E ciency Evaluation of Selected Parallelization Tools
نویسندگان
چکیده
Although the processing power of uniprocessor sys tems is signi cantly increasing every year there is a never ending quest for peak performance The so called grand challenges in computing require more computing power than a single uniprocessor machine can ever o er Recent advances in parallel processing however has brought the solution of grand challenges into reach What is usually neglected when reporting on peak performance and Tera ops is the e ort which would be necessary for implementing an e cient application Handling all the communication which is needed in a parallel algorithm is probably not the most di cult part for the programmer but often the most annoy ing part To support or even automate this task a variety of programming tools and libraries has been developed An interesting question is now to investigate and compare the performance obtained when using such tools and the e ort for using these tools For that pur pose an evaluation project has been started on a clus ter of nine workstations using the ve most commonly used tools namely PVM x Version PVM x Version p Version patches Express Version and Linda Version The Numerical Aerodynamic Simulation NAS parallel benchmark suite which consists of ve ker nels and three simulated application benchmarks was used for the performance evaluation Four of the ker nels were implemented on the cluster An Embar rassingly Parallel Benchmark to get used with the
منابع مشابه
Storage Mapping Optimization for Parallel Programs
Data dependences are known to hamper e cient parallelization of programs. Memory expansion is a general method to remove dependences in assigning distinct memory locations to dependent writes. Parallelization via memory expansion requires both moderation in the expansion degree and e ciency at run-time. We present a general storage mapping optimization framework for imperative programs, applica...
متن کاملExperiences in Developing Numerical Subprograms on a Parallel Shared Memory Computer
This Technical Report describes work carried out on the parallel shared memory SGI Power Challenge system with MIPS R MHz processors under OS IRIX at UNI C Denmark The main thrust of this work was an investigation of the scalability and e ciency of a range of numerical routines The routines studied were grouped according to the mode of parallelism they employed Simple parallelism concurrent exe...
متن کاملMixed Large-Eddy Simulation Model for Turbulent Flows across Tube Bundles Using Parallel Coupled Multiblock NS Solver
In this study, turbulent flow around a tube bundle in non-orthogonal grid is simulated using the Large Eddy Simulation (LES) technique and parallelization of fully coupled Navier – Stokes (NS) equations. To model the small eddies, the Smagorinsky and a mixed model was used. This model represents the effect of dissipation and the grid-scale and subgrid-scale interactions. The fully coupled NS eq...
متن کاملMixed Large-Eddy Simulation Model for Turbulent Flows across Tube Bundles Using Parallel Coupled Multiblock NS Solver
In this study, turbulent flow around a tube bundle in non-orthogonal grid is simulated using the Large Eddy Simulation (LES) technique and parallelization of fully coupled Navier – Stokes (NS) equations. To model the small eddies, the Smagorinsky and a mixed model was used. This model represents the effect of dissipation and the grid-scale and subgrid-scale interactions. The fully coupled NS eq...
متن کاملParallelizing Molecular Dynamics Codes using Parti Software Primitives
This paper is concerned with the implementation of the molecular dynamics code, CHARMM, on massively parallel distributed-memory computer architectures using a data-parallel approach. The implementation is carried out by creating a set of software tools, which provide an interface between the parallelization issues and the sequential code. Large practical MD problems is solved on the Intel iPSC...
متن کامل