[UR] Update Level Zero tag for fetching

[KseniyaTikhomirova]

No description provided.

[KseniyaTikhomirova]

preview mode is also failing here #20186

https://github.com/intel/llvm/actions/runs/17977216100/job/51136612200?pr=20186

[KseniyaTikhomirova]

there is a stable fail with new L0. @nrspruit FYI

[----------] Global test environment set-up.
[----------] 1 test from urCommandBufferAppendMemBufferReadRectTestWithParam
[ RUN ] urCommandBufferAppendMemBufferReadRectTestWithParam.Success/UR_BACKEND_LEVEL_ZERO__Intel_R__oneAPI_Unified_Runtime_over_Level_Zero_V2__Intel_R__Data_Center_GPU_Max_1100_ID0ID______________________write_2d_3d
Segmentation fault from GPU at 0x80abffc01000, ctx_id: 1 (CCS) type: 0 (NotPresent), level: 0 (PTE), access: 1 (Write), banned: 1, aborting.
Segmentation fault from GPU at 0x80abffc01000, ctx_id: 1 (CCS) type: 0 (NotPresent), level: 0 (PTE), access: 1 (Write), banned: 1, aborting.
Abort was called at 290 line in file:
../../neo/shared/source/os_interface/linux/drm_neo.cpp

--
exit: -6

[nrspruit]

there is a stable fail with new L0. @nrspruit FYI

[----------] Global test environment set-up. [----------] 1 test from urCommandBufferAppendMemBufferReadRectTestWithParam [ RUN ] urCommandBufferAppendMemBufferReadRectTestWithParam.Success/UR_BACKEND_LEVEL_ZERO__Intel_R__oneAPI_Unified_Runtime_over_Level_Zero_V2__Intel_R__Data_Center_GPU_Max_1100_ID0ID______________________write_2d_3d Segmentation fault from GPU at 0x80abffc01000, ctx_id: 1 (CCS) type: 0 (NotPresent), level: 0 (PTE), access: 1 (Write), banned: 1, aborting. Segmentation fault from GPU at 0x80abffc01000, ctx_id: 1 (CCS) type: 0 (NotPresent), level: 0 (PTE), access: 1 (Write), banned: 1, aborting. Abort was called at 290 line in file: ../../neo/shared/source/os_interface/linux/drm_neo.cpp

-- exit: -6

This has nothing to do with the v1.24.3 change, that is a driver failure.

oneapi-src/level-zero@v1.24.2...v1.24.3

Nothing functionally changed in the loader so this is unrelated, the driver either changed or the test system's device failed.

[pbalcer]

I think this started failing after the latest CI driver update. We need to disable those tests and triage the issue.

[KseniyaTikhomirova]

UR precommit failure - seems to be unrelated, reported and created PR to disable unstable test #20209

/__w/llvm/llvm/build/test/conformance/usm/usm-test --gtest_filter=urUSMHostAllocTest.Success/UR_BACKEND_LEVEL_ZERO__Intel_R__oneAPI_Unified_Runtime_over_Level_Zero_V2__Intel_R__Data_Center_GPU_Max_1100_ID0ID______________________UsePoolDisabled_0_0_UR_USM_ADVICE_FLAG_SET_ACCESSED_BY_DEVICE
--
/__w/llvm/llvm/unified-runtime/test/conformance/usm/urUSMHostAlloc.cpp:64: Failure
Expected equality of these values:
  *reinterpret_cast<int *>(ptr)
    Which is: 16843009
  set_data
    Which is: 16843009

SYCL precommit failure has already been reported here #19468

Intel Arc:

Failed Tests (1):
SYCL :: USM/memops2d/memcpy2d_host_to_device.cpp

@intel/llvm-gatekeepers please consider merging, I believe the CI failures are unrelated.

[sarnex]

@KseniyaTikhomirova Seeing bindless_images/copies/copy_subregion_2D.cpp fail in postcommit after this PR, can you take a look? I see the existing GH issue but it wasn't failing before this PR and now it is, so maybe something here caused it to start failing again.

https://github.com/intel/llvm/actions/runs/18039930728/job/51337407732
https://github.com/intel/llvm/actions/runs/18044941614/job/51354109471
https://github.com/intel/llvm/actions/runs/18043294055/job/51348923722