Flash Attention in ~100 lines of CUDA (forward pass only)
☆1,084Dec 30, 2024Updated last year
Alternatives and similar repositories for flash-attention-minimal
Users that are interested in flash-attention-minimal are comparing it to the libraries listed below
Sorting:
- flash attention tutorial written in python, triton, cuda, cutlass☆490Jan 20, 2026Updated last month
- Implement Flash Attention using Cute.☆102Dec 17, 2024Updated last year
- FlashInfer: Kernel Library for LLM Serving☆5,101Updated this week
- Tile primitives for speedy kernels☆3,202Feb 24, 2026Updated last week
- a minimal cache manager for PagedAttention, on top of llama3.☆136Aug 26, 2024Updated last year
- how to optimize some algorithm in cuda.☆2,841Feb 28, 2026Updated last week
- Fastest kernels written from scratch☆550Sep 18, 2025Updated 5 months ago
- ☆262Jul 11, 2024Updated last year
- This is a series of GPU optimization topics. Here we will introduce how to optimize the CUDA kernel in detail. I will introduce several…☆1,244Jul 29, 2023Updated 2 years ago
- Examples of CUDA implementations by Cutlass CuTe☆272Jul 1, 2025Updated 8 months ago
- 📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉☆9,815Feb 25, 2026Updated last week
- Material for gpu-mode lectures☆5,800Feb 1, 2026Updated last month
- Fast CUDA matrix multiplication from scratch☆1,080Sep 2, 2025Updated 6 months ago
- Step-by-step optimization of CUDA SGEMM☆433Mar 30, 2022Updated 3 years ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆79Aug 12, 2024Updated last year
- A Easy-to-understand TensorOp Matmul Tutorial☆409Updated this week
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆194Jan 28, 2025Updated last year
- Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruct…☆529Sep 8, 2024Updated last year
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆46Jun 11, 2025Updated 8 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆44Feb 27, 2025Updated last year
- Flash Attention in 300-500 lines of CUDA/C++☆36Aug 22, 2025Updated 6 months ago
- 🚀 Efficient implementations of state-of-the-art linear attention models☆4,474Updated this week
- Benchmark tests supporting the TiledCUDA library.☆18Nov 19, 2024Updated last year
- Ring attention implementation with flash attention☆987Sep 10, 2025Updated 5 months ago
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆5,330Updated this week
- Fast and memory-efficient exact attention☆22,460Updated this week
- CUDA Templates and Python DSLs for High-Performance Linear Algebra☆9,348Mar 2, 2026Updated last week
- 使用 cutlass 实现 flash-attention 精简版，具有教学意义☆58Aug 12, 2024Updated last year
- A throughput-oriented high-performance serving framework for LLMs☆947Oct 29, 2025Updated 4 months ago
- Fast low-bit matmul kernels in Triton☆436Feb 1, 2026Updated last month
- ☆116May 16, 2025Updated 9 months ago
- Efficient Triton Kernels for LLM Training☆6,189Updated this week
- A simple high performance CUDA GEMM implementation.☆426Jan 4, 2024Updated 2 years ago
- 📚A curated list of Awesome LLM/VLM Inference Papers with Codes: Flash-Attention, Paged-Attention, WINT8/4, Parallelism, etc.🎉☆5,040Feb 27, 2026Updated last week
- A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.☆595Aug 12, 2025Updated 6 months ago
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆1,035Sep 4, 2024Updated last year
- Optimizing SGEMM kernel functions on NVIDIA GPUs to a close-to-cuBLAS performance.☆408Jan 2, 2025Updated last year
- Development repository for the Triton language and compiler☆18,573Updated this week
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆255Feb 13, 2026Updated 3 weeks ago