cutlass 开源项目推荐（60 个）

bytedance /flux

A fast communication-overlapping library for tensor/expert parallelism on GPUs.

NVlabs /vibetensor

Our first fully AI generated deep learning system

coderonion /awesome-cuda-and-hpc

🚀🚀🚀 该仓库列出了一些优秀的公共 CUDA、cuda-python、cuBLAS、cuDNN、CUTLASS、TensorRT、TensorRT-LLM、Triton、TVM、MLIR、PTX 和高性能计算（HPC）项目。

DD-DuDa /Cute-Learning

Examples of CUDA implementations by Cutlass CuTe

leimao /CUTLASS-Examples

CUTLASS and CuTe Examples

psmarter /CUDA-Practice

CUDA编程练习项目-Hands-on CUDA kernels and performance optimization, covering GEMM, FlashAttention, Tensor Cores, CUTLASS, quantization, KV cache, NCCL, and profiling.

yassa9 /dvlt.cu

Suckless no dependencies CUDA/C++ port of NVIDIA's DVLT, feed it images, get a 3D point cloud + camera poses. NO python. One fast 5MB binary.

bikrammajhi /100-days-of-GPU

This is my 🔥 100 Days of GPU — a wild, hands-on journey through CUDA/CUTLASS kernels, Triton spells, and PTX sorcery.

Bruce-Lee-LY /flash_attention_inference

Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.

lavawolfiee /mini-flash-attention

Minimal FlashAttention in CUDA C++/CuTe: readable WMMA/CuTe kernels, no NxN workspace, up to 4.5x faster than naive PyTorch

YashasSamaga /ConvolutionBuildingBlocks

GEMM and Winograd based convolutions using CUTLASS

lna-lab /blackwell-geforce-nvfp4-gemm

NVFP4 inference on Blackwell GeForce (RTX 5090/5080/5070 Ti/RTX PRO 6000) — SM120 patches for vLLM + FlashInfer + CUTLASS. 175 tok/s on Qwen3.6-35B MoE.

sgl-project /whl

SGLang Kernel Wheel Index

Bruce-Lee-LY /cuda_auto_tune

NCU-driven iterative optimization workflow for CUDA/CUTLASS/Triton/CuTe DSL kernels.

yester31 /Cutlass_EX

study of cutlass

qdLMF /LightGlue-with-FlashAttentionV2-TensorRT

用于 LightGlue 的 headdim-64 flashattentionv2 TensorRT 插件的 cutlass cute 实现。在搭载 TensorRT 8.5.2 的 Jetson Orin NX 8GB 上运行。

Bruce-Lee-LY /cutlass_gemm

Multiple GEMM operators are constructed with cutlass to support LLM inference.

chiendb97 /oasr

High-performance speech-to-text (ASR) inference and serving engine for NVIDIA GPUs: Whisper, Conformer, Zipformer, RNN-T, Paraformer, Qwen2-Audio. OpenAI-compatible API, streaming, CUDA kernels.

minho-fan /kernel-opt-agent

CUDA/Triton/CuTe DSL/CUTLASS kernel optimization agent

RishiShah99 /lethe

native blackwell training backward for gated linear recurrences, and a verifier that catches broken llm-generated gpu kernels