Sergei-Lebedev

nccl

Optimized primitives for collective multi-GPU communication

param

PArametrized Recommendation and Ai Model benchmark is a repository for development of numerous uBenchmarks as well as end to end nets for evaluation of training and inference platforms.

agile-development-course

Course on Agile Development (in Russian)

ComputationalMathCourse

devtools-course

Course on Development Tools (in Russian)

dlrm

An implementation of a deep learning recommendation model (DLRM)

gloo

Collective communications library with various primitives for multi-machine training.

horovod

Distributed training framework for TensorFlow, Keras, and PyTorch.

ucc

Unified Communication Collectives Library

lapack

LAPACK development repository

mccl

nccl-rdma-sharp-plugins

RDMA and SHARP plugins for nccl library

ompi

Open MPI main development repository

pcx

Persistent Collectives X- A collective communication library for high performance, low cost persistent collectives over RDMA devices.

PowerToys

Windows system utilities to maximize productivity

practice1-devtools

Practice, Day 1 -- Development Tools

pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

sparse-tools

SparseMM

Sparse matrix multiplication in CCS format

tensorflow

An Open Source Machine Learning Framework for Everyone

torch-ucc

pytorch ucc plugin

torch_ucc

Pytorch process group third-party plugin for UCC

ucx

Unified Communication X (mailing list - https://elist.ornl.gov/mailman/listinfo/ucx-group):