Skip to content

Mirror intel/llvm commits #2763

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
May 19, 2025
Merged

Mirror intel/llvm commits #2763

merged 5 commits into from
May 19, 2025

Conversation

kbenzie
Copy link
Contributor

@kbenzie kbenzie commented May 16, 2025

Automated changes by create-pull-request GitHub action

EwanC and others added 5 commits May 16, 2025 00:40
Closes issue intel/llvm#18330 to enhance the
semantics of the UR command-buffer in-order property such that
sync-point dependencies will always be ignored.

Also fixes up an artifact in the UR command-buffer spec where it says
that descriptors to the command-buffer creation API can be optional,
when after intel/llvm#17058 they are mandatory.
We no longer need to create a UR sampler object in order to create a
sampled image. The backends can simply use the sampler description to
create the sampled image.

Note this is an ABI break in the currently still experimental bindless
images extension for Unified Runtime.

Addresses the following issue:
#1463
Relevant discussion from
#2640 (comment)

Additionally, avoiding the allocation of a sampler object itself fixes
memory leak(s) in all adapters where the UR sampler handle was never
released which was caused by the use of the API in the SYCL RT only
calling urSamplerCreate and not Release for the sampler associated with
the image. Also, while HIP and Cuda do not further create backend
specific sampler objects in the respective adapters, Level Zero was
creating one that never got released upon freeing the image memory which
lead to leaking the Level Zero sampler memory as well. Now that Level
Zero can create a sampled image from a descriptor only, the sampler
object is not needed altogether, allowing to rid of the above mentioned
leaking memory allocations and simplify the runtime.
This reapplies some of the changes from
intel/llvm#17536 (without changes to
leak_checker which broke internal testing).

This fixes: intel/llvm#18463
@kbenzie kbenzie requested a review from a team as a code owner May 16, 2025 00:40
@github-actions github-actions bot added ci/cd Continuous integration/devliery loader Loader related feature/bug conformance Conformance test suite issues. images UR images specification Changes or additions to the specification experimental Experimental feature additions/changes/specification level-zero L0 adapter specific issues cuda CUDA adapter specific issues hip HIP adapter specific issues opencl OpenCL adapter specific issues native-cpu Native CPU adapter specific issues command-buffer Command Buffer feature addition/changes/specification auto-close labels May 16, 2025
Copy link
Contributor

Unified Runtime -> intel/llvm Repo Move Notice

Information

The source code of Unified Runtime has been moved to intel/llvm under the unified-runtime top-level directory,
all future development will now be carried out there. This was done in intel/llvm#17043.

The code will be mirrored to oneapi-src/unified-runtime and the specification will continue to be hosted at oneapi-src.github.io/unified-runtime.

The contribution guide will be updated with new instructions for contributing to Unified Runtime.

PR Migration

All open PRs including this one will be marked with the auto-close label and shall be automatically closed after 30 days.

Should you wish to continue with your PR you will need to migrate it to intel/llvm.
We have provided a script to help automate this process.

If your PR should remain open and not be closed automatically, you can remove the auto-close label.


This is an automated comment.

@kbenzie kbenzie merged commit e91179d into main May 19, 2025
1 check passed
@kbenzie kbenzie deleted the mirror-commits- branch May 19, 2025 08:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
auto-close ci/cd Continuous integration/devliery command-buffer Command Buffer feature addition/changes/specification conformance Conformance test suite issues. cuda CUDA adapter specific issues experimental Experimental feature additions/changes/specification hip HIP adapter specific issues images UR images level-zero L0 adapter specific issues loader Loader related feature/bug native-cpu Native CPU adapter specific issues opencl OpenCL adapter specific issues specification Changes or additions to the specification
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants