Cperkins sycl unregister lib 002 #68
Draft
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
…19537) Makes most operations including memory copies asynchronous except `enqueueBufferMap` Worker threads wait for dependent events instead of the main thread. The state struct is now constructed by the threads instead of being constructed by the main thread and copied which 1. reduces the capturing overhead in the enqueue lambda 2. enabled removing `mutable` from enqueue lambda Also removing `#ifdef NATIVECPU_USE_OCK`
…stant (intel#19800) * For builtin func like "__sycl_getComposite2020SpecConstantValue", if structs which are larger than 64b will be returned via sret arguments and will be initialized inside the function. So we need to unpoison the sret arguments. * Always set ZE_AFFINITY_MAST to 0 when running device sanitizer tests since device sanitizer only support one GPU card now.
…ntel#19820) intel#19719 and intel#19761 added pre-commit jobs to run E2E tests pre-built with latest "open-source" releases against the newly built sycl-toolchain libraries. Those can fail if either an actual break is happenning or if the test was doing some `FileCheck`ing and that output has changed in some way (which might not be an actual ABI break). However, I think the testing is still good enough to require an explicit approvals by folks in charge of ABI breaking changes. For the case of just output change the author should be able to convince owners that the change isn't ABI-breaking relatively easily.
This test is currently failing on Windows DG2 machines. The minimal reproducer is: [minimal_reproducer.cpp](https://github.com/user-attachments/files/21798993/minimal_reproducer.cpp) It turns out that removing "ZEX_NUMBER_OF_CCS=0:4" from the minimal reproducer makes it pass. This matches the comments in intel#18576 : the root cause here is that DG2 never supported `ZEX_NUMBER_OF_CCS`. Therefore, XFAIL-ing this test on all DG2 machines (previously it is only XFAIL-ed on Linux DG2 machines). Signed-off-by: Hu, Peisen <peisen.hu@intel.com>
Signed-off-by: Chris Perkins <chris.perkins@intel.com>
Signed-off-by: Chris Perkins <chris.perkins@intel.com>
Signed-off-by: Chris Perkins <chris.perkins@intel.com>
Signed-off-by: Chris Perkins <chris.perkins@intel.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
8 participants
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.
No description provided.