In this paper we describe the approach we are pursuing on coping with parallelism in symbolic applications. Our purpose is to build a unified and user friendly parallel symbolic programming environment suited to Distributed Memory Parallel Systems (DMPS). This system is based on the Sabot's Paralation model (in particular its Lisp implementation, Paralation Lisp), whose main feature is to clearly differentiate the computational aspect from the communication one. In our opinion this distinction allows a clearer and cleaner exploitation of the opportunities of parallelism in symbolic applications. The simple semantics of the Paralation model parallel constructs seems to us well suited to run on a loosely coupled Lisp interpreter community. We chose to implement the Paralation model extensions on top of the Parallel Virtual Machine (PVM) package because this tool allows to easily use a set of Lan interconnected powerful workstations as a general purpose parallel computational resource.
A PVM Based Paralation Lisp Implementation on DMPS
C Di Napoli;M Giordano;M Mango Furnari
1994
Abstract
In this paper we describe the approach we are pursuing on coping with parallelism in symbolic applications. Our purpose is to build a unified and user friendly parallel symbolic programming environment suited to Distributed Memory Parallel Systems (DMPS). This system is based on the Sabot's Paralation model (in particular its Lisp implementation, Paralation Lisp), whose main feature is to clearly differentiate the computational aspect from the communication one. In our opinion this distinction allows a clearer and cleaner exploitation of the opportunities of parallelism in symbolic applications. The simple semantics of the Paralation model parallel constructs seems to us well suited to run on a loosely coupled Lisp interpreter community. We chose to implement the Paralation model extensions on top of the Parallel Virtual Machine (PVM) package because this tool allows to easily use a set of Lan interconnected powerful workstations as a general purpose parallel computational resource.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


