ziyuhuang123

ziyuhuang123's Stars

• gty111/GEMM_WMMA

  GEMM by WMMA (tensor core)

  Language:Cuda87

• meta-llama/llama

  Inference code for Llama models

  Language:Python55.5k9.5k

• TiledTensor/TiledCUDA

  TiledCUDA is a highly efficient kernel template library designed to elevate CUDA C’s level of abstraction for processing tiles.

  Language:C++1149

• ColfaxResearch/cfx-article-src

  Language:C++158

• galeselee/microbenchmark

  Some microbenchmark practices

  Language:Cuda1

• te42kyfo/gpu-benches

  collection of benchmarks to measure basic GPU capabilities

  Language:Jupyter Notebook24138

• HazyResearch/ThunderKittens

  Tile primitives for speedy kernels

  Language:Cuda1.5k57

• forhaoliu/ringattention

  Transformers with Arbitrarily Large Context

  Language:Python61247

• NVIDIA/cutlass

  CUDA Templates for Linear Algebra Subroutines

  Language:C++5.4k902

• google-research/maxvit

  [ECCV 2022] Official repository for "MaxViT: Multi-Axis Vision Transformer". SOTA foundation models for classification, detection, segmentation, image quality, and generative modeling...

  Language:Jupyter Notebook43530

• d2l-ai/d2l-zh

  《动手学深度学习》：面向中文读者、能运行、可讨论。中英文版被70多个国家的500多所大学用于教学。

  Language:Python61.3k10.9k

• microsoft/cusync

  Language:C++163

• nuno-azevedo/floyd-warshall-mpi

  Parallel Computing - Floyd-Warshall MPI

  Language:TeX35

• vllm-project/vllm

  A high-throughput and memory-efficient inference and serving engine for LLMs

  Language:Python26.7k3.9k

• curtisseizert/CUDASieve

  A GPU accelerated implementation of the sieve of Eratosthenes

  Language:Cuda5915

• KnowingNothing/Domino

  Language:Python8

• Accelergy-Project/timeloop-accelergy-exercises

  Exercises for exploring the Fibertree, Timeloop and Accelergy tools

  Language:Jupyter Notebook7828

• pku-liang/TileFlow

  TileFlow is a performance analysis tool based on Timeloop for fusion dataflows

  Language:C++535

• gpgpu-sim/gpgpu-sim_distribution

  GPGPU-Sim provides a detailed simulation model of contemporary NVIDIA GPUs running CUDA and/or OpenCL workloads. It includes support for features such as TensorCores and CUDA Dynamic Parallelism as well as a performance visualization tool, AerialVisoin, and an integrated energy model, GPUWattch.

  Language:C++1.1k500

• maffinnn/consistent_hash

  Language:Go1

• tspeterkim/flash-attention-minimal

  Flash Attention in ~100 lines of CUDA (forward pass only)

  Language:Cuda55749

• amd/xdna-driver

  Language:C28939

• crosetto/cupq

  a CUDA implementation of a priority queue

  Language:C++804

• MatanHamilis/one_stencil

  Multiple 1-stencil implementations using nvidia cuda.

  Language:Cuda134

• apache/tvm

  Open deep learning compiler stack for cpu, gpu and specialized accelerators

  Language:Python11.6k3.4k

• IST-DASLab/marlin

  FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

  Language:Python55845

• parasailteam/cusync

  Language:C++52

• doongz/aics

  智能计算系统 AI Computing Systems 陈云霁

  Language:C++10815

• KnowingNothing/MatmulTutorial

  A Easy-to-understand TensorOp Matmul Tutorial

  Language:C++25828

• cli99/llm-analysis

  Latency and Memory Analysis of Transformer Models for Training and Inference

  Language:Python33739