10 @@ foreach(blas_test ${SYCL_EXPRTEST_SRCS}) if(is_computecpp) set_property(TARGET ${test_exec} PROPERTY CXX_STANDARD 14) endif() - # -DTUNING_TARGET is needed when using SYCL-BLAS in header only mode. + # -DTUNING_TARGET is needed when using portBLAS in header only mode. target_compile_definitions(${test_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_TEST_INDEX_TYPE} -D${TUNING_TARGET}) - target_link_libraries(${test_exec} PRIVATE gtest_main blas::blas sycl_blas) - target_include_directories(${test_exec} PRIVATE "${CMAKE_CURRENT_SOURCE_DIR}/.." ${CBLAS_INCLUDE} ${SYCLBLAS_SRC} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_link_libraries(${test_exec} PRIVATE gtest_main blas::blas portblas) + target_include_directories(${test_exec} PRIVATE "${CMAKE_CURRENT_SOURCE_DIR}/.." ${CBLAS_INCLUDE} ${PORTBLAS_SRC} ${PORTBLAS_COMMON_INCLUDE_DIR}) add_sycl_to_target( TARGET ${test_exec} SOURCES ${blas_test} diff --git a/test/exprtest/blas1_axpy_copy_test.cpp b/test/exprtest/blas1_axpy_copy_test.cpp index 839a925cb..0f991d40a 100644 --- a/test/exprtest/blas1_axpy_copy_test.cpp +++ b/test/exprtest/blas1_axpy_copy_test.cpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_axpy_copy_test.cpp * **************************************************************************/ #include "blas_test.hpp" -#include "sycl_blas.hpp" +#include "portblas.hpp" // inputs combination template @@ -59,7 +59,7 @@ void run_test(const combination_t combi) { reference_blas::copy(y_dim, v_y.data(), 1, v_cpu_y.data(), 1); reference_blas::axpy(size, alpha, v_xcopy.data(), incX, v_cpu_y.data(), incY); - // SYCL-BLAS implementation + // portBLAS implementation auto q = make_queue(); blas::SB_Handle sb_handle(q); diff --git a/test/exprtest/blas1_iface_test.cpp b/test/exprtest/blas1_iface_test.cpp index 5caedd4fc..90e7b5a3e 100644 --- a/test/exprtest/blas1_iface_test.cpp +++ b/test/exprtest/blas1_iface_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_iface_test.cpp * diff --git a/test/exprtest/blas1_scal_asum_test.cpp b/test/exprtest/blas1_scal_asum_test.cpp index e631739d6..204f8ae95 100644 --- a/test/exprtest/blas1_scal_asum_test.cpp +++ b/test/exprtest/blas1_scal_asum_test.cpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_scal_asum_test.cpp * * **************************************************************************/ #include "blas_test.hpp" -#include "sycl_blas.hpp" +#include "portblas.hpp" // inputs combination template @@ -53,7 +53,7 @@ void run_test(const combination_t combi) { reference_blas::scal(size, alpha, v_cpu_x.data(), incX); cpu_y = reference_blas::asum(size, v_cpu_x.data(), incX); - // SYCL-BLAS implementation + // portBLAS implementation auto q = make_queue(); blas::SB_Handle sb_handle(q); diff --git a/test/exprtest/collapse_nested_tuple.cpp b/test/exprtest/collapse_nested_tuple.cpp index 46a076fbe..4053042f8 100644 --- a/test/exprtest/collapse_nested_tuple.cpp +++ b/test/exprtest/collapse_nested_tuple.cpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename collapse_nested_tuple.cpp * **************************************************************************/ #include "blas_test.hpp" -#include "sycl_blas.hpp" +#include "portblas.hpp" // inputs combination template diff --git a/test/exprtest/main.cpp b/test/exprtest/main.cpp index 1b36b26ab..c8fe633f7 100644 --- a/test/exprtest/main.cpp +++ b/test/exprtest/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/test/unittest/CMakeLists.txt b/test/unittest/CMakeLists.txt index ff566fcf3..9022d141b 100644 --- a/test/unittest/CMakeLists.txt +++ b/test/unittest/CMakeLists.txt @@ -17,62 +17,62 @@ # * See the License for the specific language governing permissions and # * limitations under the License. # * -# * SYCL-BLAS: BLAS implementation using SYCL +# * portBLAS: BLAS implementation using SYCL # * # * @filename CMakeLists.txt # * # **************************************************************************/ -set(SYCLBLAS_UNITTEST ${CMAKE_CURRENT_SOURCE_DIR}) +set(PORTBLAS_UNITTEST ${CMAKE_CURRENT_SOURCE_DIR}) -include_directories(${SYCLBLAS_TEST} ${BLAS_INCLUDE_DIRS}) +include_directories(${PORTBLAS_TEST} ${BLAS_INCLUDE_DIRS}) # compiling tests set(SYCL_UNITTEST_SRCS # Blas 1 tests - ${SYCLBLAS_UNITTEST}/blas1/blas1_axpy_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_copy_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_scal_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_rot_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_rotm_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_rotmg_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_rotg_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_sdsdot_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_axpy_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_copy_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_scal_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rot_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rotm_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rotmg_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rotg_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_sdsdot_test.cpp # Blas 2 tests - ${SYCLBLAS_UNITTEST}/blas2/blas2_gbmv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_gemv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_ger_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_sbmv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_spmv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_syr_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_spr_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_spr2_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_syr2_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_symv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_tpmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_gbmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_gemv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_ger_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_sbmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_spmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_syr_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_spr_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_spr2_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_syr2_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_symv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_tpmv_test.cpp # Blas 3 tests - ${SYCLBLAS_UNITTEST}/blas3/blas3_gemm_test.cpp - ${SYCLBLAS_UNITTEST}/blas3/blas3_gemm_batched_test.cpp - ${SYCLBLAS_UNITTEST}/blas3/blas3_trsm_test.cpp - ${SYCLBLAS_UNITTEST}/blas3/blas3_symm_test.cpp + ${PORTBLAS_UNITTEST}/blas3/blas3_gemm_test.cpp + ${PORTBLAS_UNITTEST}/blas3/blas3_gemm_batched_test.cpp + ${PORTBLAS_UNITTEST}/blas3/blas3_trsm_test.cpp + ${PORTBLAS_UNITTEST}/blas3/blas3_symm_test.cpp # Blas extension - ${SYCLBLAS_UNITTEST}/extension/omatcopy_test.cpp - ${SYCLBLAS_UNITTEST}/extension/omatcopy2_test.cpp - ${SYCLBLAS_UNITTEST}/extension/omatadd_test.cpp + ${PORTBLAS_UNITTEST}/extension/omatcopy_test.cpp + ${PORTBLAS_UNITTEST}/extension/omatcopy2_test.cpp + ${PORTBLAS_UNITTEST}/extension/omatadd_test.cpp ) if(${BLAS_ENABLE_EXTENSIONS}) - list(APPEND SYCL_UNITTEST_SRCS ${SYCLBLAS_UNITTEST}/extension/transpose_test.cpp - ${SYCLBLAS_UNITTEST}/extension/reduction_test.cpp + list(APPEND SYCL_UNITTEST_SRCS ${PORTBLAS_UNITTEST}/extension/transpose_test.cpp + ${PORTBLAS_UNITTEST}/extension/reduction_test.cpp ) endif() # Enable testing of the sycl 2020 routines just for Intel DPC++ if(is_dpcpp) set(SYCL_UNITTEST_SRCS ${SYCL_UNITTEST_SRCS} - ${SYCLBLAS_UNITTEST}/blas2/blas2_trsv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_tbsv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_tpsv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_trsv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_tbsv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_tpsv_test.cpp ) endif() @@ -80,29 +80,29 @@ endif() if(is_computecpp) set(SYCL_UNITTEST_SRCS ${SYCL_UNITTEST_SRCS} # Blas 1 tests - ${SYCLBLAS_UNITTEST}/blas1/blas1_swap_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_asum_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_dot_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_nrm2_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_iamax_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_iamin_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_swap_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_asum_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_dot_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_nrm2_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_iamax_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_iamin_test.cpp # Blas 2 tests - ${SYCLBLAS_UNITTEST}/blas2/blas2_trmv_test.cpp - ${SYCLBLAS_UNITTEST}/blas2/blas2_tbmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_trmv_test.cpp + ${PORTBLAS_UNITTEST}/blas2/blas2_tbmv_test.cpp # Blas buffer tests - ${SYCLBLAS_UNITTEST}/buffers/sycl_buffer_test.cpp + ${PORTBLAS_UNITTEST}/buffers/sycl_buffer_test.cpp ) endif() # Contains tests that fail if compiled with -ffast-math set(SYCL_UNITTEST_NOFASTMATH - ${SYCLBLAS_UNITTEST}/blas1/blas1_rotg_test.cpp - ${SYCLBLAS_UNITTEST}/blas1/blas1_rotmg_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rotg_test.cpp + ${PORTBLAS_UNITTEST}/blas1/blas1_rotmg_test.cpp ) if(GEMM_TALL_SKINNY_SUPPORT) - list(APPEND SYCL_UNITTEST_SRCS ${SYCLBLAS_UNITTEST}/blas3/blas3_gemm_tall_skinny_test.cpp) + list(APPEND SYCL_UNITTEST_SRCS ${PORTBLAS_UNITTEST}/blas3/blas3_gemm_tall_skinny_test.cpp) endif() foreach(blas_test ${SYCL_UNITTEST_SRCS}) @@ -115,8 +115,8 @@ foreach(blas_test ${SYCL_UNITTEST_SRCS}) target_compile_definitions(${test_exec} PRIVATE STRESS_TESTING) endif() target_compile_definitions(${test_exec} PRIVATE -DBLAS_INDEX_T=${BLAS_TEST_INDEX_TYPE}) - target_link_libraries(${test_exec} PRIVATE