[SYCL][Graph] Update to UR cmd-buf enqueue rename

[EwanC]

Reflects change in name of UR entry-point from
urCommandBufferEnqueueExp to urEnqueueCommandBufferExp in oneapi-src/unified-runtime#2606

[EwanC]

ping @intel/llvm-reviewers-runtime for review

[EwanC]

@aelovikov-intel All the prebuilt e2e CI runs are failing here with tests that look too generic for my path. From reading #16793 & #16682 I don't understand what could be causing this, do you have any insight?

[aelovikov-intel]

@aelovikov-intel All the prebuilt e2e CI runs are failing here with tests that look too generic for my path. From reading #16793 & #16682 I don't understand what could be causing this, do you have any insight?

Hard to say right now, let's get #16866 merged and rebase this on top of trunk after.

[EwanC]

Hi @EwanC , sorry for messing up your PR. I don't have permission to force-push and clear up my commit. Please help remove my accidentally added commit. Really sorry for that.

No worries 🙂 I've removed your commits now

[EwanC]

Closing this PR and reverting the breaking UR changes in oneapi-src/unified-runtime#2688 Giving more time to investigate what's going with CI without blocking the UR merge queue

[yingcong-wu]

I think there are two direct causes here leading to the test failures.

The first one, as @aelovikov-intel pointed out, is the missing feature preview-breaking-changes-supported in the build stage, which lead to many conditional builds not being executed.

The second one is that some other failures( SYCL :: AtomicRef/device_has_aspect_atomic64_level_zero.cpp for example ), are listed as unsupported in the build stage, which also leads to missing executables in run stage. But the root cause is still under investigation.

[yingcong-wu]

The second one is that some other failures( SYCL :: AtomicRef/device_has_aspect_atomic64_level_zero.cpp for example ), are listed as unsupported in the build stage, which also leads to missing executables in run stage. But the root cause is still under investigation.

That is because level_zero_dev_kit is also missing from build stage global features, together with cuda_dev_kit.

[yingcong-wu]

I notice that those features are all tested in llvm/sycl/test-e2e/lit.cfg.py using subprocess.getstatusoutput. I added a print to those tests and they all show the following error:

2025-02-11T15:09:01.2540004Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:328: note: /usr/bin/ld: /__w/llvm/llvm/toolchain/bin/../lib/libsycl.so: undefined reference to `urEnqueueCommandBufferExp@LIBUR_LOADER_0.12'
2025-02-11T15:09:01.2540580Z clang++: error: linker command failed with exit code 1 (use -v to see invocation)
2025-02-11T15:09:01.2541173Z lit.py: /__w/llvm/llvm/llvm/sycl/test-e2e/lit.cfg.py:356: note: /usr/bin/ld: /__w/llvm/llvm/toolchain/bin/../lib/libsycl-preview.so: undefined reference to `urEnqueueCommandBufferExp@LIBUR_LOADER_0.12'

(In https://github.com/intel/llvm/actions/runs/13265484428/job/37031762080, line 104607).

I think there is a problem with those feature checks(maybe linking to an out-dated UR libs?), and I think it would impact any UR ABI breaking changes.

Tagging @sarnex .

I would stop my investigation here. Hope this is helpful.

[EwanC]

Thanks for taking the time to investigate, I've created an issue for this so that it can be better tracked and resolved #16982

[sarnex]

I assigned the issue to myself and I'll take a look today, follow the issue for updates.