NVIDIA / cuopt
GPU accelerated decision optimization
See what the GitHub community is most excited about this month.
GPU accelerated decision optimization
RTP-LLM: Alibaba's high-performance LLM inference engine for diverse applications.
LLM training in simple, raw C/CUDA
Instant neural graphics primitives: lightning fast NeRF and more
CUDA Kernel Benchmarking Library
Tile primitives for speedy kernels
[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.
Efficient GPU kernels for block-sparse matrix multiplication and convolution
how to optimize some algorithm in cuda.
CUDA Library Samples
Causal depthwise conv1d in CUDA, with a PyTorch interface
Mirage Persistent Kernel: Compiling LLMs into a MegaKernel
Distributed multigrid linear solver library on GPU
NCCL Tests
DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling
cuVS - a library for vector search and clustering on the GPU
[ARCHIVED] Cooperative primitives for CUDA C++. See https://github.com/NVIDIA/cccl
DeepEP: an efficient expert-parallel communication library