4 ms·
Semi-relatedly and unfortunately, AI research is locked-in to NVIDIA (on Linux) because all the big lienar algebra and autodifferentiation frameworks work only
by rhaps0dy 9y ago
Semi-relatedly and unfortunately, AI research is locked-in to NVIDIA (on Linux) because all the big lienar algebra and autodifferentiation frameworks work only with CUDA.
- lloeki 9y agoThis is totally not my field so it may very well be the stupidest question ever: is there technically anything that prevents implementing CUDA on top of non-NVIDIA hardware?
- FnuGk 9y agoCUDA is made by nvidia
- baldfat 9y agoNot sure why you were down voted. CUDA is developed by NVIDIA. CUDA is a registered trademark by NVIDIA - https://trademarks.justia.com/850/30/cuda-85030071.html https://trademarks.justia.com/850/30/cuda-85030071.html Here is the CUDA webpage https://developer.nvidia.com/cuda-zone https://developer.nvidia.com/cuda-zone
- iforgotpassword 9y agoBecause it's a pointless answer that basically just repeats part of the question, providing absolutely no new information.
- baldfat 9y agoWell I think it makes a short answer NVIDIA doesn't let their proprietary CUDA technology work on non-NVIDIA devices.
- orbifold 9y agoNo there isn't, Google developed a CUDA frontend for clang/llvm which targets PTX (nvidias virtual machine instruction set), in principle someone could come along and implement a different backend for this frontend targeting for example AMDs gpu instruction architecture or something different. The other ingredient missing is a reimplementation of CUDAs runtime libraries.
- cjalmeida 9y agoThe last part is particularly important. Most deep learning libs rely on things like cuDNN which implement by hand basic operations very efficiently. The ROCm initiative by AMD already "transpile" most CUDA kernels to being AMD compatible, but it seems performance is not yet comparable in real word benchmarks.
- bryanlarsen 9y agoAMD is working on a tool to convert CUDA code: https://github.com/ROCm-Developer-Tools/HIP https://github.com/ROCm-Developer-Tools/HIP
- rodburns 9y agoWe are working on an OpenCL based layer that enables AMD GPUs, see these links for more info https://developer.codeplay.com/computecppce/latest/getting-started-with-tensorflow https://developer.codeplay.com/computecppce/latest/getting-s... https://github.com/tensorflow/tensorflow/issues/22 https://github.com/tensorflow/tensorflow/issues/22
- kuschku 9y agoNot for long. Some of them actually work with AMD now, and with NVIDIA now banning use of GeForce cards for AI research (https://news.ycombinator.com/item?id=15983587 https://news.ycombinator.com/item?id=15983587), NVIDIA is going to lose that market very quickly.
- baldfat 9y agoAMD is working on ROCm https://rocm.github.io/ https://rocm.github.io/ It will be an open source answer to CUDA and I hope this will now get more support. "We are excited to present ROCm, the first open-source HPC/Hyperscale-class platform for GPU computing that’s also programming-language independent. We are bringing the UNIX philosophy of choice, minimalism and modular software development to GPU computing. The new ROCm foundation lets you choose or even develop tools and a language run time for your application."
- Drdrdrq 9y agoDoes anyone here know how fqr along this effort is? Is it usable today (in the sense that one could develop TF backend which uses ROCm)?