I think that when you read benchmark tests, Gar, they are done on stand-alone computers and the computer is rebooted before each test. This give the 'cleanest' memory setup for repeatable results.

Barbara

>Thanks Barbara, I was afraid of that. I still cannot help wondering though. I see all sorts of testing software out there. Is there some program out there that will prevent all the clever disk and memory caching that NT does so that a program will not run faster the second time.
>
>For that matter is there some way of configuring NT to disable all that (usually) great stuff. I know any such configuration would be slower than snot (a technical term) but if it would let me compare run times without rebooting, it would be well worth it.
>
>I think I'll cross post this question to the windows forum.