Georgi Mirazchiyski
Georgi Mirazchiyski
After last rebase, there's a: _SYCL :: Regression/device_num.cpp_ e2e failure that seems unrelated.
> We already catch all exceptions in ur_libapi.cpp That is great, thanks a bunch for pointing it out! I admit I didn't know the loader already did that. I am...
> @GeorgeWeb Could you please create an intel/llvm PR for this? Thanks! @omarahmed1111 I can certainly add one. As a historical info, this has previously already been tested in a...
CC @0x12CC @nrspruit @JackAKirk I am reusing this issue as it seems that we still need a clarification for an appropriate mapping to CUDA. ### Result granularity of `urKernelSuggestMaxCooperativeGroupCountExp` 1....
@0x12CC @nrspruit I have updated my comment now after recent findings invalidated one part of it but the semantic of the query whether it is per compute unit or per...
> > Hello, AFAIK, `OffloadingActionBuilder::SYCLActionBuilder::initialize()` is the path taken when we do not pass `--offload-new-driver` option. Currently, `--offloard-arch` option is supported only in the new driver model ( i.e with...
@intel/llvm-reviewers-runtime / @sergey-semenov Hi, is it possible to have a look at this. I've forgotten it's been up for while. It's not a prio, just that the test fails on...
@intel/llvm-gatekeepers This one looks good to merge now. Thanks!
This is required for CUTLASS. I am converting to DRAFT for now as there is a unified-runtime dependency: https://github.com/oneapi-src/unified-runtime/pull/1796. **Update:** Ready for review.
@gmlueck > Can you update the wording in the PR description? It talks about `max_num_num_work_group_sync`, which I guess is an old name for the query? I will update it. The...