湖南大学学报(自然科学版)Issue(8):100-107,8.
一种面向不可靠网络的快速 RDMA 通信方法∗
A Fast RDMA Offload Method for Unreliable Interconnection Networks
摘要
Abstract
Large data RDMA (Remote Data Memory Access)transport is the most commonly used par-allel communication mode for parallel computers,which has great impact on the whole system perform-ance.As the system size increases,the fault-tolerate architecture design faces new challenges.The inter-connection network usually uses the adaptive routing mode and becomes more unreliable.This paper pro-posed a fast RDMA offload method for unreliable interconnection networks,which can be efficiently imple-mented on the NIC hardware and provides reliable RDMA communication for upper driver and programs. Compared with the traditional approaches,the hardware overhead is greatly reduced.Another benefit is that it can partially retransmit the fault data,which greatly reduces the whole RDMA delay.Simulation results show that the RDMA delay is greatly reduced,compared with the traditional methods.关键词
远程内存访问/RDMA/MPI/滑动窗口Key words
remote data memory access/RDMA/MPI/sliding window approach分类
信息技术与安全科学引用本文复制引用
王绍刚,徐炜遐,吴丹,庞征斌,夏军..一种面向不可靠网络的快速 RDMA 通信方法∗[J].湖南大学学报(自然科学版),2015,(8):100-107,8.基金项目
国家自然科学基金资助项目(61202024,61202126),National Natural Science Foundation of China(61202024,61202126) (61202024,61202126)