Skip to content

[SYCL][E2E][Joint Matrix] Remove some workarounds #17717

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.

Sign up for GitHub

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

Merged
merged 1 commit into from
Mar 31, 2025
Merged

[SYCL][E2E][Joint Matrix] Remove some workarounds #17717

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 0 additions & 3 deletions sycl/test-e2e/Matrix/Inputs/joint_matrix_bf16_fill_k_cache_impl.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -483,13 +483,10 @@ int main(
MCache1, NCache1, KCache1, MCache2, NCache2, KCache2>(matrix_size);
test<bfloat16, float, VnniFactor, /*TM*/ 32, /*TN*/ 64, /*TK*/ 16,
MCache1, NCache1, KCache1, MCache2, NCache2, KCache2>(matrix_size);
// `#ifndef PREFETCH` is a workaround for GSD-10535.
#ifndef PREFETCH
// The test is commented out due flaky results: GSD-10537.
// test<bfloat16, float, VnniFactor, /*TM*/ 1, /*TN*/ 64, /*TK*/ 32,
// MCache1,
// NCache1, /*KCache1*/ 32, MCache2, NCache2, KCache2>(matrix_size);
#endif // PREFETCH
test<bfloat16, float, VnniFactor, /*TM*/ 32, /*TN*/ 64, /*TK*/ 32,
MCache1, NCache1, /*KCache1*/ 32, MCache2, NCache2, KCache2>(
matrix_size);
Expand Down
3 changes: 1 addition & 2 deletions sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache_unroll.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,12 +11,11 @@

// REQUIRES: aspect-ext_intel_matrix

// RUN: %{build} -mllvm -inline-threshold=5000 %fp-model-precise -o %t.out -DMANUAL_UNROLL -DVNNI
// RUN: %{build} -mllvm -inline-threshold=2000 %fp-model-precise -o %t.out -DMANUAL_UNROLL -DVNNI
// RUN: %{run} %t.out

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -inline-threshold increased to 5000 to workaround bug in IGC: GSD-10534
// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
Expand Down
3 changes: 1 addition & 2 deletions sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache_unroll_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,12 +11,11 @@

// REQUIRES: aspect-ext_intel_matrix, gpu

// RUN: %{build} -mllvm -inline-threshold=5000 %fp-model-precise -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL -DVNNI
// RUN: %{build} -mllvm -inline-threshold=2000 %fp-model-precise -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL -DVNNI
// RUN: %{run} %t_gpu.out

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -inline-threshold increased to 5000 to workaround bug in IGC: GSD-10534
// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
Expand Down
Loading