gtest_main Clara::Clara blas::blas sycl_blas) - target_include_directories(${test_exec} PRIVATE ${CBLAS_INCLUDE} ${SYCLBLAS_COMMON_INCLUDE_DIR}) + target_link_libraries(${test_exec} PRIVATE gtest_main Clara::Clara blas::blas portblas) + target_include_directories(${test_exec} PRIVATE ${CBLAS_INCLUDE} ${PORTBLAS_COMMON_INCLUDE_DIR}) list (FIND SYCL_UNITTEST_NOFASTMATH ${blas_test} _index) if (${_index} GREATER -1) diff --git a/test/unittest/blas1/blas1_asum_test.cpp b/test/unittest/blas1/blas1_asum_test.cpp index 2c64630f5..5fc3d74ac 100644 --- a/test/unittest/blas1/blas1_asum_test.cpp +++ b/test/unittest/blas1/blas1_asum_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_asum_test.cpp * diff --git a/test/unittest/blas1/blas1_axpy_test.cpp b/test/unittest/blas1/blas1_axpy_test.cpp index 1df8671ec..bde6bf676 100644 --- a/test/unittest/blas1/blas1_axpy_test.cpp +++ b/test/unittest/blas1/blas1_axpy_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_axpy_test.cpp * diff --git a/test/unittest/blas1/blas1_copy_test.cpp b/test/unittest/blas1/blas1_copy_test.cpp index caa68a5d9..d5b73fc1b 100644 --- a/test/unittest/blas1/blas1_copy_test.cpp +++ b/test/unittest/blas1/blas1_copy_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_copy_test.cpp * diff --git a/test/unittest/blas1/blas1_dot_test.cpp b/test/unittest/blas1/blas1_dot_test.cpp index deb1a40c7..9bea0738d 100644 --- a/test/unittest/blas1/blas1_dot_test.cpp +++ b/test/unittest/blas1/blas1_dot_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_dot_test.cpp * diff --git a/test/unittest/blas1/blas1_iamin_test.cpp b/test/unittest/blas1/blas1_iamin_test.cpp index df71a66d9..1d01fe251 100644 --- a/test/unittest/blas1/blas1_iamin_test.cpp +++ b/test/unittest/blas1/blas1_iamin_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_iamin_test.cpp * diff --git a/test/unittest/blas1/blas1_iaminmax_common.hpp b/test/unittest/blas1/blas1_iaminmax_common.hpp index 3c28c4059..09b858dc8 100644 --- a/test/unittest/blas1/blas1_iaminmax_common.hpp +++ b/test/unittest/blas1/blas1_iaminmax_common.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_iaminmax_common.hpp * diff --git a/test/unittest/blas1/blas1_nrm2_test.cpp b/test/unittest/blas1/blas1_nrm2_test.cpp index efa83fcae..d422424e9 100644 --- a/test/unittest/blas1/blas1_nrm2_test.cpp +++ b/test/unittest/blas1/blas1_nrm2_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_nrm2_test.cpp * diff --git a/test/unittest/blas1/blas1_rot_test.cpp b/test/unittest/blas1/blas1_rot_test.cpp index b33b701d7..58e5fb84c 100644 --- a/test/unittest/blas1/blas1_rot_test.cpp +++ b/test/unittest/blas1/blas1_rot_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_rot_test.cpp * diff --git a/test/unittest/blas1/blas1_rotg_test.cpp b/test/unittest/blas1/blas1_rotg_test.cpp index 0fb5c18dc..00742a0b4 100644 --- a/test/unittest/blas1/blas1_rotg_test.cpp +++ b/test/unittest/blas1/blas1_rotg_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_rotg_test.cpp * diff --git a/test/unittest/blas1/blas1_rotm_test.cpp b/test/unittest/blas1/blas1_rotm_test.cpp index f6eeffcb2..0a5de30f7 100644 --- a/test/unittest/blas1/blas1_rotm_test.cpp +++ b/test/unittest/blas1/blas1_rotm_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_rotm_test.cpp * diff --git a/test/unittest/blas1/blas1_rotmg_test.cpp b/test/unittest/blas1/blas1_rotmg_test.cpp index 0d710f101..f8a014fd1 100644 --- a/test/unittest/blas1/blas1_rotmg_test.cpp +++ b/test/unittest/blas1/blas1_rotmg_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_rotmg_test.cpp * @@ -52,13 +52,13 @@ struct RotmgTest { RotmgTest(scalar_t d1, scalar_t d2, scalar_t x1, scalar_t y1) : input{d1, d2, x1, y1} {} - void run_sycl_blas_rotmg(); + void run_portblas_rotmg(); void validate_with_reference(); void validate_with_rotm(); }; template -void RotmgTest::run_sycl_blas_rotmg() { +void RotmgTest::run_portblas_rotmg() { auto q = make_queue(); blas::SB_Handle sb_handle(q); @@ -202,7 +202,7 @@ void run_test(const combination_t combi) { std::tie(d1_input, d2_input, x1_input, y1_input, will_overflow) = combi; RotmgTest test{d1_input, d2_input, x1_input, y1_input}; - test.run_sycl_blas_rotmg(); + test.run_portblas_rotmg(); /* Do not test with things that might overflow or underflow. Results will not * make sense if that happens */ diff --git a/test/unittest/blas1/blas1_scal_test.cpp b/test/unittest/blas1/blas1_scal_test.cpp index b648ae20d..f08d8d590 100644 --- a/test/unittest/blas1/blas1_scal_test.cpp +++ b/test/unittest/blas1/blas1_scal_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_scal_test.cpp * diff --git a/test/unittest/blas1/blas1_sdsdot_test.cpp b/test/unittest/blas1/blas1_sdsdot_test.cpp index 8b676d8ab..283c9bf41 100644 --- a/test/unittest/blas1/blas1_sdsdot_test.cpp +++ b/test/unittest/blas1/blas1_sdsdot_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_sdsdot_test.cpp * diff --git a/test/unittest/blas1/blas1_swap_test.cpp b/test/unittest/blas1/blas1_swap_test.cpp