3 ms·

Are you sure? I'm pretty sure I can branch on core-specific measures like (in CUDA) TheadIdx and BlockIdx

• gpuhacker 8y ago
  It's true that the programming model allows that, but underneath all threads within the warp will execute the same instructions. However if there is a branch, some threads can be predicated so their instructions have no effect. This is called warp divergence and can become a performance issue. If possible branch only on threadidx using multiples of the warp size. There's a cool slide deck on implementing a parallel sum algorithm that explains this really well.
  • dialecticDolt 8y ago
    Do you have any idea what the source of the slide deck could have been? It sounds very interesting and I'd love to see it.
    • gpuhacker 8y ago
      Sure! I meant this one: https://developer.download.nvidia.com/assets/cuda/files/reduction.pdf https://developer.download.nvidia.com/assets/cuda/files/redu...
• 0-_-0 8y ago
  Yes but can you do that from Julia without writing CUDA/OpenCL code? I'm happy that Julia supports GPU programming for simple code, but I don't see how you can run algorithms with inter-thread communication.
  • simondanisch 8y ago
    That's how you do it: https://github.com/JuliaGPU/GPUArrays.jl/blob/master/src/linalg.jl#L29 https://github.com/JuliaGPU/GPUArrays.jl/blob/master/src/lin... Or directly with cuda shfl intrinsics: https://github.com/JuliaGPU/CUDAnative.jl/blob/b249dfd145501632695cc6e03cdf98e49f5c8708/examples/reduce/reduce.jl#L22 https://github.com/JuliaGPU/CUDAnative.jl/blob/b249dfd145501...
    • 0-_-0 8y ago
      Oh cool! Thanks!
  • byt143 8y ago
    Yes you can.