meta-pytorch / tritonbenchLinks
Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.
☆214Updated this week
Alternatives and similar repositories for tritonbench
Users that are interested in tritonbench are comparing it to the libraries listed below
Sorting:
- Applied AI experiments and examples for PyTorch☆292Updated last week
- A Quirky Assortment of CuTe Kernels☆435Updated this week
- Cataloging released Triton kernels.☆252Updated 7 months ago
- ☆234Updated last week
- PyTorch bindings for CUTLASS grouped GEMM.☆110Updated 3 months ago
- ☆88Updated 9 months ago
- ☆229Updated last year
- A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.☆274Updated this week
- Collection of kernels written in Triton language☆152Updated 4 months ago
- extensible collectives library in triton☆88Updated 5 months ago
- A collection of memory efficient attention operators implemented in the Triton language.☆278Updated last year
- Fast low-bit matmul kernels in Triton☆356Updated last week
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆261Updated last month
- ☆110Updated last year
- kernels, of the mega variety☆481Updated 3 months ago
- Fast Hadamard transform in CUDA, with a PyTorch interface☆224Updated last year
- ☆111Updated 8 months ago
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆113Updated last year
- DeeperGEMM: crazy optimized version☆70Updated 3 months ago
- ☆98Updated 11 months ago
- Fastest kernels written from scratch☆318Updated 5 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆95Updated 2 months ago
- This repository contains the experimental PyTorch native float8 training UX☆224Updated last year
- ring-attention experiments☆149Updated 10 months ago
- TritonParse: A Compiler Tracer, Visualizer, and mini-Reproducer(WIP) for Triton Kernels☆146Updated this week
- Framework to reduce autotune overhead to zero for well known deployments.☆81Updated 2 weeks ago
- ☆81Updated 7 months ago
- A lightweight design for computation-communication overlap.☆160Updated last week
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆211Updated 3 weeks ago
- 🚀 Collection of components for development, training, tuning, and inference of foundation models leveraging PyTorch native components.☆209Updated last week