Parallel discrete event simulation
Communications of the ACM - Special issue on simulation
PVM: a framework for parallel distributed computing
Concurrency: Practice and Experience
ParaSol: a multithreaded system for parallel simulation based on mobile threads
WSC '95 Proceedings of the 27th conference on Winter simulation
SimKit: a high performance logical process simulation class library in C++
WSC '95 Proceedings of the 27th conference on Winter simulation
Automatic parallelization of discrete event simulation programs
WSC '93 Proceedings of the 25th conference on Winter simulation
Iterative design of efficient simulations using Maisie
WSC '91 Proceedings of the 23rd conference on Winter simulation
CSIM: a C-based process-oriented simulation language
WSC '86 Proceedings of the 18th conference on Winter simulation
Simulation with GPSS and Gpssv
Simulation with GPSS and Gpssv
ON EXTENDING PARALLELISM TO SERIAL SIMULATORS
ON EXTENDING PARALLELISM TO SERIAL SIMULATORS
Hi-index | 0.00 |
Much of the research in parallel discrete-event simulation (PDES) has resulted in new experimental simulation languages or toolkits. Meanwhile, the simulation community continues to use existing (serial) commercial tools which are reportedly more powerful and flexible from a modeler's point of view. A possible way to increase the impact of PDES in the simulation community is to make existing simulation packages execute in parallel. Towards this end, we present a parallelization of the GPSS simulation language. We implement parallel GPSS as a GPSS-to-C++ translator and execute the transformed code with the help of the ParaSol parallel simulation system. The mapping from GPSS to ParaSol is simple because, unlike other parallel simulation systems, ParaSol is transaction oriented. On the other hand, because GPSS was not designed with parallelism in mind 'there are GPSS constructs that can behave poorly in a parallel environment. We present details on the mapping, some of the challenges we faced in this task, and key solutions that we adopted to enhance parallelism.