Construction of Hybrid MPI-OpenMP Solutions for SMP Clusters
スポンサーリンク
概要
- 論文の詳細を見る
This paper proposes a middle-grain approach to construct hybrid MPI-OpenMP solutions for SMP clusters from an existing MPI algorithm. Experiments on different cluster platforms show that our solutions exceed the solutions that are based on the de-facto MPI model in most cases, and occasionally by as much as 40% of performance. We also prove an automatic outperformance of a thread-to-thread communication model over a traditional process-to-process communication model in hybrid solutions. In addition, the paper performs a detailed analysis on the hardware and software factors affecting the performance of MPI in comparison to hybrid models.
- 一般社団法人 情報処理学会の論文
著者
-
Yoshinaga Tsutomu
Graduate School Of Information Systems University Of Electro-communications
-
Viet Ta
Graduate School Of Information Systems University Of Electro-communications
-
Abderazek Ben
Graduate School Of Information Systems University Of Electro-communications
-
Sowa Masahiro
Graduate School Of Information Systems University Of Electro-communications
-
Sowa Masahiro
Graduate School of Information Systems, University of Electro-Communications
関連論文
- Improving Linpack Performance on SMP Clusters with Asynchronous MPI Programming
- Improving Linpack Performance on SMP Clusters with Asynchronous MPI Programming(数値計算)
- Optimization for Hybrid MPI-OpenMP Programs with Thread-to-thread Communication (2004年並列/分散/協調処理に関する「青森」サマーワークショップ(SWoPP青森2004))
- Construction of Hybrid MPI-OpenMP Solutions for SMP Clusters(System Software)
- D-6-8 Hybrid Compiler-Controlled Self-Adjustable Parallelism-Independent Scheduling Algorithm for Cluster of Workstations
- An Ambiguous, Context-Free Grammar for Deterministic Parsing In Queue-Java Compiler
- Fast, Effective Instruction Generation Algorithm For Queue-Java Compiler (QJAVAC)
- Design of Producer-order Parallel Queue Processor Architecture
- Design of Producer-order Parallel Queue Processor Architecture
- Design of Producer-order Parallel Queue Processor Architecture
- A Reduced Bit-Width Instruction Set Architecture for FQM Execution in Hybrid Processor Architecture (FaRM-rq)
- Computation-Communication Overlap of Linpack on a GPU-Accelerated PC Cluster
- Design and Implementation of a Handshake Join Architecture on FPGA
- Improving Linpack Performance on SMP Clusters with Asynchronous MPI Programming
- Construction of Hybrid MPI-OpenMP Solutions for SMP Clusters
- Construction of Hybrid MPI-OpenMP Solutions for SMP Clusters