xlite-dev / ffpa-attnLinks
⚡️FFPA: Extend FlashAttention-2 with Split-D, achieve ~O(1) SRAM complexity for large headdim, 1.8x~3x↑ vs SDPA.
☆186Updated last month
Alternatives and similar repositories for ffpa-attn
Users that are interested in ffpa-attn are comparing it to the libraries listed below
Sorting:
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆183Updated 4 months ago
- ☆86Updated 3 months ago
- ☆96Updated 9 months ago
- Implement Flash Attention using Cute.☆87Updated 6 months ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆80Updated last month
- Examples of CUDA implementations by Cutlass CuTe☆197Updated 4 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆38Updated 3 months ago
- A lightweight design for computation-communication overlap.☆143Updated this week
- Standalone Flash Attention v2 kernel without libtorch dependency☆110Updated 9 months ago
- ☆135Updated last year
- A collection of memory efficient attention operators implemented in the Triton language.☆272Updated last year
- ☆77Updated last month
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆128Updated 2 months ago
- ☆60Updated 2 months ago
- flash attention tutorial written in python, triton, cuda, cutlass☆377Updated last month
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆71Updated 10 months ago
- ☆148Updated 5 months ago
- A Easy-to-understand TensorOp Matmul Tutorial☆364Updated 9 months ago
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆109Updated 11 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆94Updated 6 years ago
- ☆123Updated 6 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆99Updated 3 weeks ago
- ☆212Updated 11 months ago
- Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.☆136Updated this week
- ☆141Updated 3 months ago
- ☆139Updated last year
- ☆71Updated last month
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆252Updated 7 months ago
- Puzzles for learning Triton, play it with minimal environment configuration!☆367Updated 6 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆92Updated 3 weeks ago