Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

A Quirky Assortment of CuTe Kernels

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

FlashInfer: Kernel Library for LLM Serving

Accelerating MoE with IO and Tile-aware Optimizations

FlashMLA: Efficient Multi-head Latent Attention Kernels

PyTorch media decoding and encoding

DeepEP: an efficient expert-parallel communication library