Distributed N-body Simulation on the Grid Using Dedicated Hardware

Statistics – Computation

Scientific paper

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

Scientific paper

We present performance measurements of direct gravitational N-body simulation on the grid, with and without specialized (GRAPE-6) hardware. Our inter-continental virtual-organization consists of three sites, one in Tokyo, one in Philadelphia and one in Amsterdam. We run simulations with up to 131072 particles for a variety of topologies. As expected, high performance simulations over the entire planet are dominated by network latency. Based on these timings we construct a model to predict the performance of a hypothetical grid consisting of all the 582 GRAPE-6 modules on Earth. Such setup reaches break-even between network communication and computation for N˜10^8 particles, resulting in a compute time of about one year for 1~N-body time unit. Much better performance can be obtained using only a local cluster of GRAPEs. Equipping the Netherlands DAS-3 grid with GRAPEs would achieve break-even at a million particles, resulting in a compute time of about half an hour for 1~N-body time unit.

No associations

LandOfFree

Say what you really think

Search LandOfFree.com for scientists and scientific papers. Rate them and share your experience with other people.

Rating

Distributed N-body Simulation on the Grid Using Dedicated Hardware 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 Distributed N-body Simulation on the Grid Using Dedicated Hardware, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Distributed N-body Simulation on the Grid Using Dedicated Hardware will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFWR-SCP-O-1812662

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.