Alternatives and similar repositories for flash-attention

Users that are interested in flash-attention are comparing it to the libraries listed below

• InternLM / turbomind
  ☆96Updated 9 months ago
• infinigence / Semi-PD
  A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.
  ☆123Updated 2 weeks ago
• ademeure / DeeperGEMM
  DeeperGEMM: crazy optimized version
  ☆74Updated 8 months ago
• usyd-fsalab / fp6_llm
  An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).
  ☆276Updated 5 months ago
• Ascend / triton-ascend
  Triton adapter for Ascend. Mirror of https://gitee.com/ascend/triton-ascend
  ☆98Updated this week
• feifeibear / LLMRoofline
  Compare different hardware platforms via the Roofline Model for LLM inference tasks.
  ☆120Updated last year
• infinigence / FlashOverlap
  A lightweight design for computation-communication overlap.
  ☆209Updated 2 weeks ago
• madsys-dev / deepseekv2-profile
  ☆153Updated 10 months ago
• AlibabaPAI / FLASHNN
  ☆104Updated last year
• microsoft / vattention
  Dynamic Memory Management for Serving LLMs without PagedAttention
  ☆454Updated 7 months ago
• xlite-dev / HGEMM
  ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.
  ☆142Updated 8 months ago
• AlibabaResearch / flash-llm
  Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity
  ☆231Updated 2 years ago
• LLMServe / SwiftTransformer
  High performance Transformer implementation in C++.
  ☆148Updated 11 months ago
• xlite-dev / ffpa-attn
  🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.
  ☆242Updated last month
• flagos-ai / FlagCX
  FlagCX is a scalable and adaptive cross-chip communication library.
  ☆166Updated this week
• stepfun-ai / StepMesh
  ☆338Updated last week
• microsoft / sarathi-serve
  A low-latency & high-throughput serving engine for LLMs
  ☆464Updated this week
• triton-lang / kernels
  ☆100Updated last year
• perplexityai / pplx-kernels
  Perplexity GPU Kernels
  ☆552Updated 2 months ago
• tgale96 / grouped_gemm
  PyTorch bindings for CUTLASS grouped GEMM.
  ☆139Updated 7 months ago
• flagos-ai / FlagAttention
  A collection of memory efficient attention operators implemented in the Triton language.
  ☆287Updated last year
• DeepLink-org / DLSlime
  DLSlime: Flexible & Efficient Heterogeneous Transfer Toolkit
  ☆91Updated last week
• HandH1998 / QQQ
  QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.
  ☆153Updated 4 months ago
• Victarry / PP-Schedule-Visualization
  Pipeline Parallelism Emulation and Visualization
  ☆74Updated this week
• PKU-SEC-Lab / HybriMoE
  [DAC'25] Official implement of "HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"
  ☆96Updated 3 weeks ago
• flashinfer-ai / cutlass-viz
  ☆65Updated 8 months ago
• tlc-pack / libflash_attn
  Standalone Flash Attention v2 kernel without libtorch dependency
  ☆113Updated last year
• EfficientMoE / MoE-Infinity
  PyTorch library for cost-effective, fast and easy serving of MoE models.
  ☆273Updated 2 months ago
• alibaba / easydist
  Automated Parallelization System and Infrastructure for Multiple Ecosystems
  ☆82Updated last year
• antgroup / DeepXTrace
  DeepXTrace is a lightweight tool for precisely diagnosing slow ranks in DeepEP-based environments.
  ☆85Updated 3 weeks ago