index 1cb0ab463..99abde18f 100644 --- a/test/unittest/blas1/blas1_swap_test.cpp +++ b/test/unittest/blas1/blas1_swap_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas1_swap_test.cpp * diff --git a/test/unittest/blas2/blas2_gbmv_test.cpp b/test/unittest/blas2/blas2_gbmv_test.cpp index a6886c52b..d9bbed370 100644 --- a/test/unittest/blas2/blas2_gbmv_test.cpp +++ b/test/unittest/blas2/blas2_gbmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_gbmv_test.cpp * diff --git a/test/unittest/blas2/blas2_gemv_test.cpp b/test/unittest/blas2/blas2_gemv_test.cpp index 75c2a9853..c710b6e3a 100644 --- a/test/unittest/blas2/blas2_gemv_test.cpp +++ b/test/unittest/blas2/blas2_gemv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_gemv_test.cpp * diff --git a/test/unittest/blas2/blas2_ger_test.cpp b/test/unittest/blas2/blas2_ger_test.cpp index bcef2fa3d..46d2c5615 100644 --- a/test/unittest/blas2/blas2_ger_test.cpp +++ b/test/unittest/blas2/blas2_ger_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_ger_test.cpp * diff --git a/test/unittest/blas2/blas2_sbmv_test.cpp b/test/unittest/blas2/blas2_sbmv_test.cpp index 5988662d2..b1f0201f4 100644 --- a/test/unittest/blas2/blas2_sbmv_test.cpp +++ b/test/unittest/blas2/blas2_sbmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_sbmv_test.cpp * diff --git a/test/unittest/blas2/blas2_spmv_test.cpp b/test/unittest/blas2/blas2_spmv_test.cpp index 79d383ee5..353547835 100644 --- a/test/unittest/blas2/blas2_spmv_test.cpp +++ b/test/unittest/blas2/blas2_spmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_spmv_test.cpp * diff --git a/test/unittest/blas2/blas2_spr2_test.cpp b/test/unittest/blas2/blas2_spr2_test.cpp index 0fe5996e9..85edb0982 100644 --- a/test/unittest/blas2/blas2_spr2_test.cpp +++ b/test/unittest/blas2/blas2_spr2_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_spr2_test.cpp * diff --git a/test/unittest/blas2/blas2_spr_test.cpp b/test/unittest/blas2/blas2_spr_test.cpp index 94177b2b2..4cd0aa242 100644 --- a/test/unittest/blas2/blas2_spr_test.cpp +++ b/test/unittest/blas2/blas2_spr_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_spr_test.cpp * diff --git a/test/unittest/blas2/blas2_symv_test.cpp b/test/unittest/blas2/blas2_symv_test.cpp index f59da57ca..7524d9b0d 100644 --- a/test/unittest/blas2/blas2_symv_test.cpp +++ b/test/unittest/blas2/blas2_symv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_symv_test.cpp * diff --git a/test/unittest/blas2/blas2_syr2_test.cpp b/test/unittest/blas2/blas2_syr2_test.cpp index d08e80f36..ee627eb22 100644 --- a/test/unittest/blas2/blas2_syr2_test.cpp +++ b/test/unittest/blas2/blas2_syr2_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_syr2_test.cpp * diff --git a/test/unittest/blas2/blas2_syr_test.cpp b/test/unittest/blas2/blas2_syr_test.cpp index 3bb9f01b6..e239b0fc5 100644 --- a/test/unittest/blas2/blas2_syr_test.cpp +++ b/test/unittest/blas2/blas2_syr_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_syr_test.cpp * diff --git a/test/unittest/blas2/blas2_tbmv_test.cpp b/test/unittest/blas2/blas2_tbmv_test.cpp index e6808f773..e1cbc99f8 100644 --- a/test/unittest/blas2/blas2_tbmv_test.cpp +++ b/test/unittest/blas2/blas2_tbmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_tbmv_test.cpp * diff --git a/test/unittest/blas2/blas2_tbsv_test.cpp b/test/unittest/blas2/blas2_tbsv_test.cpp index e9c0536c7..aa85afa97 100644 --- a/test/unittest/blas2/blas2_tbsv_test.cpp +++ b/test/unittest/blas2/blas2_tbsv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_tbsv_test.cpp * diff --git a/test/unittest/blas2/blas2_tpmv_test.cpp b/test/unittest/blas2/blas2_tpmv_test.cpp index 28ee9bbd0..892c1b1b6 100644 --- a/test/unittest/blas2/blas2_tpmv_test.cpp +++ b/test/unittest/blas2/blas2_tpmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_tpmv_test.cpp * diff --git a/test/unittest/blas2/blas2_tpsv_test.cpp b/test/unittest/blas2/blas2_tpsv_test.cpp index 5797f163a..e1ae91ab4 100644 --- a/test/unittest/blas2/blas2_tpsv_test.cpp +++ b/test/unittest/blas2/blas2_tpsv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_tpsv_test.cpp * diff --git a/test/unittest/blas2/blas2_trmv_test.cpp b/test/unittest/blas2/blas2_trmv_test.cpp index 529cb177d..e0324cc57 100644 --- a/test/unittest/blas2/blas2_trmv_test.cpp +++ b/test/unittest/blas2/blas2_trmv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_trmv_test.cpp * diff --git a/test/unittest/blas2/blas2_trsv_test.cpp b/test/unittest/blas2/blas2_trsv_test.cpp index 326163bf1..8ada7505e 100644 --- a/test/unittest/blas2/blas2_trsv_test.cpp +++ b/test/unittest/blas2/blas2_trsv_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas2_trsv_test.cpp * diff --git a/test/unittest/blas3/blas3_gemm_batched_test.cpp b/test/unittest/blas3/blas3_gemm_batched_test.cpp index f38922025..c9ed81c60 100644 --- a/test/unittest/blas3/blas3_gemm_batched_test.cpp +++ b/test/unittest/blas3/blas3_gemm_batched_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_gemm_batched_test.cpp * diff --git a/test/unittest/blas3/blas3_gemm_common.hpp b/test/unittest/blas3/blas3_gemm_common.hpp index 1eca8bfec..c921c96b7 100644 --- a/test/unittest/blas3/blas3_gemm_common.hpp +++ b/test/unittest/blas3/blas3_gemm_common.