SUARA: A scalable universal allreduce communication algorithm for acceleration of parallel deep learning applications

TitleSUARA: A scalable universal allreduce communication algorithm for acceleration of parallel deep learning applications
Publication TypeJournal Article
Year of Publication2024
AuthorsNuriyev, E., R. R. Manumachu, S. Aseeri, M. Verma, and A. Lastovetsky
Journal TitleJournal of Parallel and Distributed Computing
Volume183
Pages15
Journal Date01/2024
DOI10.1016/j.jpdc.2023.104767
AttachmentSize
jpdc-suara.pdf2.46 MB