GitHub Cuda Monthly Trending Monthly Trending of Cuda in GitHub Sun, 17 May 2026 01:52:24 GMT http://mshibanami.github.io/GitHubTrendingRSS deepseek-ai/DeepGEMM https://github.com/deepseek-ai/DeepGEMM <p>DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling</p><hr> deepseek-ai/DeepEP https://github.com/deepseek-ai/DeepEP <p>DeepEP: an efficient expert-parallel communication library</p><hr> alibaba/rtp-llm https://github.com/alibaba/rtp-llm <p>RTP-LLM: Alibaba's high-performance LLM inference engine for diverse applications.</p><hr> karpathy/llm.c https://github.com/karpathy/llm.c <p>LLM training in simple, raw C/CUDA</p><hr> NVIDIA/AMGX https://github.com/NVIDIA/AMGX <p>Distributed multigrid linear solver library on GPU</p><hr> NVIDIA/nvbench https://github.com/NVIDIA/nvbench <p>CUDA Kernel Benchmarking Library</p><hr> NVIDIA/CUDALibrarySamples https://github.com/NVIDIA/CUDALibrarySamples <p>CUDA Library Samples</p><hr> NVIDIA/cub https://github.com/NVIDIA/cub <p>[ARCHIVED] Cooperative primitives for CUDA C++. See https://github.com/NVIDIA/cccl</p><hr> NVIDIA/nccl-tests https://github.com/NVIDIA/nccl-tests <p>NCCL Tests</p><hr> princeton-vl/lietorch https://github.com/princeton-vl/lietorch <p style="color:#586069;"><em>No description/README provided.</em></p> mirage-project/mirage https://github.com/mirage-project/mirage <p>Mirage Persistent Kernel: Compiling LLMs into a MegaKernel</p><hr> NVlabs/instant-ngp https://github.com/NVlabs/instant-ngp <p>Instant neural graphics primitives: lightning fast NeRF and more</p><hr> ashawkey/diff-gaussian-rasterization https://github.com/ashawkey/diff-gaussian-rasterization <p style="color:#586069;"><em>No description/README provided.</em></p> NVIDIA/cuopt https://github.com/NVIDIA/cuopt <p>GPU accelerated decision optimization</p><hr> BBuf/how-to-optim-algorithm-in-cuda https://github.com/BBuf/how-to-optim-algorithm-in-cuda <p>how to optimize some algorithm in cuda.</p><hr> rahul-goel/fused-ssim https://github.com/rahul-goel/fused-ssim <p>Lightning fast differentiable SSIM.</p><hr> thu-ml/SageAttention https://github.com/thu-ml/SageAttention <p>[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.</p><hr> HazyResearch/ThunderKittens https://github.com/HazyResearch/ThunderKittens <p>Tile primitives for speedy kernels</p><hr>