Dounia Khaldi
Dounia Khaldi
> Since you moved the deprecated version of the API spec to "doc/extensions/deprecated", there isn't any need for the "doc/extensions/experimental/sycl_ext_oneapi_matrix" directory. Can you move the spec from: > > sycl/doc/extensions/experimental/sycl_ext_oneapi_matrix/sycl_ext_oneapi_matrix.asciidoc...
@intel/dpcpp-specification-reviewers, please help merge
> If it is preferred for reviewing purposes I could add the temporary/initial fp19 implementation that uses uint32_t directly to this PR? Hopefully the uint32_t fp19 should be a bit...
The reason to use size_t for "joint_matrix" shape `rows` and `cols` and things like: stride and coordinates in the `joint_matrix_apply`, is to be consistent with what is currently using in...
@MrSidims , as far as I remember, you are already fixing this at the SPIRV level to make sure things are aligned with the cooperative matrix spec. But I don't...
@ggojska, you should create a fork from the repo first.
@gmlueck, can you please help review this?
int4 representation is replaced by https://github.com/intel-restricted/applications.compilers.llvm-project/pull/22863
> > > What I see is that no one from @intel/sycl-matrix-reviewers, @intel/dpcpp-esimd-reviewers and @intel/dpcpp-devops-reviewers groups look at the CI failures in these PRs. > > > GHA script creates...
> There are often other failures and the IGC team won't investigate them until they're in a release. How would you feel it we changed dev-igc testing here to only...