Experiences and results multitasking a hydrodynamics code on global and local memory machines
A one-dimensional, time-dependent Lagrangian hydrodynamics code using a Godunov solution method has been multitasked for the Cray X-MP/48, the Intel iPSC hypercube, the Alliant FX series and the IBM RP3 computers. Actual multitasking results have been obtained for the Cray, Intel and Alliant computers and simulated results were obtained for the Cray and RP3 machines. The differences in the methods required to multitask on each of the machines is discussed. Results are presented for a sample problem involving a shock wave moving down a channel. Comparisons are made between theoretical speedups, predicted by Amdahl's law, and the actual speedups obtained. The problems of debugging on the different machines are also described.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 6190736
- Report Number(s):
- LA-UR-86-4356; CONF-870831-1; ON: DE87003759
- Resource Relation:
- Conference: International conference on parallel processing, St. Charles, IL, USA, 17 Aug 1987
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallel processing a real code: A case history
Fast algorithms for structural optimization and least squares. Annual interim report, 15 July 1986-14 July 1987
Related Subjects
SUPERCOMPUTERS
COMPARATIVE EVALUATIONS
PARALLEL PROCESSING
FINITE DIFFERENCE METHOD
FORTRAN
HYDRODYNAMIC MODEL
ONE-DIMENSIONAL CALCULATIONS
COMPUTERS
DIGITAL COMPUTERS
ITERATIVE METHODS
MATHEMATICAL MODELS
NUMERICAL SOLUTION
PARTICLE MODELS
PROGRAMMING
PROGRAMMING LANGUAGES
STATISTICAL MODELS
THERMODYNAMIC MODEL
990210* - Supercomputers- (1987-1989)