50 points by learning_enthusiast 7 months ago flag hide 10 comments
gnarthacker 7 months ago next
Here are some great resources for learning GPU programming: \n1. CUDA C Programming - This book by Simon Kang is perfect for beginners. It provides a clear introduction to GPU programming using CUDA C. \n2. OpenCL Programming Guide - This book by Tomasz Kozinkiewicz and Marek Kubala is also very helpful. It covers both OpenCL and OpenGL programming. \n3. GPU Gems - This series of books by NVIDIA is a great resource for advanced GPU programmers. It covers a wide range of topics in GPU programming and has contributions from many leading experts in the field.
deeplearningdave 7 months ago next
Great recommendations! I would also add the NVIDIA CUDA C Programming Guide as a resource for beginners. It's free and available on the NVIDIA website.
gnarthacker 7 months ago next
That's a great addition, thanks! I would also add the Parallel Programming and Optimization with CUDA book by Sunil R. Gayen for advanced GPU programmers.
parallelprogrammer 7 months ago prev next
I would highly recommend the DirectX 12 Programming Guide for GPU programming. It provides a great introduction to DirectX 12 and it's written by Microsoft so it's an official resource.
openglguy 7 months ago next
DirectX is a great choice, but for cross-platform GPU programming I would recommend OpenGL or Vulkan. I've had great success with the OpenGL Programming Guide of Tomas Akenine-Möller and Wolfgang Engel.
deeplearningdave 7 months ago next
Vulkan is definitely worth checking out too. It's newer than OpenGL and DirectX but it's gaining a lot of popularity and momentum. The Vulkan Programming Guide by Wojciech Słoniowski is a good starting point.
parallelprogrammer 7 months ago next
I've also heard good things about the Khronos Group's Vulkan Tutorial, it's an open source tutorial that covers the basics of Vulkan and provides examples to get you started.
ml_enthusiast 7 months ago prev next
For GPU programming in machine learning, I would suggest starting with TensorFlow or PyTorch. Both of these frameworks have great documentation and supportive communities.
deeplearningdave 7 months ago next
I agree! TensorFlow and PyTorch are both great for GPU programming in machine learning. I also recommend the Caffe2 framework for those looking for a simple entry into deep learning with GPUs.
parallelprogrammer 7 months ago next
I've heard that the Apache MXNet framework is a good choice for GPU programming in machine learning as well. It's fast, scalable, and has a great community.