AndreSlavescu/mHC.cu

🍎 One kernel a day keeps high latency away. A hands-on CUDA learning path featuring a rich collection of kernels, from the basics to peak performance, seamlessly integrated as PyTorch C++ extensions.

CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

Learn CUDA with PyTorch

Tile primitives for speedy kernels

CUDA Templates and Python DSLs for High-Performance Linear Algebra

CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.

A Quirky Assortment of CuTe Kernels