Show HN: GPULlama3.java Llama Compilied to PTX/OpenCL Now Integrated in Quarkus

6 comments

• mikepapadim 3 days ago

  https://github.com/beehive-lab/GPULlama3.java

• lostmsu 3 days ago

  Does it support flash attention? Use tensor cores? Can I write custom kernels?

  UPD. found no evidence that it supports tensor cores, so it's going to be many times slower than implementations that do.

  [-]
  • mikepapadim 2 days ago

    Yes, when you use the PTX backend it supports Tensor Cores.It has also implementation for flash attention. You can also write your own kernels, have a look here: https://github.com/beehive-lab/GPULlama3.java/blob/main/src/... https://github.com/beehive-lab/GPULlama3.java/blob/main/src/...

    [-]
    • lostmsu 2 days ago

      TornadoVM GitHub has no mentions of tensor cores or WMMA instructions. The only mention of tensor cores is in 2024 and states they are not used: https://github.com/beehive-lab/TornadoVM/discussions/393

      [-]
      • mikepapadim 2 days ago

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

        [-]
        • lostmsu 11 hours 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:IL...

          Good article: https://alexarmbr.github.io/2024/08/10/How-To-Write-A-Fast-M...