全部/科技/Cuda / 月榜

GitHub · Cuda / 月榜

26分钟前更新
HISTORY近 30 天历史柱高表示当天去重热搜数量
1307
09/08—10/07 有历史数据
  • 018634
    DeepGEMM
    DeepGEMM: clean and efficient BLAS kernel library on GPUdeepseek-ai
  • 026552
    flashinfer
    FlashInfer: Kernel Library for LLM Servingflashinfer-ai
  • 031.0万
    DeepEP
    DeepEP: an efficient expert-parallel communication librarydeepseek-ai
  • 043960
    SageAttention
    [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.thu-ml
  • 053745
    ThunderKittens
    Tile primitives for speedy kernelsHazyResearch
  • 062241
    cugraph
    cuGraph - RAPIDS Graph Analytics Libraryrapidsai
  • 07860
    cuvs
    cuVS - a library for vector search and clustering on the GPUNVIDIA
  • 081053
    cuopt
    GPU accelerated decision optimizationNVIDIA
  • 093302
    how-to-optim-algorithm-in-cuda
    how to optimize some algorithm in cuda.BBuf
  • 101671
    nccl-tests
    NCCL TestsNVIDIA
  • 111.8万
    instant-ngp
    Instant neural graphics primitives: lightning fast NeRF and moreNVlabs
  • 122534
    mirage
    Mirage Persistent Kernel: Compiling LLMs into a MegaKernelmirage-project
  • 133.1万
    llm.c
    LLM training in simple, raw C/CUDAkarpathy