ghZCCL: Advancing GPU-aware Collective Communications with Homomorphic Compression
Open Access
- 8 June 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 43-56
Abstract
No abstract availableFunding Information
- DOE (DE-AC02-06CH11357,DE-SC0024207)
- NSF (OAC-2104023,OAC-2311875)
This publication has 34 references indexed in Scilit:
- Exploring Wavelet Transform Usages for Error-bounded Scientific Data Compression2023
- Accelerating MPI All-to-All Communication with Online Compression on Modern GPU Clusters2022
- Designing High-Performance MPI Libraries with On-the-fly Compression for Modern GPU Clusters2021
- Optimizing Error-Bounded Lossy Compression for Scientific Data by Dynamic Spline Interpolation2021
- Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation2019
- SALaR: Scalable and Adaptive Designs for Large Message Reduction Collectives2018
- Exploration of Lossy Compression for Application-Level Checkpoint/Restart2015
- Direct numerical simulations of autoignition in stratified dimethyl-ether (DME)/air turbulent mixturesCombustion and Flame, 2015
- A direct numerical simulation study of turbulence and flame structure in transverse jets analysed in jet-trajectory based coordinatesJournal of Fluid Mechanics, 2012
- Bandwidth optimal all-reduce algorithms for clusters of workstationsJournal of Parallel and Distributed Computing, 2009