Skip to content

Conversation

@rafbiels
Copy link
Contributor

@rafbiels rafbiels commented Jan 7, 2025

Add a new e2e test to validate queue::fill outputs for any pattern size between 1 and 32 bytes.

Backend implementations of this feature commonly branch on the pattern size to implement an optimized path for specific values. This test ensures a wide range of cases is tested, including also odd numbers of bytes.

Re-adds #15991 reverted in #16465 with changes to fix #16434.

Add a new e2e test to validate queue::fill outputs for any
pattern size between 1 and 32 bytes.

Backend implementations of this feature commonly branch on
the pattern size to implement an optimized path for specific
values. This test ensures a wide range of cases is tested,
including also odd numbers of bytes.
Move "clang-format: off" to ensure UNSUPPORTED-TRACKER follows directly after UNSUPPORTED

Co-authored-by: Udit Kumar Agarwal <[email protected]>
Copy link
Contributor

@uditagarwal97 uditagarwal97 left a comment

Choose a reason for hiding this comment

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

Test LGTM. I'd seek an additional review from @maarquitos14 since he reviewed this test originally in #15991.

@dm-vodopyanov dm-vodopyanov merged commit fad00cc into intel:sycl Jan 8, 2025
17 checks passed
@uditagarwal97
Copy link
Contributor

Hi @rafbiels,
This test is failing on Gen12 post-commit: https://github.com/intel/llvm/actions/runs/12692767650/job/35379943179

FAIL: SYCL :: USM/fill_any_size.cpp (1994 of 2244)
******************** TEST 'SYCL :: USM/fill_any_size.cpp' FAILED ********************
Exit Code: -6

Command Output (stdout):
--
# RUN: at line 1
/__w/llvm/llvm/toolchain/bin//clang++  -Werror  -fsycl -fsycl-targets=spir64  /__w/llvm/llvm/llvm/sycl/test-e2e/USM/fill_any_size.cpp -o /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# executed command: /__w/llvm/llvm/toolchain/bin//clang++ -Werror -fsycl -fsycl-targets=spir64 /__w/llvm/llvm/llvm/sycl/test-e2e/USM/fill_any_size.cpp -o /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# note: command had no output on stdout or stderr
# RUN: at line 2
env ONEAPI_DEVICE_SELECTOR=level_zero:gpu  /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# executed command: env ONEAPI_DEVICE_SELECTOR=level_zero:gpu /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# .---command stdout------------
# | All 64 tests passed
# `-----------------------------
# RUN: at line 2
env ONEAPI_DEVICE_SELECTOR=opencl:fpga  /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# executed command: env ONEAPI_DEVICE_SELECTOR=opencl:fpga /__w/llvm/llvm/build-e2e/USM/Output/fill_any_size.cpp.tmp1.out
# .---command stderr------------
# | terminate called after throwing an instance of 'sycl::_V1::exception'
# |   what():  Enqueue process failed.
# `-----------------------------

Another example: https://github.com/intel/llvm/actions/runs/12690937422/job/35374577967. I think the failure is rather flaky.

rafbiels added a commit to rafbiels/llvm that referenced this pull request Jan 10, 2025
Test added in intel#16544 was originally flagged as UNSUPPORTED for
OpenCL CPU device only, but has now been observed to fail also
for other OpenCL devices. Flag the test as UNSUPPORTED for all
OpenCL devices.
@rafbiels
Copy link
Contributor Author

Thank you for flagging this @uditagarwal97! It looks like this issue may affect any OpenCL device, so I'm changing the UNSUPPORTED flag to cover them all in #16588

aelovikov-intel pushed a commit that referenced this pull request Jan 10, 2025
…#16588)

Test added in #16544 was originally flagged as UNSUPPORTED for OpenCL
CPU device only, but has now been observed to fail also for other OpenCL
devices. Flag the test as UNSUPPORTED for all OpenCL devices.
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.

USM/fill_any_size.cpp XPASS in OCL CPU in Nightly

4 participants