xlite-dev / ffpa-attnLinks
🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.
☆223Updated 2 months ago
Alternatives and similar repositories for ffpa-attn
Users that are interested in ffpa-attn are comparing it to the libraries listed below
Sorting:
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆121Updated 5 months ago
- ☆97Updated 7 months ago
- Implement Flash Attention using Cute.☆96Updated 10 months ago
- ☆100Updated last year
- A collection of memory efficient attention operators implemented in the Triton language.☆282Updated last year
- ☆120Updated 2 months ago
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆186Updated 8 months ago
- ☆107Updated 5 months ago
- flash attention tutorial written in python, triton, cuda, cutlass☆428Updated 5 months ago
- Examples of CUDA implementations by Cutlass CuTe☆241Updated 3 months ago
- A lightweight design for computation-communication overlap.☆181Updated 2 weeks ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆41Updated 7 months ago
- ☆148Updated 7 months ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆112Updated last year
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆144Updated 2 months ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆76Updated last year
- ☆65Updated 6 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆94Updated last month
- ☆56Updated 3 months ago
- ☆150Updated 9 months ago
- A Easy-to-understand TensorOp Matmul Tutorial☆385Updated 2 weeks ago
- PyTorch bindings for CUTLASS grouped GEMM.☆125Updated 4 months ago
- ☆141Updated last year
- ☆101Updated 5 months ago
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆45Updated 4 months ago
- llm theoretical performance analysis tools and support params, flops, memory and latency analysis.☆108Updated 3 months ago
- Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.☆264Updated this week
- [DAC'25] Official implement of "HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"☆75Updated 4 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆102Updated 7 years ago
- Tutorials for writing high-performance GPU operators in AI frameworks.☆132Updated 2 years ago