Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Integrate llvm-project @f1595ecfdce5387e41826fd72ff930a1a39ae398 #18897

Merged
merged 8 commits into from
Oct 31, 2024

Conversation

Max191
Copy link
Contributor

@Max191 Max191 commented Oct 25, 2024

@Max191
Copy link
Contributor Author

Max191 commented Oct 25, 2024

Waiting on stablehlo and torch-mlir to bump llvm, and then will bump them here to fix build failures.

EDIT: Both landed and bumped.

@Max191 Max191 force-pushed the integrates/llvm-20241024 branch 3 times, most recently from a46d4bd to bee3ab3 Compare October 30, 2024 14:18
@Max191 Max191 marked this pull request as ready for review October 31, 2024 14:31
@@ -380,6 +379,7 @@
"onnx/node/generated/test_sce_none_weights_expanded",
"onnx/node/generated/test_sce_none_weights_log_prob",
"onnx/node/generated/test_sce_none_weights_log_prob_expanded",
"onnx/node/generated/test_shape_clip_start",
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This shape test fails on top of main too, so I don't think the failure is related to the integrate. There is an issue filed in torch-mlir for this test here: llvm/torch-mlir#3841.

@Max191 Max191 merged commit df2b8a4 into main Oct 31, 2024
39 checks passed
@Max191 Max191 deleted the integrates/llvm-20241024 branch October 31, 2024 16:04
@benvanik
Copy link
Collaborator

@Max191 this has a typo in one of the tests (tile_and_distribute_to_workgroups.mlir):

        %cl_4 = hal.interface.constant.load layout(#pipeline_layout) ordinal(4) : i32
        %cl_5 = hal.interface.constant.load layout(#pipeline_layout) ordinal(5) : i32
        %0 = arith.index_castui %cl_0 : i32 to index
        %1 = arith.index_castui %cl_1 : i32 to index
        %2 = arith.index_castui %cl_2 : i32 to index
        %3 = arith.index_castui %cl_3 : i32 to index
        %tile0 = arith.index_castui %cl_3 : i32 to index
        %tile1 = arith.index_castui %cl_3 : i32 to index

that should be

        %tile0 = arith.index_castui %cl_4 : i32 to index
        %tile1 = arith.index_castui %cl_5 : i32 to index

(the new ones that were added)

@Max191
Copy link
Contributor Author

Max191 commented Oct 31, 2024

@Max191 this has a typo in one of the tests (tile_and_distribute_to_workgroups.mlir):

Fixed here: #18982

Max191 added a commit that referenced this pull request Oct 31, 2024
Fixes a typo in a lit test that was added in
#18897.

Signed-off-by: Max Dawkins <[email protected]>
@ScottTodd
Copy link
Member

@ScottTodd
Copy link
Member

Build error (warning as error) on gcc: https://github.com/iree-org/iree/actions/runs/11627279570/job/32380341216#step:4:6312

Left a comment upstream: llvm/torch-mlir#3813 (comment)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants