[CSC 435] Oops - FLOPS

[Andrew J. Pounds]

Guys --- there is a part of the Dusty Deck assignment that asks you to...

"Also show the performance results for your completely tuned program and 
graph the flop ratio as a function of problem size."

So counting the flops is different than just estimating  O(n) -- aka 
"big oh" -- and dividing by the time. The tool I usually use for this - 
called PaPi, was working on hammer (which we can't use) and has some 
problems with double precision flop counting on the systems in GSC 218.  
For this assignment just don't worry about counting the flops.

We will go over this in some detail on a later date (including using 
PaPi and other tools) because it is a skill you must have for HPC work 
and benchmarking.


-- 
*/Andrew J. Pounds, Ph.D./*
/Professor of Chemistry and Computer Science/
/Director of the Computational Science Program/
/Mercer University, Macon, GA 31207 (478) 301-5627/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://theochem.mercer.edu/pipermail/csc435/attachments/20260129/9837b45f/attachment.htm>