Issue No. 03 - May/June (2012 vol. 14)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MCSE.2011.117
Ali Khajeh-Saeed , University of Massachusetts, Amherst
J. Blair Perot , University of Massachusetts, Amherst
<p>In this scenario, computational fluid dynamics simulations of turbulence are performed with 64 GPUs and an optimized CFD algorithm using communication/computation overlapping. Detailed timings reveal that the GPUs' internal calculations are so efficient that operations related to data exchange between compute nodes now cause a scaling bottleneck on all but the largest problems.</p>
Computational fluid dynamics, CFD, multi-GPU, conjugate gradient, Laplacian, scientific computing
A. Khajeh-Saeed and J. B. Perot, "Computational Fluid Dynamics Simulations Using Many Graphics Processors," in Computing in Science & Engineering, vol. 14, no. , pp. 10-19, 2011.