A pilot study to compare programming effort for two parallel programming models

作     者：Hochstein, Lorin Basili, Victor R. Vishkin, Uzi Gilbert, John 

作者机构：Univ Nebraska Dept Comp Sci & Engn Lincoln NE 68583 USA Univ Maryland Dept Comp Sci College Pk MD 20742 USA Univ Maryland Inst Adv Comp Studies College Pk MD 20742 USA Univ Calif Santa Barbara Dept Comp Sci Santa Barbara CA 93106 USA 

出 版 物：《JOURNAL OF SYSTEMS AND SOFTWARE》 (系统与软件杂志)

年 卷 期：2008年第81卷第11期

页      面：1920-1930页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Department of Energy [DE-FG02-04ER25633, DE-CFC02-01 ER25489] Air Force Rome Labs [FA8750-05-1-0100] 

主　　题：MPI XMT Message-passing PRAM Empirical study Parallel programming Effort 

摘      要：Context: Writing software for the current generation of parallel systems requires significant programmer effort, and the community is seeking alternatives that reduce effort while still achieving good performance. Objective: Measure the effect of parallel programming models (message-passing vs. PRAM-like) oil programmer effort. Design, setting. and subjects: One group of subjects implemented sparse-matrix dense-vector multiplication using message-passing (MPI), and a second group solved the same problem using a PRAM-like model (XMTC). The subjects were students in two graduate-level classes: one class was taught MPI and the other was taught XMTC. Main outcome measures: Development time, program correctness. Results: Mean XMTC development time was 4.8 h less than mean MPI development time (95% confidence interval, 2.0-7.7), a 46% reduction. XMTC programs were more likely to be correct, but the difference in correctness rates was not statistically significant (p = .16). Conclusions: XMTC Solutions for this particular problem required less effort than MPI equivalents, but further Studies are necessary which examine different types of problems and different levels of programmer experience. (C) 2008 Elsevier Inc. All rights reserved.