Add thread sanitizer CI pipeline

ci/.gitlab-ci.yml

@@ -1,19 +1,20 @@
 include:
   - local: 'ci/cpu/asan_ubsan_lsan.yml'
-  - local: 'ci/cpu/clang15_release_cxx20.yml'
-  - local: 'ci/cpu/clang15_release_stdexec.yml'
-  - local: 'ci/cpu/clang15_release.yml'
-  - local: 'ci/cpu/clang16_release.yml'
-  - local: 'ci/cpu/clang18_release.yml'
-  - local: 'ci/cpu/gcc11_release_stdexec.yml'
-  - local: 'ci/cpu/gcc11_debug_stdexec.yml'
-  - local: 'ci/cpu/gcc12_release_cxx20.yml'
-  - local: 'ci/cpu/gcc13_codecov.yml'
-  - local: 'ci/cpu/gcc13_release.yml'
-  - local: 'ci/cuda/gcc11_release.yml'
-  - local: 'ci/cuda/gcc11_release_scalapack.yml'
-  - local: 'ci/cuda/gcc11_codecov.yml'
-  - local: 'ci/cuda/gcc11_debug_scalapack.yml'
-  - local: 'ci/cuda/gcc13_release_stdexec.yml'
-  - local: 'ci/rocm/clang14_release.yml'
-  - local: 'ci/rocm/clang14_release_stdexec.yml'
+  - local: 'ci/cpu/tsan.yml'
+  # - local: 'ci/cpu/clang15_release_cxx20.yml'
+  # - local: 'ci/cpu/clang15_release_stdexec.yml'
+  # - local: 'ci/cpu/clang15_release.yml'
+  # - local: 'ci/cpu/clang16_release.yml'
+  # - local: 'ci/cpu/clang18_release.yml'
+  # - local: 'ci/cpu/gcc11_release_stdexec.yml'
+  # - local: 'ci/cpu/gcc11_debug_stdexec.yml'
+  # - local: 'ci/cpu/gcc12_release_cxx20.yml'
+  # - local: 'ci/cpu/gcc13_codecov.yml'
+  # - local: 'ci/cpu/gcc13_release.yml'
+  # - local: 'ci/cuda/gcc11_release.yml'
+  # - local: 'ci/cuda/gcc11_release_scalapack.yml'
+  # - local: 'ci/cuda/gcc11_codecov.yml'
+  # - local: 'ci/cuda/gcc11_debug_scalapack.yml'
+  # - local: 'ci/cuda/gcc13_release_stdexec.yml'
+  # - local: 'ci/rocm/clang14_release.yml'
+  # - local: 'ci/rocm/clang14_release_stdexec.yml'

ci/common-ci.yml

@@ -34,7 +34,7 @@ stages:
     reports:
       dotenv: build.env
   variables:
-    SPACK_SHA: 0905edf592752742eb4ddab3a528d3aee8f92930
+    SPACK_SHA: develop-2024-10-06
     SPACK_DLAF_REPO: ./spack
     DOCKER_BUILD_ARGS: '[
         "BASE_IMAGE",

ci/cpu/tsan.yml

@@ -0,0 +1,35 @@
+include:
+  - local: 'ci/common-ci.yml'
+
+cpu tsan deps:
+  extends: .build_deps_common
+  variables:
+    EXTRA_APTGET: "clang-18 libclang-rt-18-dev libomp-18-dev"
+    COMPILER: clang@18
+    SPACK_ENVIRONMENT: ci/docker/tsan.yaml
+    USE_MKL: "ON"
+    DEPS_IMAGE: $CSCS_REGISTRY_PATH/cpu-tsan/deps
+
+cpu tsan build:
+  extends:
+    - .build_common
+    - .build_for_daint-mc
+  needs:
+    - cpu tsan deps
+  variables:
+    DLAF_IMAGE: $CSCS_REGISTRY_PATH/cpu-tsan/dlaf:$CI_COMMIT_SHA
+
+cpu tsan test:
+  extends: .run_common
+  needs:
+    - cpu tsan build
+  variables:
+    TSAN_OPTIONS: "ignore_noninstrumented_modules=1:halt_on_error=1:history_size=0"
+    PIKA_SMALL_STACK_SIZE: "0x100000"
+    PIKA_MEDIUM_STACK_SIZE: "0x100000"
+    # Override use of libSegFault, not necessary with sanitizers
+    LD_PRELOAD: ""
+  trigger:
+    include:
+      - artifact: pipeline.yml
+        job: cpu tsan build

ci/ctest_to_gitlab.sh

@@ -61,7 +61,7 @@ JOB_TEMPLATE="
     USE_MPI: 'YES'
     DISABLE_AFTER_SCRIPT: 'YES'
     DLAF_HDF5_TEST_OUTPUT_PATH: \$CI_PROJECT_DIR
-  script: mpi-ctest -L {{LABEL}}
+  script: mpi-ctest --timeout 420 -E eigensolver -L {{LABEL}}
   artifacts:
     paths:
       - codecov-reports/"
@@ -93,14 +93,14 @@ JOB_TEMPLATE="
   variables:
     SLURM_CPUS_PER_TASK: {{CPUS_PER_TASK}}
     SLURM_NTASKS: {{NTASKS}}
-    SLURM_TIMELIMIT: '20:00'
+    SLURM_TIMELIMIT: '2:00:00'
     SLURM_UNBUFFEREDIO: 1
     SLURM_WAIT: 0
     PULL_IMAGE: 'YES'
     USE_MPI: 'YES'
     DISABLE_AFTER_SCRIPT: 'YES'
     DLAF_HDF5_TEST_OUTPUT_PATH: \$CI_PROJECT_DIR
-  script: mpi-ctest -L {{LABEL}}"
+  script: mpi-ctest --timeout 420 -E eigensolver -L {{LABEL}}"
 fi
 
 JOBS=""

ci/docker/tsan.yaml

@@ -0,0 +1,34 @@
+#
+# Distributed Linear Algebra with Future (DLAF)
+#
+# Copyright (c) 2018-2024, ETH Zurich
+# All rights reserved.
+#
+# Please, refer to the LICENSE file in the root directory.
+# SPDX-License-Identifier: BSD-3-Clause
+#
+
+spack:
+  include:
+  - /spack_environment/common.yaml
+
+  view: false
+  concretizer:
+    unify:
+      true
+
+  specs:
+    - dla-future@master build_type=Debug cxxflags="-fsanitize=thread -fno-omit-frame-pointer" ldflags="-fsanitize=thread" +miniapps +ci-test
+
+  packages:
+    all:
+      variants:
+        - 'build_type=RelWithDebInfo'
+    pika:
+      require:
+        - 'build_type=Debug'
+        - 'malloc=system'
+        - '+sanitizers'
+        - '+stdexec'
+        - 'cxxflags="-fsanitize=thread -fno-omit-frame-pointer"'
+        - 'ldflags="-fsanitize=thread"'