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_gemm_common.hpp * @@ -142,7 +142,7 @@ inline void verify_gemm(const gemm_arguments_t arguments) { blas::helper::copy_to_device(sb_handle.get_queue(), c_m_gpu.data(), m_c_gpu, buffer_size_c); - // SYCL BLAS GEMM implementation + // portBLAS GEMM implementation if (batch == 1) { _gemm(sb_handle, transa, transb, m, n, k, alpha, m_a_gpu + offset, lda, m_b_gpu + offset, ldb, beta, m_c_gpu + offset, ldc); @@ -255,7 +255,7 @@ inline void verify_gemm( blas::helper::copy_to_device(sb_handle.get_queue(), c_m_gpu.data(), m_c_gpu, buffer_size_c); - // SYCL BLAS GEMM STRIDED BATCHED implementation + // portBLAS GEMM STRIDED BATCHED implementation _gemm_strided_batched(sb_handle, transa, transb, m, n, k, alpha, m_a_gpu + offset, lda, stride_a, m_b_gpu + offset, ldb, stride_b, beta, m_c_gpu + offset, ldc, stride_c, batch); diff --git a/test/unittest/blas3/blas3_gemm_tall_skinny_test.cpp b/test/unittest/blas3/blas3_gemm_tall_skinny_test.cpp index c717f05c2..d4359f1b9 100644 --- a/test/unittest/blas3/blas3_gemm_tall_skinny_test.cpp +++ b/test/unittest/blas3/blas3_gemm_tall_skinny_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_gemm_tall_skinny_test.cpp * diff --git a/test/unittest/blas3/blas3_gemm_test.cpp b/test/unittest/blas3/blas3_gemm_test.cpp index e7c14c9de..36d4a55e3 100644 --- a/test/unittest/blas3/blas3_gemm_test.cpp +++ b/test/unittest/blas3/blas3_gemm_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_gemm_test.cpp * diff --git a/test/unittest/blas3/blas3_symm_test.cpp b/test/unittest/blas3/blas3_symm_test.cpp index 136ff3bf6..856389f40 100644 --- a/test/unittest/blas3/blas3_symm_test.cpp +++ b/test/unittest/blas3/blas3_symm_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename blas3_symm_test.cpp * @@ -85,7 +85,7 @@ inline void verify_symm(const symm_arguments_t arguments) { blas::helper::copy_to_device(sb_handle.get_queue(), c_m_gpu.data(), m_c_gpu, size_c); - // SYCL BLAS SYMM implementation + // portBLAS SYMM implementation _symm(sb_handle, side, uplo, m, n, alpha, m_a_gpu, lda, m_b_gpu, ldb, beta, m_c_gpu, ldc); diff --git a/test/unittest/blas3/blas3_trsm_test.cpp b/test/unittest/blas3/blas3_trsm_test.cpp index ceb38cd00..e3ccc7102 100644 --- a/test/unittest/blas3/blas3_trsm_test.cpp +++ b/test/unittest/blas3/blas3_trsm_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * **************************************************************************/ diff --git a/test/unittest/buffers/sycl_buffer_test.cpp b/test/unittest/buffers/sycl_buffer_test.cpp index d3f78bf69..5fc333307 100644 --- a/test/unittest/buffers/sycl_buffer_test.cpp +++ b/test/unittest/buffers/sycl_buffer_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename sycl_buffer_test.cpp * diff --git a/test/unittest/extension/extension_reference.hpp b/test/unittest/extension/extension_reference.hpp index f6a80f85b..ffa9bad0d 100644 --- a/test/unittest/extension/extension_reference.hpp +++ b/test/unittest/extension/extension_reference.hpp @@ -17,13 +17,13 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename extension_reference.hpp * **************************************************************************/ -#ifndef SYCL_BLAS_EXTENSION_REFERENCE_IMPLEMENTATION_HPP -#define SYCL_BLAS_EXTENSION_REFERENCE_IMPLEMENTATION_HPP +#ifndef PORTBLAS_EXTENSION_REFERENCE_IMPLEMENTATION_HPP +#define PORTBLAS_EXTENSION_REFERENCE_IMPLEMENTATION_HPP #include diff --git a/test/unittest/extension/omatcopy2_test.cpp b/test/unittest/extension/omatcopy2_test.cpp index f3ed06f8b..46584efaf 100644 --- a/test/unittest/extension/omatcopy2_test.cpp +++ b/test/unittest/extension/omatcopy2_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy2_test.cpp * diff --git a/test/unittest/extension/omatcopy_test.cpp b/test/unittest/extension/omatcopy_test.cpp index 0ed4f6559..77d8307e4 100644 --- a/test/unittest/extension/omatcopy_test.cpp +++ b/test/unittest/extension/omatcopy_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatcopy_test.cpp * diff --git a/test/unittest/extension/reduction_test.cpp b/test/unittest/extension/reduction_test.cpp index f2cfcbd7d..2531eed94 100644 --- a/test/unittest/extension/reduction_test.cpp +++ b/test/unittest/extension/reduction_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reduction_test.cpp * diff --git a/test/unittest/extension/transpose_test.cpp b/test/unittest/extension/transpose_test.cpp index 829045af6..6d89cf215 100644 --- a/test/unittest/extension/transpose_test.cpp +++ b/test/unittest/extension/transpose_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename transpose.cpp * diff --git a/test/unittest/main.cpp b/test/unittest/main.cpp index cd102f803..02c7ab414 100644 --- a/test/unittest/main.cpp +++ b/test/unittest/main.