Abstract/Description

Networks of Workstations (NOWs) based on various operating systems are globally accepted as the standard for computing environments in the IT industry. To harness the tremendous potential for computing capability represented by these NOWs, various new tools are being developed. Parallel Virtual Machine (PVM) and Message Passing Interface (MPI) have existed on UNIX workstations for some time, and are maturing in their capability for handling Distributed Parallel Processing (DPP). This research is aimed to explore each of these two vehicles for DPP, considering capability, ease of use, and availability, and compares their distinguishing features; and also explores programmer interface and their utilisation for solving real world parallel processing applications. The research conducted herein concludes that each API has its unique features of strength, hence has potential to remain active into the foreseeable future. This work recommends a potential research issue, that is, to study the feasibility of creating a programming environment that allows access to the virtual machine features of PVM and the message passing features of MPI.

Location

Crystal Ball Room A, Hotel Pearl Continental, Karachi, Pakistan

Session Theme

Poster Session B: Computing Systems [CS]

Session Type

Poster Session

Session Chair

Dr. Anila Usman

Start Date

28-8-2005 2:50 PM

End Date

28-8-2005 3:10 PM

Share

COinS
 
Aug 28th, 2:50 PM Aug 28th, 3:10 PM

A comparative study and analysis of PVM and MPI for parallel and distributed systems

Crystal Ball Room A, Hotel Pearl Continental, Karachi, Pakistan

Networks of Workstations (NOWs) based on various operating systems are globally accepted as the standard for computing environments in the IT industry. To harness the tremendous potential for computing capability represented by these NOWs, various new tools are being developed. Parallel Virtual Machine (PVM) and Message Passing Interface (MPI) have existed on UNIX workstations for some time, and are maturing in their capability for handling Distributed Parallel Processing (DPP). This research is aimed to explore each of these two vehicles for DPP, considering capability, ease of use, and availability, and compares their distinguishing features; and also explores programmer interface and their utilisation for solving real world parallel processing applications. The research conducted herein concludes that each API has its unique features of strength, hence has potential to remain active into the foreseeable future. This work recommends a potential research issue, that is, to study the feasibility of creating a programming environment that allows access to the virtual machine features of PVM and the message passing features of MPI.