8800GT will be here Thursday!!!

[_dangtx_]

try the 6700 at 2900 mhz ull see it matches or close teh 2160 at 3 gigs

[highoctane]

try the 6700 at 2900 mhz ull see it matches or close teh 2160 at 3 gigs

I'll give it a run later, it almost seems to me like the crazy card is cpu bound at the default resolution for 06. From what I had read a e2140 at 3ghz was kinda the equivalent of a e6700 at stock 2.66, should be interesting to see what the cpu score turns out to be as well as the sm2 & sm3 scores.

Hmmm, its about lunchtime, maybe I'll head home and give it a shot.

[highoctane]

Ok here is the result of e6700 now at 8x333 (2.66ghz), overall score is about the same as the e2140 at 3ghz. I'm guessing this video card is very cpu bound at lower 1280 and lower resolutions. I guess things could be different with AA/AF cranked up and higher resolutions.

[highoctane]

Ok a little more testing, now with the cpu at 400x9 (3.6ghz), this is actually the first time I've seen real gains from OC'ing the video card.

e6700 3600mhz, 8800gt sc default clocks 650/950


e6700 3600mhz, 8800gt sc default clocks 675/987


e6700 3600mhz, 8800gt sc default clocks 702/987

[_dangtx_]

nice

[Brandito]

mem doesn't seem to clock very high? my gts does 1000 pretty easy, but that's up from the stock 800.

i guess the mem is probably about he sme on both cards though?

[Poci]

I dont even need to oc my 8800

Ill have fun doing that in the near future though

[COMPn00b09]

mem doesn't seem to clock very high? my gts does 1000 pretty easy, but that's up from the stock 800.

i guess the mem is probably about he sme on both cards though?

Your's is stock 800? Mine is 900, and I don't know if it's power issues or temps or what, but I can't get it much further than 940.

[Brandito]

Your's is stock 800? Mine is 900, and I don't know if it's power issues or temps or what, but I can't get it much further than 940.

seems to be the same quality mem, maybe yields are better so they decided the higher mem clock would be alright, 1000 seems about tops though.

[highoctane]

Well I wasn't topped out earlier but I just ran out of time, 2hr lunch was pushing it.

Anyways here's where I'm currently at with the card and still climbing:

738/1782/1026, I stopped clocking the shader because it moves in big steps, the next one being up to 1836mhz.

[Brandito]

you should be able to change the shader clock independantly with 1mhz increments in riva tuner?

what are you using to verify stability. i found ati tool is the best, heats card up much more thn most games. run the fuzzy cube thing to heat the card up then i ran the test for artifacts thing for at least an hour and if i had no artifacts i considered it good to go. if you can run a card on full tilt like that at temps it'll rarely reach in the real world for an hour i think you'll be good to go with 0 artifacts.

[highoctane]

Just tinkering with running 3dmark06 at the moment, just trying to find the absolute max's right now and then trying working backwards from there I guess for stability. I would say on water cooling where I'm at right now could be a very realistic clock but I don't know if it will game at these speeds or I should say I don't think I would try long duration gaming on stock air cooling at the current speed.

I have the clocks unlinked but the shader clock wants to make a big jump from where it is currently. Even though I can adjust in 1mhz increments the shader clock stays the same until you reach a point and it makes the next big jump, in my case I either run 1782 or 1836, no inbetween increments.

[Poci]

you should be able to change the shader clock independantly with 1mhz increments in riva tuner?

what are you using to verify stability. i found ati tool is the best, heats card up much more thn most games. run the fuzzy cube thing to heat the card up then i ran the test for artifacts thing for at least an hour and if i had no artifacts i considered it good to go. if you can run a card on full tilt like that at temps it'll rarely reach in the real world for an hour i think you'll be good to go with 0 artifacts.

Ive seen BF2 artifact where atitool wouldnt

[Brandito]

Ive seen BF2 artifact where atitool wouldnt

i haven't found that, i don't even see the artifacts ati-tool finds, so i'd say it'd be pretty difficult in game.

i have found in the past when i used to use 3dmark03 to find artifacts that the memory artifacts looked much different than ones from the core. the core usually had little speckles while the memory would look like the image was tearing.

i play a lot of 2142 and that game is just a bit harder on the system i believe (based on lower framerates at the same settings used in bf2 overall) and haven't seen any artifacting in it.

however it could easily be possible that one system artifacts easier with ati tool while another with something else.

i like using ati tool because it finds the artifacts i would never even see, so for ultimate stability i like that.

also it REALLY depends on how you use ati tool. if you're using the "find max" settings you'll get overinflated oc's a lot of the time, the artifact finder doesn't run long enough to really test for artifacting. many times it will oc the memory or core far beyond their stability without finding artifacts at all. best way is to allow the fuzzy cube to be animated till the card reaches its max temp then run the artifact tester for a good hour or more, well that's my method at least. longer or shorter periods may be fine, i'm sure everyone has their own methodology.

[NvMatrix]

3Dmark06
11697marks
5236 SM2.0
5596 SM3.0
2646 CPU