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename main.cpp * diff --git a/tools/auto_tuner/CMakeLists.txt b/tools/auto_tuner/CMakeLists.txt index a3e895ae5..57f6a5b53 100644 --- a/tools/auto_tuner/CMakeLists.txt +++ b/tools/auto_tuner/CMakeLists.txt @@ -1,6 +1,6 @@ cmake_minimum_required(VERSION 3.7) -set(SYCLBLAS_INCLUDE ${CMAKE_CURRENT_SOURCE_DIR}/../../include) -set(SYCLBLAS_SRC ${CMAKE_CURRENT_SOURCE_DIR}/../../src) +set(PORTBLAS_INCLUDE ${CMAKE_CURRENT_SOURCE_DIR}/../../include) +set(PORTBLAS_SRC ${CMAKE_CURRENT_SOURCE_DIR}/../../src) list(APPEND CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/../../cmake/ @@ -11,7 +11,7 @@ set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-deprecated-copy-with-user-provided # Setup datatypes, workgroup sizes and other options. # NB: This has to be included before CmakeFunctionHelper as it declares various options. -include(ConfigureSYCLBLAS) +include(ConfigurePORTBLAS) include(SYCL) find_package(PythonInterp 3 REQUIRED) find_package(SystemBLAS REQUIRED) @@ -77,8 +77,8 @@ add_custom_target(tuner_generate_kernels DEPENDS ${tuner_kernel_srcs}) add_library(tuner_kernel_lib ${tuner_kernel_srcs}) target_include_directories(tuner_kernel_lib PRIVATE - ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_SRC} + ${PORTBLAS_INCLUDE} + ${PORTBLAS_SRC} include/ ${CMAKE_CURRENT_BINARY_DIR} ) @@ -106,8 +106,8 @@ foreach(blas_tuner ${SYCL_AUTO_TUNNER_SRCS}) add_executable(${tuner_exec} ${blas_tuner}) target_link_libraries(${tuner_exec} PRIVATE blas::blas tuner_kernel_lib) target_include_directories(${tuner_exec} PRIVATE - ${SYCLBLAS_INCLUDE} - ${SYCLBLAS_SRC} + ${PORTBLAS_INCLUDE} + ${PORTBLAS_SRC} include/ ${CMAKE_CURRENT_BINARY_DIR} ) diff --git a/tools/auto_tuner/README.md b/tools/auto_tuner/README.md index d96453f33..35633ec88 100644 --- a/tools/auto_tuner/README.md +++ b/tools/auto_tuner/README.md @@ -10,7 +10,7 @@ potential configurations and then prints a list of them and their performance. Building -------- -1. Clone the SYCL-BLAS repository, making sure to pass the `--recursive` option, in order to clone submodule(s). +1. Clone the portBLAS repository, making sure to pass the `--recursive` option, in order to clone submodule(s). 2. Create a build directory as `tools/auto_tuner/build`. 3. Run `CMake` and `Ninja` from the build directory: diff --git a/tools/auto_tuner/include/gemm_tuner.hpp b/tools/auto_tuner/include/gemm_tuner.hpp index 1fabcf26f..b3e67b702 100644 --- a/tools/auto_tuner/include/gemm_tuner.hpp +++ b/tools/auto_tuner/include/gemm_tuner.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename gemm_tuner.hpp * @@ -28,7 +28,7 @@ #include "utils.hpp" #include "reference_gemm.hpp" -#include "sycl_blas.hpp" +#include "portblas.hpp" using namespace cl::sycl; using namespace blas; @@ -67,11 +67,11 @@ inline std::vector interleaved_to_strided( } template -static TestResultEntry tune_syclblas(int r, char transA, char transB, +static TestResultEntry tune_portblas(int r, char transA, char transB, GemmArgs a, ::blas::gemm_batch_type_t batch_type) { - TestResultEntry result("SYCL-BLAS gemm"); - auto sb_handle = get_sycl_blas_handle(); + TestResultEntry result("portBLAS gemm"); + auto sb_handle = get_portblas_handle(); { auto event = blas::helper::copy_to_device( sb_handle.get_queue(), a.init_c.data(), a.c, a.init_c.size()); @@ -157,7 +157,7 @@ void run_tune_gemm(int seed, int m, int k, int n, int batch_size, int rep, device_c, result_c, ldc, batch_size, expected_c}; { - auto result = tune_syclblas(rep, *ta_str, *tb_str, args, batch_type); + auto result = tune_portblas(rep, *ta_str, *tb_str, args, batch_type); results.push_back(result); } @@ -173,7 +173,7 @@ void run_tune_gemm(int seed, int m, int k, int n, int batch_size, int rep, #undef BENCH_PARAMS std::cout << "SIZE : " << results.size() << std::endl; - get_sycl_blas_handle().wait(); + get_portblas_handle().wait(); std::sort(results.begin(), results.end()); results.print_all(); } diff --git a/tools/auto_tuner/include/reference_gemm.hpp b/tools/auto_tuner/include/reference_gemm.hpp index 4050d593c..eaa52bbdb 100644 --- a/tools/auto_tuner/include/reference_gemm.hpp +++ b/tools/auto_tuner/include/reference_gemm.hpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename reference_gemm.hpp * diff --git a/tools/auto_tuner/include/tune.hpp b/tools/auto_tuner/include/tune.hpp index bc1b0f67f..ea6e75ea8 100644 --- a/tools/auto_tuner/include/tune.hpp +++ b/tools/auto_tuner/include/tune.