Marcos Maronas
Marcos Maronas
> This PR so far only contains changes to the 'add' atomic tests to get initial feedback. I will add other tests later so that review feedback won't potentially require...
> No problem. I've clang-formatted my changes and added `.` to the comments in the new code. I can do it for the existing code as well as I go....
> @intel/dpcpp-devops-reviewers, @intel/llvm-reviewers-runtime, any ideas what could cause this error: It is weird that the first run works and the second doesn't. My guess is that the driver/hardware is left...
I just saw the `sycl-ls` issue again in https://github.com/intel/llvm/actions/runs/8180367243/job/22368716100?pr=12925. Note that the runner is different this time: `intel_sycl05`.
Another failure with runner `intel_sycl05`: https://github.com/intel/llvm/actions/runs/8156710255/job/22297115170?pr=12582
Friendly ping @AlexeySachkov
The CI failure was already found in https://github.com/intel/llvm/pull/12791, and it's unrelated to this PR. This is ready to merge @intel/llvm-gatekeepers
The pre-commit failure is known and has been reported already in https://github.com/intel/llvm/pull/12791. I don't think it's related to these changes.
> > I think we also should check whether the device supports `aspect::usm_device_allocations`. I think you can do that by adding `aspect-usm_device_allocations` to the `REQUIRES` lines for each test. >...
> > > > I think we also should check whether the device supports `aspect::usm_device_allocations`. I think you can do that by adding `aspect-usm_device_allocations` to the `REQUIRES` lines for each...