[Beowulf] benchmaster

Robert G. Brown rgb at phy.duke.edu
Thu Dec 16 18:06:34 EST 2004


OK, I spent a day and a half cleaning up cpu_rate, renaming it (as it no
longer just tests CPU per se anyway), making it a neat little website,
and packaging it all up to go.  I even made the repository
yum-distributable, so in FC3 you can drop a benchmaster.repo into your
yum repo directory and automagically update from it.

Of course the primary purpose for a benchmark harness isn't to be
installed via a package, it is to play with it.  So you might want to
get the tarball instead of the binary rpm so you can e.g. change
compilers and flags and so forth.

This is fully GPL code with no restrictions on it other than the
purely nominal "beverage" clause.  It is designed for you to be able to
add your own code fragments and time them fairly easily.

Its other main virtue is that it makes it very easy to generate a list
of benchmark results for different values of the parameters (such as
vector length in stream).  This lets you make plots, and plots carry a
lot more information than a short table of numbers.  For example, look
at the jpegs on the lucifer.html page below.

Main page:

http://www.phy.duke.edu/~rgb/General/benchmaster.php

Sample/example page:

http://www.phy.duke.edu/~rgb/General/benchmaster/lucifer.html

I've worked on this enough that I'm HOPING that it is largely decrufted
and reasonably bug free and even tolerably documented.

But no guarantees.  It IS free software, after all...;-)

    rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu


_______________________________________________
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