[Wien] Benchmarks

Peter Blaha pblaha at theochem.tuwien.ac.at
Fri Jul 8 17:31:30 CEST 2005


No, this pgf90 compiler is very bad (and version 6.0 is quite buggy)
At the beginning we did not haveanything else than this pgi compiler.
What is very good is the gotolib for our Xeon machines. It even beats the
newest Intel mkl by quite some amount:

test_case.output1_ifort9_mkl:         TIME HAMILT (CPU)  =    21.8, HNS =    16.3, DIAG =   118.5
test_case.output1_ifort9_gotopthreads:TIME HAMILT (CPU)  =    20.9, HNS =    12.6, DIAG =    88.1
test_case.output1_pgi_gotopthreads:   TIME HAMILT (CPU)  =    43.8, HNS =    12.2, DIAG =    86.4
test_case.output1_pgi_goto_dual:      TIME HAMILT (CPU)  =    45.1, HNS =    16.9, DIAG =   146.0

So clearly ifort is more than twice as fast in hamilt (vector-sin/cos operations),
and gotolib is much better than mkl.

Regards

> I noticed that the benchmark pages now contains some rather fast timings
> with pgf90 and 64-bit goto libraries. Has the Portland Group finally got
> their act togethor to compete with ifort?
> 
> Note: if you have an old email address for me, please note that "nwu" has
> been changed to "northwestern".
> 
> -----------------------------------------------
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2220 N Campus Drive
> Northwestern University
> Evanston, IL 60201, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> email: L - marks @ northwestern . edu
> http://www.numis.northwestern.edu
> -----------------------------------------------
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 


                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------


More information about the Wien mailing list