Originally posted by: wineo
, excellent scores with the GT
. Do you think the gtx would work ok with the 4c0re? with like 10% loss?
: Also did you do a run of aquamark3?
When I got my system stable with my 2gb ram, I also tried upping the fsb. I tried up to 320, win would boot but restart during testing (no errors). I then did more testing got it down to 303 passed for 17h testing but then I saw a restart. This is not enough vcore right?
Can't think of any reason against using a 8800gtx with a 4coredual-sata2 mobo! However, no way of really knowing how big the performance hit will scale when using a "gtx" or even "ultra" untill someone trys either, or both?
Personally, for this mobo, I would choose to go for the 8800gt, or 8800gts, and overclock them hell out of it ---> makes more economical sense !
And Unlikely to achieve stable 320 fsb without setting Vagp of around 2V --> then need to heatsink Vagp mosfet and add fan to n/b heatsink.
Also, I think a restart after 17h of testing could have happened for a "million" different reasons, including even a windows "glitch". Try more Vcore if you must, but personally, for now, I would leave your restart after 17hours of stress testing as being down to a random unknown fault/failure ---> windows/hardware/software ---> who knows ?
The following are my scores using C2D E6700 @ 2.66GHz, Geil PC2-6400 @ 266MHz/3-3-3-9-1T, Palit 8800gt Sonic, with a 4coredual-sata2. Completely removed 7800gs agp drivers and did clean re-install of 169.06, which is the driver supplied with 8800gt. And just for interest, 3dMark01se and Aquamark are more systemwide orientated, so we would expect a 8800gt to be stressed more and have a larger performance impact during 3dMark03, 3dMark05 and 3dMark06 testing.
The results below should however only be taken as typical, and not best achievable, since my testing was carried out as a simple one time only "continious run of tests" in the order shown,
3dMark05 = 14789
3dMark06 = 9265
3dMark03 = 33091
3dMark01se = 43156
Aquamark03 = 138550
Using newer 174 series drivers and clocking E6700 @ 3GHz-plus and/or overclocking the hell out of the 8800gt would undoubtably have achieved MUCH higher scores, but that would defeat the point of trying to show what is achievable from the kind of midish performance setup that most ordinary people, not into serious overclocking, are more likely to have or aim for ?
With everything exact same except now using my agp 7800gs @ 545/1600, the following was achieved,
3dMark05 = 8547
3dMark06 = 4612
3dMark03 = 19780
3dMark01se = 38410
Aquamark03 = 118120.
As expected, being that 3dMark01se and Aquamark03 are more systemwide orientated, the performance margin increase when using a 8800gt was greater for the 3dMark03, 3dMark05 and 3dMark06 tests.
Mind you, maybe a more important point to note from my testing is that if using a 8800gt in the 4x PCI-E slot of a 4coredual mobo is likely to have a 10-15% hit on performance, then just maybe a 4coredual setup would be a much better platform for using a ATI 3850/3870 AGP card since presumably there would be no, or much less, performance loss ----> in other words, regardless of the fact that a 8800gt will perform better than ATI's 3850 and 3870 in 16x PCI-E setups, a 3850/3870 could become king in a 4coredual setup when used in the AGP slot ---> just a thought??