3 ms·

https://github.com/beehive-lab/TornadoVM/pull/732 https://github.com/beehive-lab/TornadoVM/pull/732 https://github.com/beehive-lab/TornadoVM/pull/313 https://gi

• lostmsu 10mo ago
  I believe these are SIMD. Tensor cores require MMA family of instructions. Ask me how I know. :) https://github.com/m4rs-mt/ILGPU/compare/master...lostmsu:ILGPU:wgmma/make-descriptor https://github.com/m4rs-mt/ILGPU/compare/master...lostmsu:IL... Good article: https://alexarmbr.github.io/2024/08/10/How-To-Write-A-Fast-Matrix-Multiplication-From-Scratch-With-Tensor-Cores.html#tensor-core-vs-ffma https://alexarmbr.github.io/2024/08/10/How-To-Write-A-Fast-M...