Skip to content

Conversation

@bb-sycl
Copy link
Contributor

@bb-sycl bb-sycl commented Nov 4, 2025

Scheduled drivers uplift

@bb-sycl bb-sycl requested a review from a team as a code owner November 4, 2025 03:22
@sarnex
Copy link
Contributor

sarnex commented Nov 4, 2025

@nrspruit In this PR we update the L0 to 1.25.2 and the GPU runtime to 25.40.35563.4, but it looks like we are still seeing the failures from the symbol issue in the other PR. Any ideas?

Failed Tests (14):
  SYCL :: Adapters/level_zero/event-leak.cpp
  SYCL :: Adapters/level_zero/interop-image-get-native-mem.cpp
  SYCL :: Adapters/level_zero/interop-image-ownership.cpp
  SYCL :: Adapters/level_zero/interop-image.cpp
  SYCL :: Adapters/level_zero/interop-thread.cpp
  SYCL :: Adapters/level_zero/interop.cpp
  SYCL :: Adapters/level_zero/interop_memcpy.cpp
  SYCL :: Adapters/level_zero/track_indirect_access_memory.cpp
  SYCL :: DeprecatedFeatures/DiscardEvents/discard_events_l0_inorder.cpp
  SYCL :: DeviceImageBackendContent/L0_interop_test.cpp
  SYCL :: Matrix/joint_matrix_bfloat16_accumulator.cpp
  SYCL :: Matrix/joint_matrix_half_accumulator.cpp
  SYCL :: Regression/barrier_waitlist_with_interop_event.cpp
  SYCL :: Regression/barrier_with_work.cpp

@sarnex sarnex force-pushed the ci/update_gpu_driver-linux-25.40.35563.4 branch from 4f5775c to b9d6ff5 Compare November 6, 2025 22:10
@sarnex
Copy link
Contributor

sarnex commented Nov 7, 2025

@intel/sycl-matrix-reviewers Seeing new failures on this driver, two questions:

  1. Is it okay if I disable them so I can bump the driver?
  2. Are these known failures or should I make an issue?

Ex from BMG

  SYCL :: Matrix/joint_matrix_bfloat16_accumulator.cpp
  SYCL :: Matrix/joint_matrix_half_accumulator.cpp

@dkhaldi
Copy link
Contributor

dkhaldi commented Nov 7, 2025

Yes, please mark them as xfail on BMG and create an issue for:
SYCL :: Matrix/joint_matrix_bfloat16_accumulator.cpp
SYCL :: Matrix/joint_matrix_half_accumulator.cpp

@sarnex
Copy link
Contributor

sarnex commented Nov 7, 2025

@dkhaldi Thanks, but there are also failures on Arc and PVC. Okay to XFAIL those too?

Arc

  SYCL :: Matrix/joint_matrix_bf16_fill_k_cache_SLM.cpp

PVC

  SYCL :: Matrix/joint_matrix_bf16_fill_k_cache_arg_dim.cpp
  SYCL :: Matrix/joint_matrix_bf16_fill_k_cache_runtime_dim.cpp

@dkhaldi
Copy link
Contributor

dkhaldi commented Nov 7, 2025

Correct, these should fail as they are already failing with igc-dev . No need for issue as it should be the same as igc-dev one

SYCL :: Matrix/joint_matrix_bf16_fill_k_cache_arg_dim.cpp
SYCL :: Matrix/joint_matrix_bf16_fill_k_cache_runtime_dim.cpp

but they should xpass with igc-dev now

@dkhaldi
Copy link
Contributor

dkhaldi commented Nov 7, 2025

As for Matrix/joint_matrix_bf16_fill_k_cache_SLM.cpp on arc. This is new to me. So please create an issue.

@sarnex
Copy link
Contributor

sarnex commented Nov 7, 2025

Will do, thanks!

Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
@sarnex sarnex requested a review from a team as a code owner November 7, 2025 16:20
// XFAIL: linux && gpu-intel-dg2 && !igc-dev
// XFAIL-TRACKER: aspect-ext_intel_matrix, gpu

https://github.com/intel/llvm/issues/20595
Copy link
Contributor

Choose a reason for hiding this comment

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

I think this link should be added to the XFAIL-TRACKER, and aspect-ext_intel_matrix, gpu be added to REQUIRES lol.

Copy link
Contributor

Choose a reason for hiding this comment

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

the technical term for this is 'whoops', thanks lol

Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
@sarnex sarnex requested a review from a team as a code owner November 7, 2025 20:55
@sarnex sarnex requested a review from uditagarwal97 November 7, 2025 20:55
Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>
@sarnex
Copy link
Contributor

sarnex commented Nov 10, 2025

@intel/sycl-matrix-reviewers Can someone please take a final look at this? Thanks!

Copy link
Contributor

@dkhaldi dkhaldi left a comment

Choose a reason for hiding this comment

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

LGTM

Copy link
Contributor

@YixingZhang007 YixingZhang007 left a comment

Choose a reason for hiding this comment

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

LGTM! Thanks!

@sarnex
Copy link
Contributor

sarnex commented Nov 10, 2025

Thanks!

@sarnex sarnex merged commit ccd981d into sycl Nov 10, 2025
38 of 39 checks passed
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.

6 participants