This repository was archived by the owner on Mar 28, 2023. It is now read-only.
[SYCL][ESIMD] Increase USM allocation size for Stencil test. #189
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@kbobrovs, I see some failing tests unrelated to my change. Are these known regressions? [2021-03-19T20:22:10.988Z] Failed Tests (4): [2021-03-19T20:22:10.988Z] SYCL :: Basic/enqueue_barrier.cpp [2021-03-19T20:22:10.988Z] SYCL :: Functor/kernel_functor.cpp [2021-03-19T20:22:10.988Z] SYCL :: InorderQueue/in_order_dmemll.cpp [2021-03-19T20:22:10.988Z] SYCL :: Reduction/reduction_transparent.cpp |
|
@kbobrovs, nevermind, I restarted the test and all passed now. So it was a glitch. |
DenisBakhvalov
approved these changes
Mar 19, 2021
myler
pushed a commit
to myler/llvm-test-suite
that referenced
this pull request
Apr 12, 2022
…-xmain CMPLRTST-14236: Adjust test driver to make it support child suites of llvm_test_suite_sycl
aelovikov-intel
pushed a commit
to aelovikov-intel/llvm
that referenced
this pull request
Mar 27, 2023
…lvm-test-suite#189) The Stencil test showed sporadic failures before last change. Investigation shows it works fine on Linux. The problem only showed up on Windows and could be voided by adjusting SLM allocation size. The GPU assembly dump is the same in the passing and failing case. Semenov, Sergey helped look into USM allocation issue on Windows vs. Linux and didn't find anything suspicious. So the problem could be still due to out-of-order block access. As discussed, in this patch the kernel SLM allocation size is increased by +16 to avoid OOB access.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The Stencil test showed sporadic failures before last change. Investigation shows it works fine on Linux. The problem only showed up on Windows and could be voided by adjusting SLM allocation size. The GPU assembly dump is the same in the passing and failing case. Semenov, Sergey helped look into USM allocation issue on Windows vs. Linux and didn't find anything suspicious. So the problem could be still due to out-of-order block access. As discussed, in this patch the kernel SLM allocation size is increased by +16 to avoid OOB access.