RichBa5tard | 12th June 2003 18:16 | Quote: Originally posted by TeuS I also think so, but it's known that your scores go down after a few benchmarks, how are you going to solve that? | i need betatesters for measuring how big of difference it is. =)
If you're talking about the 3Dmark score drop when you bench a second (/third/...) time without closing 3Dmark: that problem is already solved! The MadMark tool kills the application after each run, and starts a new instance for the next run. There is a slight drop in performance due to MadMark hogging a little cpu power/memory in the background, but it's pretty limited (~0.20 sec slower on pifast). As long as the performanceloss is consistent, MadMark will represent comparable results.
MadMark is not ment to be used to reach a 25k 3Dmark score, it's ment to compare HW/software settings. :)
I'm cleaning up the code of MadMark ver 0.05, if someone wants to betatest tonight, please do! |