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune.hpp * **************************************************************************/ -#ifndef SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ -#define SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ +#ifndef PORTBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ +#define PORTBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ #include "tuner_types.hpp" @@ -32,4 +32,4 @@ template TestResultEntry tune(int r, GemmArgs a); -#endif // SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ +#endif // PORTBLAS_TOOLS_AUTO_TUNER_TUNE_HPP_ diff --git a/tools/auto_tuner/include/tune_impl.hpp b/tools/auto_tuner/include/tune_impl.hpp index eb3203baa..bd2f15685 100644 --- a/tools/auto_tuner/include/tune_impl.hpp +++ b/tools/auto_tuner/include/tune_impl.hpp @@ -17,19 +17,19 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_impl.hpp * **************************************************************************/ -#ifndef SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ -#define SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ +#ifndef PORTBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ +#define PORTBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ #include "tuner_types.hpp" #include "utils.hpp" -#include "sycl_blas.hpp" +#include "portblas.hpp" template @@ -41,7 +41,7 @@ TestResultEntry tune(int r, GemmArgs a) { static_cast(Config::ShapeMode), static_cast(Config::VecType), VecSize, static_cast(Config::BatchType)>; TestResultEntry result(Gemm::get_type_string()); - auto sb_handle = get_sycl_blas_handle(); + auto sb_handle = get_portblas_handle(); { { auto event = blas::helper::copy_to_device( @@ -74,4 +74,4 @@ TestResultEntry tune(int r, GemmArgs a) { return result; } -#endif // SYCLBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ +#endif // PORTBLAS_TOOLS_AUTO_TUNER_TUNE_IMPL_HPP_ diff --git a/tools/auto_tuner/include/tuner_types.hpp b/tools/auto_tuner/include/tuner_types.hpp index 0b987324f..4f86583dd 100644 --- a/tools/auto_tuner/include/tuner_types.hpp +++ b/tools/auto_tuner/include/tuner_types.hpp @@ -17,21 +17,21 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tuner_types.hpp * **************************************************************************/ -#ifndef SYCLBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ -#define SYCLBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ +#ifndef PORTBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ +#define PORTBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ #include #include -#include "sycl_blas.hpp" +#include "portblas.hpp" -using sycl_blas_handle_t = ::blas::SB_Handle; +using portblas_handle_t = ::blas::SB_Handle; template using HostContainer = std::vector; @@ -118,4 +118,4 @@ struct GemmArgs { int stride_c; }; -#endif // SYCLBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ +#endif // PORTBLAS_TOOLS_AUTO_TUNER_TUNER_TYPES_HPP_ diff --git a/tools/auto_tuner/include/utils.hpp b/tools/auto_tuner/include/utils.hpp index 2f12f5262..9d1322880 100644 --- a/tools/auto_tuner/include/utils.hpp +++ b/tools/auto_tuner/include/utils.hpp @@ -17,14 +17,14 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename utils.hpp * **************************************************************************/ -#ifndef SYCLBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ -#define SYCLBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ +#ifndef PORTBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ +#define PORTBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ #include "tuner_types.hpp" @@ -34,7 +34,7 @@ #include #include -inline sycl_blas_handle_t make_sycl_blas_handle() { +inline portblas_handle_t make_portblas_handle() { cl::sycl::queue q([=](cl::sycl::exception_list ex_list) { try { for (auto &e_ptr : ex_list) { @@ -48,12 +48,12 @@ inline sycl_blas_handle_t make_sycl_blas_handle() { << q.get_device().get_info() << std::endl; - sycl_blas_handle_t sb_handle(q); + portblas_handle_t sb_handle(q); return sb_handle; } -inline sycl_blas_handle_t &get_sycl_blas_handle() { - static sycl_blas_handle_t sb_handle = make_sycl_blas_handle(); +inline portblas_handle_t &get_portblas_handle() { + static portblas_handle_t sb_handle = make_portblas_handle(); return sb_handle; } @@ -111,4 +111,4 @@ static void run_tune(int rep, double flop_cnt, TestResultEntry &result, result.gflops = gigaflop_count / seconds_per_iter.count(); } -#endif // SYCLBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ +#endif // PORTBLAS_TOOLS_AUTO_TUNER_UTILS_HPP_ diff --git a/tools/auto_tuner/src/tune_all.cpp b/tools/auto_tuner/src/tune_all.cpp index d243d3aee..499f8235b 100644 --- a/tools/auto_tuner/src/tune_all.cpp +++ b/tools/auto_tuner/src/tune_all.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_all.cpp * diff --git a/tools/auto_tuner/src/tune_nn.cpp b/tools/auto_tuner/src/tune_nn.cpp index 422b61cc0..36265a9d3 100644 --- a/tools/auto_tuner/src/tune_nn.cpp +++ b/tools/auto_tuner/src/tune_nn.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_nn.cpp * diff --git a/tools/auto_tuner/src/tune_nt.cpp b/tools/auto_tuner/src/tune_nt.cpp index 185841dfe..7fc3a0b14 100644 --- a/tools/auto_tuner/src/tune_nt.cpp +++ b/tools/auto_tuner/src/tune_nt.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_nt.cpp * diff --git a/tools/auto_tuner/src/tune_tn.cpp b/tools/auto_tuner/src/tune_tn.cpp index 58253e33f..d19845339 100644 --- a/tools/auto_tuner/src/tune_tn.cpp +++ b/tools/auto_tuner/src/tune_tn.