Abhishek Bagusetty
Abhishek Bagusetty
@liruipeng @waynemitchell I have pushed all the relevant updates to this PR
To address #6718
@zjin-lcf You are right, it needs the size of 64, which is here: https://github.com/intel/llvm/blob/sycl/sycl/include/sycl/ext/oneapi/sub_group_mask.hpp#L28 I am not sure how to change the sub group mask size to 64 in a...
@mdtoguchi I see that the linking issue is addressed with the PR in several applications on atleast a couple of different AMD GPU archs.
Hi @jasukhar One of the library ([Hypre](https://github.com/hypre-space/hypre)) with SYCL backend wanted to tap into oneMKL cuSparse support and wanted to check if there were plans to support cuSolver
If this is associated with oneMKL related for HIP backend, please refer to the following threads (for workaround) in addition to the above: https://github.com/oneapi-src/oneMKL/issues/336
I was trying to use the following snippet to get the API versioning and it seems to be wrong, though I was using L0 v1.13.1 based on L0-spec v1.7.0 ```bash...
@mwarusz Thanks for the snippets. I believe you were using these modules for the above: ``` module use /soft/testing/modulefiles/ module load intel-UMD23.05.25593.11/23.05.25593.11 module load dpcpp-master ``` Use of `SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS` will...
Also whether the above is using explicit or implicit scaling
Thanks so much for promptly testing in detail. If I see with future outlook, (`SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1` being default): Explicit scaling works as expected. Implicit scaling (i.e., ZE_AFFINITY_MASK=0): The test fails with...