Re: Optimising the Keen-Minsky Model

[richard]

Hi c, 

Please find attached the revised KeenMinsky02d.m 

It runs some 25% slower, but I find it easier to read. 

I tidied up a few things and some things I tried did not 
work as expected. For example, I tried replacing "Minsky" 
with E(19) and Octave complained ... I will see if I can 
find a better way of doing that, but it may be the logical 
way to make changes to the varible under test. 

"serial version with outer loop" greatly helped debugging, 
as the error messages provided by the parallel version were 
somewhat cryptic (and I did a few really dumb things). 

I may have found a bug - 
the original "serial version with outer loop" code had this 
#  Y(:, j) = X(1:end-1, 3);
this worked in the original version, but octave complained 
in the new code 
error: KeenMinsky02d: A(I,J,...) = X: dimensions mismatch

Changed to >  Y(:, j) = X(1:end-0, 3);

regards

Richard

KeenMinsky02d.m
Description: Text document