Skip to content

[SYCL][NATIVECPU] use ur_memory_scope_capability_flags_t in NativeCPU adapter #18462

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

Closed
wants to merge 10 commits into from

Conversation

uwedolinsky
Copy link
Contributor

A fix for the NativeCPU adapter to use ur_memory_scope_capability_flags_t instead of uint64_t.

Required for NativeCPU to pass AtomicRef -related e2e and SYCL-CTS tests.

@uwedolinsky uwedolinsky requested a review from a team as a code owner May 14, 2025 11:07
…ldown"

This reverts commit 8b35b01, reversing
changes made to 2ce5fef.
@uwedolinsky uwedolinsky requested review from a team and bader as code owners May 17, 2025 19:19
@uwedolinsky uwedolinsky requested a review from npmiller May 17, 2025 19:19
Copy link
Contributor

@hvdijk hvdijk left a comment

Choose a reason for hiding this comment

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

Something seems to have gone in the last merge or rebase to re-trigger CI, this PR isn't supposed to have changes in 5000+ files

@hvdijk
Copy link
Contributor

hvdijk commented May 19, 2025

The thousands of files changed here are because of a force-push to the sycl branch trying to merge #18403. There was a window of about ten minutes where commit 2b32ece was part of the sycl branch and you got unlucky that you rebased in just that window. @sarnex Is there a policy on force pushes to the sycl branch? I would not have expected that to be done under any circumstances, it also broke our internal mirrorring of the repo. If we have to assume that this branch gets force-pushed to we'll need to change our internal processes to account for that.

@uwedolinsky uwedolinsky marked this pull request as draft May 19, 2025 12:02
@uwedolinsky
Copy link
Contributor Author

Replaced by #18537

@sarnex
Copy link
Contributor

sarnex commented May 19, 2025

@hvdijk Really sorry about this. Usually when merging the pulldowns, we rely on a bot/tool to do the merge automatically. Unforunately the bot was broken so we tried to do the pulldown merge but I messed it up and the tree was broken. We decided a force-push was the best way to fix the issue. In general we never want to force push, and it's blocked by default, it just had to be done to fix the bad manual merge.

The root fix is getting the bot fixed so we never have to even try to merge manually, we've reached out to the team that owns it.

I want to say force pushes should never happen, but in rare cases like this they may, but we're trying to prevent having to do them ever.

Sorry again, and let me know if there's a way I can help.

@hvdijk
Copy link
Contributor

hvdijk commented May 19, 2025

@sarnex Thanks, if it's only ever going to happen in extremely rare cases, that's okay. We needed some manual intervention on our end to restore the mirror, but we did get it working again, and @uwedolinsky was able to sort of restore this PR by creating it anew.

@sarnex
Copy link
Contributor

sarnex commented May 19, 2025

Glad to hear, sorry again guys

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.

4 participants