Many Thanks Galen, I've been running an implementation of parallel Fast Fourier Transform here for 1 week and it works fine.
I'll try those benchmarks today and tomorrow. Cheers, LEslie 2006/3/5, Galen M. Shipman <gship...@cs.unm.edu>: > Hi Leslie, > > To start I would try running the Intel test suite, you can find this > here: http://www-unix.mcs.anl.gov/mpi/mpi-test/tsuite.html there are > also several other test suites available on this site, this will give > you correctness first. > For benchmarks you may try the NAS Parallel Benchmarks found here: > http://www.nas.nasa.gov/Software/NPB/ > > Thanks, > > Galen > > On Mar 4, 2006, at 9:23 AM, Leslie Watter wrote: > > > Hi All, > > > > Some time ago, I was trying to implement a new BTL to openmpi. > > > > Now it works here, but I need to test it. > > > > Someone (sorry, I've forgot who), told me to use NetPIPE to test > > the BTL layer. > > Ok, I used it, but now I need some parallel applications that uses > > openmpi. > > NetPIPE uses a ping-pong test between 2 machines to evaluate its test. > > > > > > If anybody could send me an application that uses openmpi, > > I can return with the results after its execution. > > > > Why this? I need to compare the execution time just changing the > > BTL protocol. > > > > Here I have a cluster with 16 machines (Athlon 1.2 128RAM with > > 100Mbits ethernet and a 1Giga Switch), using openmpi-1.0 stable (with > > modifications that include LLC BTL [ I plan to release my code in a > > website as soon as I have the results and the analysis done. ]. > > > > What kind of application I'm looking for? > > - applications that uses TCP btl > > - applications that can run in 2,4, 8 and/or 16 machines > > - preferably applications that are message passing intensive > > (optional) > > > > > > Many Thanks > > > > Leslie > > PS: I want to put the results of execution time in mine master thesis. > > > > -- > > Leslie H. Watter > > > > _______________________________________________ > > devel mailing list > > de...@open-mpi.org > > http://www.open-mpi.org/mailman/listinfo.cgi/devel > > _______________________________________________ > devel mailing list > de...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/devel > -- Leslie H. Watter
