jundaf2

ThunderKittens

Tile primitives for speedy kernels

MatmulTutorial

A Easy-to-understand TensorOp Matmul Tutorial

cutlass-kernels

remote-dataloader

PyTorch DataLoader processed in multiple remote computation machines for heavy data processings

git-filter-repo

Quickly rewrite git repository history (filter-branch replacement)

opencv

Open Source Computer Vision Library

perftest

Infiniband Verbs Performance Tests

bbpe

BPE from byte, for real

nccl-rdma-sharp-plugins

RDMA and SHARP plugins for nccl library

minbpe

Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.

depyf

depyf is a tool to help you understand and adapt to PyTorch compiler torch.compile.

marlin

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

rfcs

PyTorch RFCs (experimental)

python-bpe

Byte Pair Encoding for Python!

dataloader-benchmarks

DL Dataloader Benchmarks

tokenizers

💥 Fast State-of-the-Art Tokenizers optimized for Research and Production

CPU-Free-model

Source code for the CPU-Free model - a fully autonomous execution model for multi-GPU applications that completely excludes the involvement of the CPU beyond the initial kernel launch.

multi-gpu-programming-models

Examples demonstrating available options to program multiple GPUs in a single node or a cluster

tutorial-multi-gpu

Efficient Distributed GPU Programming for Exascale, an SC/ISC Tutorial

bcl

The Berkeley Container Library

gdrcopy

A fast GPU memory copy library based on NVIDIA GPUDirect RDMA technology

monolith

ByteDance's Recommendation System

mlir-air

ByteMLPerfWeb

cuda_scheduling_examiner_mirror

A tool for examining GPU scheduling behavior.

awesome-courses

:books: List of awesome university courses for learning Computer Science!

NCCL

Sample examples of how to call collective operation functions on multi-GPU environments. A simple example of using broadcast, reduce, allGather, reduceScatter and sendRecv operations.

open-gpu-kernel-modules

NVIDIA Linux open GPU kernel module source

gpumembench

A GPU benchmark suite for assessing on-chip GPU memory bandwidth

detectron2

Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.