>Don't think I didn't try this (and a number of similar ideas). The
>problem was that the forked copy of Prime95 doesn't seem to start
>executing until ReCache completes. Either that or ReCache
>suspends itself until Prime95 completes - *most* unsatisfactory to
>hog all that memory long-term...
That's interesting; I haven't been getting that problem. When pass 2b
starts, Prime95 launches, and ReCache continues until the pass finishes, and
then it terminates, leaving Prime95 running at peak performance. I'm running
Windows 98, and I compiled your code with Borland C++ Builder 4. Could one
of these factors make a difference?
(If you're interested in trying out my compiled executable, I'd be glad to
send it to you.)
>As you've found out, running ReCache to completion then starting
>Prime95 does have a good effect (it forces DLLs etc. which aren't
>currently active to be dumped, for a start). However, I certainly get
>more consistent results using the manual start procedure.
The automatic method has been completely consistent for me. Strange...
_______________________________________________________________
Get Free Email and Do More On The Web. Visit http://www.msn.com
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
