Ok, so iv had a chance to play with a pre overclocked Zotac 8800gt (not mine), courtesy of a friend who's pc I have just built, as I managed to pick up one of these rare as rocking horse ***t (in the uk at least) cards for his new build.
Anyways, on with the testing. As a comparison, I benched MY 8800gts 640mb before installing the gt, to get some initial numbers and make my mate feel good about his purchase
Test setup was as follows for testing both cards, overclock on cpu/ram lowered for more 'real world' results:
q6600 @ 3ghz 1.1v (to represent average joe 100% achievable everytime overclock) on p5b deluxe
2gb Crucial Ballistix pc8500 @ 334fsb (1:1 @ 4-4-4-12, 1.8v)
2x 80gb seagate 7200.9 RAID0
Windows Xp sp2 and Nvidia beta driver 169.04
169.04 driver was used as I did these tests before the 169.09 came out, and I no longer have access to the card - it was a temporary mate to mate loan.
Driver cleaner was run between card installs to make sure each driver install was clean and fresh for each card. Nvidia drivers set to force 'High Quality' and LOD bias set to 'Clamp'.
Now for the cards:
In the 'old' corner we have my Palit 640mb 8800gts @ stock clocks - 513mhz/792mhz (shader 1188mhz)
In the 'new' corner, the Zotac 512mb 8800gt @ factory OC - 660mhz/900mhz (shader 1620mhz)
Tests!!
First up Crysis........Now before anyone shouts "169.04 beta has driver cheats for crysis" - I know. Thus crysis.exe was renamed to avoid the hack, and graphics etc were therefore acurate and perfect.
Game cfg files were modded to 'enable' vista's 'Very High' settings under xp dx9, res @ 1280x1024, no AA or AF.
Crysis gpu bench mark allowed to run all 4 loops, then I took an average of the 4 average fps stats.
640mb 8800gts - Average fps: 18.96
512mb 8800gt - Average fps: 26.42
World In Conflict - Set to 'very high', 'water reflects clouds' enabled, AA and AF both OFF @ 1600x1200.
640mb 8800gts - Min 15fps, Max 73fps, Average 36fps
512mb 8800gt - Min 26fps, Max 96fps, Average 47fps
Aquamark - default settings.
640mb 8800gts - 165,133
512mb 8800gt - 165,333
3dMark 2005 - default settings.
640mb 8800gts - 16,574
512mb 8800gt - 18,795
3dMark 2006 - default settings.
640mb 8800gts - 10,721
512mb 8800gt - 13,493
Soooo, a conclusion is in order. My mates got a bargain Seriously though, the 8800gt is a very nice bit of kit, it definately showed up my 8800gts, although neither card can manage crysis at my native tft res of 1600x1200....thats a next gen 9800 etc task. But for here and now its a very capable bit of kit.
-----------------------------------------------------------------------
As an extra test I re-ran all tests on my 8800gts, but with clock speeds boosted to exactly match those of the 8800gt.
So my 640mb 8800gts was overclocked to: 660mhz core, 900mhz memory, 1620mhz shader. 100% 24/7 stable btw.
So if i copy and paste my 8800gt scores but update the 8800gts comparative numbers we get:
Crysis
640mb 8800gts Overclocked - Average fps: 24.72
512mb 8800gt - Factory OC - Average fps: 26.42
World In Conflict -
640mb 8800gts Overclocked - Min 23fps, Max 92fps, Average 43fps
512mb 8800gt - Factory OC - - Min 26fps, Max 96fps, Average 47fps
Aquamark - default settings.
640mb 8800gts Overclocked- 170,024
512mb 8800gt - Factory OC- 165,333
3dMark 2005 - default settings.
640mb 8800gts Overclocked- 18,214
512mb 8800gt - Factory OC - 18,795
3dMark 2006 - default settings.
640mb 8800gts Overclocked- 12,904
512mb 8800gt - Factory OC - 13,493
A MUCH closer result. So at same clock speeds its simply the extra pipelines giving a little extra boost to the 8800gt....
essentially it wins out simply due to higher clocks.
Ok, so we technically new this already, but I just wanted to show people that anyone with a decent overclocking 640mb gts has no real reason to swap out to a gt....its simply not much of an upgrade IF your gts can hold the same everyday clocks of a gt - if it can theyre basically the same.
640mb 8800gts dual slot stock cooler manages much better than the gt single slot POS even at identical clock speeds for what its worth too
and whilst I have no comparative numbers, the 640mb gts did much better at 1600x1200 with AA and AF turned on than the gt when at ident. clocks, simply due to larger vram.
