Min Si's repositories
openshmem-specification
OpenSHMEM Application Programming Interface
darshan
Darshan I/O characterization tool
dummy_collectives
A minimum demo for PyTorch c10d extension APIs
FAQs
Frequently Asked Questions for the reproducibility initiative of the SC Conference
gloo
Collective communications library with various primitives for multi-machine training.
libfabric
Open Fabric Interfaces
mlnx-tools
Mellanox userland tools and scripts
nccl
Optimized primitives for collective multi-GPU communication
nccl-tests
NCCL Tests
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.
pytorch
Tensors and Dynamic neural networks in Python with strong GPU acceleration
rccl
ROCm Communication Collectives Library (RCCL)
SOS
Sandia OpenSHMEM is an implementation of the OpenSHMEM specification over multiple Networking APIs, including Portals 4 and the Open Fabric Interface (OFI). Please click on the Wiki tab for help with building and using SOS.
torch_ucc
Pytorch process group third-party plugin for UCC
tutorials
PyTorch tutorials.
ucc
Unified Communication Collectives Library
yaksa
Yaksa: High-performance Noncontiguous Data Management