>Now, since my version for some reason is faster than George's (2-3% on LL
>tests, 6-7% on factoring; I have no clue why), I don't really like going
>back to a slower version.
I have played around with the Prime95 V18.1 & found a simular improvement
in compiling on a more recent compiler (MS VC++ 5.0 & 6.0). The orignal
code indicates it is a MS VC++ 4.x program -- which had a terrible
optimizer. Interestingly enough after recompiling the Recache program
doesn't seem to help at all so maybe there is a difference in the way it is
utilizing ram. I also noticed a couple of things in the GUI that was
slowing it down by another % or so. All in all I was able to squeeze about
4%-5% extra speed out of the LL. However I never allowed any of those
results to be reported to Primenet because I have not found a way to do
enough QA testing on it to completely satisfy myself. It would be very
nice if we could put together a QA procedure that we could test these
possible performance enhancers on our own then if everything seemed to be
ok to send the changes to George & QA team for possible inclusion in a new
version.
Shane Sanford
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
