Statistics – Computation
Scientific paper
Jan 2012
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2012aas...21914517r&link_type=abstract
American Astronomical Society, AAS Meeting #219, #145.17
Statistics
Computation
Scientific paper
Current numerical simulations employed by the astrophysics simulation community typically require long times to complete, particularly for three-dimensional regions. The recently developed CUDA extensions for C allow existing code to be modified to run on modern graphics cards in an efficient and parallelizeable manner. Graphics cards, particularly those developed for scientific computation, have significant speed advantages over CPUs for highly-parallelizeable applications such as magnetohydrodynamics grid solvers. Athena, a grid-based MHD code for astrophysics (Stone et al. 2008), is retooled to run on CUDA capability 2.0 or greater NVIDIA graphics cards, with effort taken to maximize performance. Performance benchmarks between graphics cards and CPUs are presented.
No associations
LandOfFree
A Modern Astrophysical MHD Solver on CUDA-Capable GPUs does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with A Modern Astrophysical MHD Solver on CUDA-Capable GPUs, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and A Modern Astrophysical MHD Solver on CUDA-Capable GPUs will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1576237