π Study GPU programming with CUDA and PyTorch
- GPU Mode: [Youtube Channel],[github],
- How to profile CUDA kernels in PyTorch
- See the profile directory for examples on how to use PyTorch Profiler with CUDA kernels.
- (Todo) Getting Started with CUDA
- (Todo) Understanding Computation and Memory Architecture
- (Todo) Fusing PyTorch Optimizer (optimizing optimizer)
... And so on
- The ML/AI Engineer's starter guide to GPU Programming
- Getting Started With CUDA for Python Programmers - Jeremy Howard. (2024, January 28)
- Optimizing MatMul from Naive to CUBLAS - Simon Boehm (Anthropic)
- Tensor Parallelism with CUDA - Multi-GPU Matrix Multiplication - Dennis Kennetz (Oracle)
- 100 days of CUDA challenge