Uccl
UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)
About
UCCL is designed to enhance communication efficiency among GPUs, focusing on collective operations, peer-to-peer transfers, and extensible performance across heterogeneous hardware. It aims to provide flexibility and portability for machine learning applications, allowing seamless integration with various GPU and networking vendors. The library includes components like UCCL-collective for optimized collective operations, UCCL-P2P for advanced data transfers, and UCCL-EP for running deep learning tasks on diverse hardware setups.
Open Source Health
- Stars
- 1,513
- Forks
- 172
- License
- Apache-2.0
- Last commit
- 27 days ago
Related Categories
Vendor
UCCL
Next-generation GPU communication
Quick Links
Open Source
Related Products
The Algorithm
Source code for the X Recommendation Algorithm
Shared categories
Recommenders
TensorFlow Recommenders is a library for building recommender system models using TensorFlow.
Shared categories
RPG_KDD2025
This repository provides the code for implementing RPG described in our KDD'25 paper "Generating Long Semantic IDs in Parallel for Recommendation".
Shared categories
Pairec
A Go web framework for quickly building recommendation online services based on JSON configuration.
Shared categories
Textrank
TextRank implementation for Python 3.
Shared categories