Running Linpack

Felix Rauch rauch at inf.ethz.ch
Thu Jun 28 10:55:49 EDT 2001


On Thu, 28 Jun 2001, Ferdinand Mahr wrote:
> Has anyone experience with the High Performance Linpack benchmark from
> netlib.org? After some trouble with the makefiles, I was able to compile
> and run it on a dual machine (unfortunately I don't have two seperate
> similar machines or even a cluster at work right now), but when it finished
> after an hour or so, I didn't get any results.
> My question is: how can I determine the MFLOP performance of my system
> (compareable to the results in the top500.org list)? I couldn't find an
> answer in the provided documentation. Is there a script that analyzes the
> output of the program? Are there any other linpack benchmark
> implementations?

On the standard output (or your output file) you should see something
like this:

T/V                N    NB     P     Q               Time             Gflops
----------------------------------------------------------------------------
W01R2L8        57000    94     8    16            2212.04          5.582e+01
----------------------------------------------------------------------------
||Ax-b||_oo / ( eps * ||A||_1  * N        ) =        0.0013793 ...... PASSED
||Ax-b||_oo / ( eps * ||A||_1  * ||x||_1  ) =        0.0017576 ...... PASSED
||Ax-b||_oo / ( eps * ||A||_oo * ||x||_oo ) =        0.0003179 ...... PASSED

My understanding is that the number below Gflops is the achieved
performance.

If you specified multiple runs, you will see multiple outputs of the
above. Choose the highest number with all operations "PASSED" (nothing
"FAILED").

- Felix
-- 
Felix Rauch                      | Email: rauch at inf.ethz.ch
Institute for Computer Systems   | Homepage: http://www.cs.inf.ethz.ch/~rauch/
ETH Zentrum / RZ H18             | Phone: ++41 1 632 7489
CH - 8092 Zuerich / Switzerland  | Fax:   ++41 1 632 1307


_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list