Parallelizing KIVA-3 to get the Right Simulation Time in Engine Design

S. Filippone, G. Bella, and P. D’Ambra (Italy)


Computational fluid dynamics, parallel computing, numerical software, sparse linear algebra.


Modern fluid dynamics simulations are very resource intensive, both in terms of memory space and of computational time; thus parallel computing platforms are often necessary to satisfy complex design goals within reasonable time. We present an account of our experience with the parallelization of the KIVA-3 fluid dynamics code, outlining the major problems encountered and the direction of future developments.

Important Links:

Go Back