2010•Microcomputer applicationsRequires access

Mixed CUDA、MPI and OpenMP in Three Mode Parallel Programming

Qingkun Liu

Abstract

Using CUDA + MPI + OpenMP parallel programming model of the three to achieve the coarse-grained parallelbetween nodes,node fine-grained parallelism as well as within the GPU as a parallel computing devices CUDA programming model.This new three parallel hybrid programming way SMP cluster provides a more efficient parallel strategy.This article discusses the three parallel programming environment to build fast and multi-granular hybrid parallel programmingmethod,and in a number of nodes in the cluster environment,the completion of testing.

About this research paper

What this paper is about

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Key concepts: CUDA, Computer science, Parallel computing, Parallel programming model, Programming paradigm, Cluster (spacecraft), GPU cluster, Parallel algorithm