I don't know the exact formula for how the total score is achieved but obviously the nature and dragothic low increases account for most of the score increase
Printable View
I don't know the exact formula for how the total score is achieved but obviously the nature and dragothic low increases account for most of the score increase
Um man you REALLY should redo your bench with the 40.41's only with your ram at DDR 400, because 3Dmark has a massive increase from performance when you feed it mem bandwidth.Quote:
Originally posted by JustWhitman
http://home.insightbb.com/~forseti/3dmark.jpg
Thats a compare between my computer with the 40.41 drivers and the 29.42 drviers. The only diffrence is that when I had the 29.42 drivers my comp ran stable with DDR400, and with the 40.41's I had to take my RAM to DDR300 (not because of the drviers, but because I had lowered my Vmem) As you can see the only real bench that got a massive gain was the Nature benchmark, but then agian, the only real bench that the 9700 pulls ahead of the GeForce4 Ti is in the nature benchmark, at default settings. In some cases I actually lost performance, so how does a 20FPS gain in Nature merit a 600pt gain overall in 3DMark? Something is definately fishy here.
You cant possibly get that score , unless your CPu was overclocked ..Quote:
Originally posted by The_Bloodguard
I increased my GF4 4200 from 260/520 to 290/550 and installed the new 4041 drivers and my 3dmark2001 went from about 11300 to 13264. Very cool and I kick Pinchy's butt by 50 pts or so! :)
Was it ???
Quote:
Originally posted by JustWhitman
In some cases I actually lost performance, so how does a 20FPS gain in Nature merit a 600pt gain overall in 3DMark? Something is definately fishy here.
JustWhitman's 3DMark is correct.Quote:
Originally posted by sapasion
I don't know the exact formula for how the total score is achieved but obviously the nature and dragothic low increases account for most of the score increase
To calculate your 3DMark you do this:
((Add up low details scores) + ((Add up high details scores)*2))*10 = Your 3DMark score
In this situation it will be:
((180.3+227.2+165.2) + ((59.8+123.7+73.8+75.1)*2))*10 = 12375
I probably will eventually, I've just been spending most of my freetime playing Mafia, hehe. I think I will tomoroow though, just to see what kind of gains I get out of the mem increase.Quote:
Originally posted by KaoTiK
Um man you REALLY should redo your bench with the 40.41's only with your ram at DDR 400, because 3Dmark has a massive increase from performance when you feed it mem bandwidth.
--edit--
BTW what kind of gain do you think I would get out of DDR400 vs DDR300 speeds.
--another edit--
I beat him by 5FPS in Nature bench though, heh.
Yea I'd like to know too, I mean his card is clocked lower than mine, yet he beats me over 1,000pts.Quote:
Originally posted by cocacola155
You cant possibly get that score , unless your CPu was overclocked ..
Was it ???
--edit--
I looked at his compare, and yea he has his 2.26 clocked @ 2.652, plus the RDRAM helps him blow me away I guess :D I think Next month I'm going to pick up a 2.53 along with some RDRAM, and a new mobo of course.
Yes, my CPU was OC to 156 FSB but my CPU speed is noted in the 3DMark compare link at the bottom of my sig.Quote:
by CocaCola You cant possibly get that score , unless your CPu was overclocked ..
What surprised me was that my last 3DMark was ~11300 at the same 156 FSB but the vid card was set at 260/520 at that time. I ran 3DMark at 260/520, no O/C on the 2.26 and the 4041 drivers and received a ~11900 score so I bumped the vid card and CPU and got the 13200 score.
With the way some people believe 3dmark i sometimes think that extra 3dmarks is all they need to enjoy their games better even if the image quality is slaughtered ;)Quote:
Originally posted by Un4given
We don't know that they have been holding out. At this time we don't know what the drivers changed. They could have just as easily changed the LOD or something else to gain these extra points. Also, 3DMark is a common benchmark. Most card makers try to optimize drivers for games/programs used most commonly in benchmarking. Unfortunately, when this is done it rarely translates into increased performance across the board for all programs.
Which i dont think it was.
Must be that RIMM4200, I guess it loves the bandwith. I mean, I have my 2.26 @ 2.55 (150FSB), and then card OCed to 300/648 and I get 13212. Pretty unbelieveable, but wow. Sweet jesus. Anyways, good job. Stupid Ti4400. Hehe.Quote:
Originally posted by The_Bloodguard
Yes, my CPU was OC to 156 FSB but my CPU speed is noted in the 3DMark compare link at the bottom of my sig.
What surprised me was that my last 3DMark was ~11300 at the same 156 FSB but the vid card was set at 260/520 at that time. I ran 3DMark at 260/520, no O/C on the 2.26 and the 4041 drivers and received a ~11900 score so I bumped the vid card and CPU and got the 13200 score.