Tile primitives for speedy kernels
Repository Index
132,241 visible repositories in the index
1–9 of 9 · Most stars
Mirage Persistent Kernel: Compiling LLMs into a MegaKernel
FlashKDA: high-performance Kimi Delta Attention kernels
Flash Attention in ~100 lines of CUDA (forward pass only)
GPU accelerated decision optimization
- gpu
- linear-programming
- optimization
- cuda
RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.
- anns
- building-blocks
- clustering
- cuda
cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.
- attention
- blackwell
- cuda-kernels
- flash-attention
cuVS - a library for vector search and clustering on the GPU
- anns
- clustering
- cuda
- distance
Graphics Processing Units Molecular Dynamics
- molecular-dynamics-simulation
- heat-transport
- cuda
- molecular-dynamics