70 likes | 141 Views
Comparing Parallel Computaion Times USING 2D SOR Dan Stalloch 4/19/2011. Languages used. MPI, OpenMP, Cilk++, Charm++, Cahpel, X10. Images Generated by running the MPI VErsion. 9 size, 3 rows, 3 cols, 5000 threshold. 14 size, 2 rows, 2 cols, 0.001 threshold. The Codes Used.
E N D
Comparing Parallel Computaion Times USING 2D SORDan Stalloch4/19/2011
Languages used MPI, OpenMP, Cilk++, Charm++, Cahpel, X10
Images Generated by running the MPI VErsion 9 size, 3 rows, 3 cols, 5000 threshold 14 size, 2 rows, 2 cols, 0.001 threshold
The Codes Used The MPI code The OpenMP code The Cilk code The Charm++ 2D SOR Module The Charm++ 2D SOR Header File The Charm++ 2D SOR Source File An example of using the Jacobi computation
Timing Numbers used 4 processors, 14 size, 2 rows, 2 cols, 0.001 threshold For MPI Time taken was: 1303255876 seconds and 638772 microseconds For OpenMP Time taken was: 1303256179 seconds and 182009 microseconds (this is currently without executing whole code)
How to Make Huge unimpressive Files 10,000 size, 5 rows, 2 cols, 0.001 threshold, size of file generated is 95.4 MB
References http://charm.cs.uiuc.edu/tutorial/BroadcastHelloWorld.htm http://blog.speedgocomputing.com/2010_08_01_archive.html http://ems.calumet.purdue.edu/mcss/kraftrl/cs590a/Successive-Over-Relaxation/2D-SOR-MPI.c http://arstechnica.com/civis/viewtopic.php?f=20&t=723453 http://www.prace-project.eu/documents/15_x10_wl.pdf http://cass-mt.pnnl.gov/docs/pubs/craytechreportrealms-deliv7.pdf http://openmp.org/wp/openmp-compilers/ http://cplusplus.syntaxerrors.info/index.php?title=%27for%27_loop_initial_declaration_used_outside_C99_mode http://x10.codehaus.org/