[CSC 435] Cough Cough

[Andrew J. Pounds]

The dusty deck has some real nastiness in it that I will go over on 
Thursday.  If you have already downloaded it and started to play with it 
you will note that the makefile starts off at the -O3 optimization 
level.  This breaks the code.  You have to start with the -O0 level, 
check the results, increase to -O1, and make sure the results are the 
same.  This process continues to higher optimization levels.   Feel free 
to drop the MAXDIM value down to something like 25 to get the code to 
run faster.

We will explicitly start timing the code in an exercise on Thursday.  
You will note that my makefile creates a timing library.   While you may 
be inclined to use language-specific timers please resist that urge.  My 
timers have been shown to give consistent results across both languages 
and the will not impact the performance.

Also -- do not try to use the random number generators from the 
different languages.  You will get inconsistent results making it 
impossible to build code for testing.  I spent a lot of time researching 
random number generators and what you have in this code is perferct for 
our purposes.

On Thursday I will discuss the best practices for re-writing this code 
to get consistent results while improving performance --- and we will 
let dusty run on the lab computers and warm things up a bit... and build 
our first (of many) profile graphs.

We will then move on to more serious code profiling tools next week.

-- 
*/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/20260106/6dc44dd0/attachment.htm>