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_tn.cpp * diff --git a/tools/auto_tuner/src/tune_tt.cpp b/tools/auto_tuner/src/tune_tt.cpp index ffd808381..245878c42 100644 --- a/tools/auto_tuner/src/tune_tt.cpp +++ b/tools/auto_tuner/src/tune_tt.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename tune_tt.cpp * From f50c3ef04f5be76c1ae20ff6deb923006eccaaf5 Mon Sep 17 00:00:00 2001 From: Alejandro Acosta Date: Thu, 17 Aug 2023 12:16:27 +0100 Subject: [PATCH 2/3] Remove comment --- benchmark/cublas/main.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/benchmark/cublas/main.cpp b/benchmark/cublas/main.cpp index 83f276729..6360cb06f 100644 --- a/benchmark/cublas/main.cpp +++ b/benchmark/cublas/main.cpp @@ -40,7 +40,6 @@ int main(int argc, char** argv) { // Initialize googlebench benchmark::Initialize(&argc, argv); - // Create a portBLAS sb_handle from the queue cublasHandle_t cublas_handle = NULL; CUBLAS_CHECK(cublasCreate(&cublas_handle)); From e8cda7fca10f051553ea510fcc9d7b2ea9c23876 Mon Sep 17 00:00:00 2001 From: Alejandro Acosta Date: Tue, 22 Aug 2023 11:57:38 +0100 Subject: [PATCH 3/3] Rebase --- benchmark/cublas/extension/omatadd.cpp | 2 +- benchmark/rocblas/extension/omatadd.cpp | 2 +- src/interface/extension/omatadd.cpp.in | 4 ++-- src/operations/extension/transpose.hpp | 16 ++++++++-------- test/unittest/extension/omatadd_test.cpp | 2 +- 5 files changed, 13 insertions(+), 13 deletions(-) diff --git a/benchmark/cublas/extension/omatadd.cpp b/benchmark/cublas/extension/omatadd.cpp index 908403ade..67cc544cf 100644 --- a/benchmark/cublas/extension/omatadd.cpp +++ b/benchmark/cublas/extension/omatadd.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatadd.cpp * diff --git a/benchmark/rocblas/extension/omatadd.cpp b/benchmark/rocblas/extension/omatadd.cpp index 7aeba6cf6..f2d1bb39b 100644 --- a/benchmark/rocblas/extension/omatadd.cpp +++ b/benchmark/rocblas/extension/omatadd.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatadd.cpp * diff --git a/src/interface/extension/omatadd.cpp.in b/src/interface/extension/omatadd.cpp.in index e1fd5be1b..eb0d3b510 100644 --- a/src/interface/extension/omatadd.cpp.in +++ b/src/interface/extension/omatadd.cpp.in @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatadd.cpp.in * @@ -25,7 +25,7 @@ #include "interface/extension_interface.hpp" #include "sb_handle/kernel_constructor.hpp" -#include "sb_handle/sycl_blas_handle.hpp" +#include "sb_handle/portblas_handle.hpp" #include "operations/extension/transpose.hpp" namespace blas { diff --git a/src/operations/extension/transpose.hpp b/src/operations/extension/transpose.hpp index 1e4e2d536..0d3adba54 100644 --- a/src/operations/extension/transpose.hpp +++ b/src/operations/extension/transpose.hpp @@ -219,7 +219,7 @@ Transpose -SYCL_BLAS_INLINE bool TransposeAdd< +PORTBLAS_INLINE bool TransposeAdd< both_trans, Tile_size, wg_size, cl_size, local_memory, in1_t, in2_t, out_t, element_t>::valid_thread(cl::sycl::nd_item<1> item) const { auto idx = item.get_global_linear_id(); @@ -229,7 +229,7 @@ SYCL_BLAS_INLINE bool TransposeAdd< template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void TransposeAdd::bind(cl::sycl::handler &cgh) { A_.bind(cgh); @@ -240,7 +240,7 @@ TransposeAdd -SYCL_BLAS_INLINE typename in1_t::index_t +PORTBLAS_INLINE typename in1_t::index_t TransposeAdd::get_size() const { // Smallest TileSize square-multiple containing input/output matrices @@ -250,7 +250,7 @@ TransposeAdd -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void TransposeAdd::adjust_access_displacement() { A_.adjust_access_displacement(); @@ -274,7 +274,7 @@ TransposeAdd -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void TransposeAdd::get_indices(cl::sycl::nd_item<1> id, index_t &in_a_idx, @@ -312,7 +312,7 @@ TransposeAdd -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void TransposeAdd::eval(cl::sycl::nd_item<1> id) { auto A = A_.get_pointer(); @@ -364,7 +364,7 @@ TransposeAdd -SYCL_BLAS_INLINE void TransposeAdd< +PORTBLAS_INLINE void TransposeAdd< both_trans, Tile_size, wg_size, cl_size, local_memory, in1_t, in2_t, out_t, element_t>::get_indices(cl::sycl::nd_item<1> id, index_t &in_a_idx, index_t &in_b_idx, index_t &in_local_idx, @@ -408,7 +408,7 @@ template template -SYCL_BLAS_INLINE void +PORTBLAS_INLINE void TransposeAdd::eval(local_memory_t local_mem, cl::sycl::nd_item<1> id) { diff --git a/test/unittest/extension/omatadd_test.cpp b/test/unittest/extension/omatadd_test.cpp index eb92e8417..382fc0f5d 100644 --- a/test/unittest/extension/omatadd_test.cpp +++ b/test/unittest/extension/omatadd_test.cpp @@ -17,7 +17,7 @@ * See the License for the specific language governing permissions and * limitations under the License. * - * SYCL-BLAS: BLAS implementation using SYCL + * portBLAS: BLAS implementation using SYCL * * @filename omatadd_test.cpp *