Nicolas Miller

So this is in a fairly good state (I do need to resolve conflicts), and it works well on CUDA platforms but the CI does flag some issues on other...

/verify with https://github.com/intel/llvm-test-suite/pull/1102

After reading through everything again, given all the legitimate concerns and the fact that the equivalent OpenMP feature has now also been reverted upstream (https://reviews.llvm.org/D118493#4178250) after similar discussions, I think...

Adding the stale label, there hasn't been updates in a while and the issue might be partially solve, feel free to comment if the issue persist, otherwise the issue will...

Hello, quick update on this so after looking into it a bit further, the implementation in this PR can easily be tweaked to support 64 bits sub-groups, however I found...

Since then we've added support for cupti for the CUDA plugin, it's possible that the cupti libraries are not on your `PATH`, which would prevent the CUDA plugin from loading...

For 3, during the runtime (and also for running `sycl-ls`)

That seems okay, the warning should be safe to ignore as well, it doesn't work properly with CUDA. Does `sycl-ls` find the GPU with the `%PATH%` fixed? You can also...

> One thing I noticed that is different from your example on the GetStartedGuide is that I need to include "LANGUAGES CXX" in my project() line in cmakelists.txt or else...

Since it's working for you now @matthewphilipdixon I'm going to close this ticket, please feel free to re-open it, or open a new one if you have further issues!