Sooooo, any thoughts or comments? Sorry if this post seems pointless to anyone.
Anyways, on with the testing. As a comparison, I benched MY 8800gts 640mb before installing the gt, to get some initial numbers and make my mate feel good about his purchase
Test setup was as follows for testing both cards, overclock on cpu/ram lowered for more 'real world' results:
q6600 @ 3ghz 1.1v (to represent average joe 100% achievable everytime overclock) on p5b deluxe
2gb Crucial Ballistix pc8500 @ 334fsb (1:1 @ 4-4-4-12, 1.8v)
2x 80gb seagate 7200.9 RAID0
Windows Xp sp2 and Nvidia beta driver 169.04
169.04 driver was used as I did these tests before the 169.09 came out, and I no longer have access to the card - it was a temporary mate to mate loan.
Driver cleaner was run between card installs to make sure each driver install was clean and fresh for each card. Nvidia drivers set to force 'High Quality' and LOD bias set to 'Clamp'.
Now for the cards:
In the 'old' corner we have my Palit 640mb 8800gts @ stock clocks - 513mhz/792mhz (shader 1188mhz)
In the 'new' corner, the Zotac 512mb 8800gt @ factory OC - 660mhz/900mhz (shader 1620mhz)
Tests!!
First up Crysis........Now before anyone shouts "169.04 beta has driver cheats for crysis" - I know. Thus crysis.exe was renamed to avoid the hack, and graphics etc were therefore acurate and perfect.
Game cfg files were modded to 'enable' vista's 'Very High' settings under xp dx9, res @ 1280x1024, no AA or AF.
Crysis gpu bench mark allowed to run all 4 loops, then I took an average of the 4 average fps stats.
640mb 8800gts - Average fps: 18.96
512mb 8800gt - Average fps: 26.42
World In Conflict - Set to 'very high', 'water reflects clouds' enabled, AA and AF both OFF @ 1600x1200.
640mb 8800gts - Min 15fps, Max 73fps, Average 36fps
512mb 8800gt - Min 26fps, Max 96fps, Average 47fps
Aquamark - default settings.
640mb 8800gts - 165,133
512mb 8800gt - 165,333
3dMark 2005 - default settings.
640mb 8800gts - 16,574
512mb 8800gt - 18,795
3dMark 2006 - default settings.
640mb 8800gts - 10,721
512mb 8800gt - 13,493
Soooo, a conclusion is in order. My mates got a bargain Seriously though, the 8800gt is a very nice bit of kit, it definately showed up my 8800gts, although neither card can manage crysis at my native tft res of 1600x1200....thats a next gen 9800 etc task. But for here and now its a very capable bit of kit.
-----------------------------------------------------------------------
As an extra test I re-ran all tests on my 8800gts, but with clock speeds boosted to exactly match those of the 8800gt.
So my 640mb 8800gts was overclocked to: 660mhz core, 900mhz memory, 1620mhz shader. 100% 24/7 stable btw.
So if i copy and paste my 8800gt scores but update the 8800gts comparative numbers we get:
Crysis
640mb 8800gts Overclocked - Average fps: 24.72
512mb 8800gt - Factory OC - Average fps: 26.42
World In Conflict -
640mb 8800gts Overclocked - Min 23fps, Max 92fps, Average 43fps
512mb 8800gt - Factory OC - - Min 26fps, Max 96fps, Average 47fps
Aquamark - default settings.
640mb 8800gts Overclocked- 170,024
512mb 8800gt - Factory OC- 165,333
3dMark 2005 - default settings.
640mb 8800gts Overclocked- 18,214
512mb 8800gt - Factory OC - 18,795
3dMark 2006 - default settings.
640mb 8800gts Overclocked- 12,904
512mb 8800gt - Factory OC - 13,493
A MUCH closer result. So at same clock speeds its simply the extra pipelines giving a little extra boost to the 8800gt....
essentially it wins out simply due to higher clocks.
Ok, so we technically new this already, but I just wanted to show people that anyone with a decent overclocking 640mb gts has no real reason to swap out to a gt....its simply not much of an upgrade IF your gts can hold the same everyday clocks of a gt - if it can theyre basically the same.
640mb 8800gts dual slot stock cooler manages much better than the gt single slot POS even at identical clock speeds for what its worth too
and whilst I have no comparative numbers, the 640mb gts did much better at 1600x1200 with AA and AF turned on than the gt when at ident. clocks, simply due to larger vram.
Sooooo, any thoughts or comments? Sorry if this post seems pointless to anyone.