PARALLEL COMPUTING EFFICIENCY FOR REAL WORLD PROBLEMS

PARALLEL COMPUTING EFFICIENCY FOR REAL WORLD PROBLEMS

Carl, Dietmar

Full Article:

The numerical simulation of a high speed short duration event like a detonation or a projectile penetration usually requires the use of sophisticated explicit computer codes. These codes apply small time steps for every computational cycle. Combined with the need to model large structures with millions or hundreds of millions of nodes, these computations can take from days to weeks to run. To shorten the computation time, modern commercial explicit codes support the use of parallelized computer systems. This approach frequently claims to reduce computation times: The use of eight processors instead of one is supposed to reduce computation time by a factor of six or seven for example. While that might be true for simple problems, it is not necessarily true for complex problems. This paper describes some real world applications and corresponding numerical simulations involving Euler and Lagrange computation techniques. Single and parallel computing times are compared.

The numerical simulation of a high speed short duration event like a detonation or a projectile penetration usually requires the use of sophisticated explicit computer codes. These codes apply small time steps for every computational cycle. Combined with the need to model large structures with millions or hundreds of millions of nodes, these computations can take from days to weeks to run. To shorten the computation time, modern commercial explicit codes support the use of parallelized computer systems. This approach frequently claims to reduce computation times: The use of eight processors instead of one is supposed to reduce computation time by a factor of six or seven for example. While that might be true for simple problems, it is not necessarily true for complex problems. This paper describes some real world applications and corresponding numerical simulations involving Euler and Lagrange computation techniques. Single and parallel computing times are compared.

Palavras-chave:

DOI: 10.5151/meceng-wccm2012-18502

Referências bibliográficas

• [1] M.L. Wilkins. Computer Simulation of dynamic Phenomena. Berlin: Springer, 1999.

Como citar:

Carl, Dietmar; "PARALLEL COMPUTING EFFICIENCY FOR REAL WORLD PROBLEMS", p-1654-1661. In: In Proceedings of the 10th World Congress on Computational Mechanics [= Blucher Mechanical Engineering Proceedings, v. 1, n. 1]. São Paulo: Blucher, 2014.
ISSN 23580828, DOI 10.5151/meceng-wccm2012-18502