ANALISIS UNJUK KERJA KOMPUTASI DISTRIBUTED SHARED MEMORY PADA SISTEM CLUSTER KOMPUTER PERSONAL
Abstract: Parallel processing
systems that support
the shared memory
abstraction are becoming widely accepted in many areas of
computing. The shared memory or single system image of address space
abstraction provides several
advantages over the
message passing abstraction.
Distributed shared memory or DSM is a memory configuration system that can be used
to provide a
coherent shared address
space for a
workstation clusters or
loosely coupled systems which have no support for shared memory in
hardware.The purpose of this paper
is to describe
issues related to
the implementation of
DSM in Personal
Computer Cluster system. The
system is developed
under Linux operating
system and cluster middleware Parallel Virtual Machine
[PVM]. In this research, an object-based Distributed Shared Memory
(DSM) system called
Adsmith is used
on the top
of PVM. An application
program for A Traveling Salesman Problem (TSP) is tested using different
scenarios to get average execution time and speedup of the computation. Each
scenario represents a different number of node cluster involved for
computation. The results show that
programs developed with Adsmith can
achieve better performance
comparable to that
developed with PVM in
high computation environment.
Penulis: Ferrianto Gozali
& Dimas Lagusto
Kode Jurnal: jptlisetrodd050002
