> single precision double precision
> ********************* ********************
> real complex real complex
>Benchmark % Clock % Clock % Clock % Clock
>========= ========= ========= ========= =========
> kSelMM 590.6 546.4 313.1 295.6
> kGenMM 97.9 97.3 97.5 97.5
> kMM_NT 96.6 95.3 96.9 95.2
> kMM_TN 97.0 95.3 96.7 95.1
> BIG_MM 572.5 561.4 302.6 298.9
> kMV_N 53.7 141.0 42.2 56.2
> kMV_T 62.4 83.9 43.3 56.5
> kGER 93.1 115.1 28.8 38.9
>
>IS it because I configured with
>../configure -b 64 -Fa alg -fPIC -D c -DPentiumCPS=3100
>--prefix=/home/ossetest
Nope, nothing wrong with the configure line that I see. My guess is that
configure doesn't recognize the chip due to changing model/CPU family
values. What does your full cpuinfo look like?
>From what wikipedia shows me, this is a Core2-based processor, so you
should be able to compare against Core2Duo64SSE3.tgz.
If you want to try to install using the Core2 architectural defaults, pass :
-A 15
to configure.
For Core2-based systems, you will get substantially better double precision
performance with 3.9.1 than you do with 3.8.2 (from the above, it looks
like you are using 3.8.2).
Cheers,
Clint
**************************************************************************
** R. Clint Whaley, PhD ** Assist Prof, UTSA ** www.cs.utsa.edu/~whaley **
**************************************************************************
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________
Math-atlas-devel mailing list
Math-atlas-devel@...
https://lists.sourceforge.net/lists/listinfo/math-atlas-devel