Principles of Parallel Computers and Some Impacts on Their Programming Models
In this paper we briefly outline some principles of parallel architectures and discuss several impacts on their programming models. First, parallel computers are generally classified. A description of the most important classes — multiprocessors and massively parallel systems — follows, with some details about chosen machines. The corresponding programming models for shared-memory and distributed-memory architectures are introduced. The special relationship between machine architecture and efficient parallel programming is emphasized here. The paper concludes with some hints for the software developer about where to use which parallel programming model.
KeywordsParallel Computer Message Passing Multiprocessor System Parallel Programming Model Master Processor
Unable to display preview. Download preview PDF.
- Hitachi Product Sheet SR2201, Hitachi Corp.Google Scholar
- KSR Parallel Programming-Manual, (Kendall Square Research Corporation, Waltham Massachusetts 1993)Google Scholar
- Cross-Bar und Snooping-Bus, iX no. 7, (1995)Google Scholar
- F. Bellosa, Memory Concious Scheduling and Processor Allocation on NUMA Architectures, TR-I4-5-95, (Computer Science Dep., Univ. Erlangen 1995)Google Scholar
- J. Dongarra, Performance of Various Computers using Standard Linear Equation Software, Tech. Report January 7, 1995, (Computer Science Department, Univ. of Tennessee, Knoxville 1995)Google Scholar
- C.A.R. Hoare, Communicating Sequential Processes, (Prentice Hall, New York 1989)Google Scholar
- K. Hwang, Advanced Computer Architecture, (McGraw-Hill, New York 1993)Google Scholar