CUDA Tile IR is an MLIR-based intermediate representation and compiler infrastructure for CUDA kernel optimization, focusing on tile-based computation patterns and optimizations targeting NVIDIA tensor core units.
Repository Index
142,273 visible repositories in the index
1–4 of 4 · Most stars
Graphics Processing Units Molecular Dynamics
- molecular-dynamics-simulation
- heat-transport
- cuda
- molecular-dynamics
This is a complete package of recent deep learning methods for 3D point clouds in pytorch (with pretrained models).
A high-performance engine for matrix math, physics simulation, and MCAD and ECAD as code. RunMat checks types, sizes and units before anything runs, executes on CPU and GPU, and runs existing MATLAB-language code as written.
- matlab
- rust
- jit-compiler
- linear-algebra