-
Notifications
You must be signed in to change notification settings - Fork 407
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
Implement UniqueToken for SYCL #3833
Conversation
Hmm.. This passes with an older SYCL-CUDA release but not with the one we are now using for testing. I'll try to check which one broke or at least if I can find one that we can compile and that works also for this pull request. |
ea0bc9b
to
cb80bda
Compare
This should work now. I'm still not quite sure about the internal compiler errors but they are related to files that shouldn't matter too much for the SYCL CI (except maybe for the header self-containment tests). I will, of course, try to dig a little more but I think the status here is acceptable for the release. As mentioned before, the tests are all passing without the workarounds for Host and Intel GPU; only the SYCL CUDA backend causes all the problems here. One of the problematic files was Other than that, I needed to rebase on top of #3815 to make all tests pass. |
Retest this please. |
2e57b56
to
a726bba
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks fine with the caveat that I don't know SYCL...
a726bba
to
254898a
Compare
Rebased. Still depends on #3815. |
Please rebase |
34b454b
to
8aa5104
Compare
This also allows running all
container
tests.AFAICT, the concurrency calculation here is an upper limit for
Intel
GPUs and should match the value we get forNvidia
GPUs. As usual, I checked that this pull request also passes onIntel
GPUs.We still don't have a good implementation for
clock_tick
so theUniqueToken
might be pretty slow.