OpenGL Fur Benchmark

[Random Murderer]

no xfire makes no difference in this test at all

yes, yes it does. it is entirely dependent on the cards used.

 

[cdawall]

yes, yes it does. it is entirely dependent on the cards used.

welll i have no difference on my 3850s whatsoever

 

[Random Murderer]

welll i have no difference on my 3850s whatsoever

the 3xx0 series seems to have no benefit, but the x1k does.
no word on the hd2k series, because apparently nobody that has crossfired hd2k cards has the balls to run this test.

 

[cdawall]

the 3xx0 series seems to have no benefit, but the x1k does.
no word on the hd2k series, because apparently nobody that has crossfired hd2k cards has the balls to run this test.

ahh ok

 

[CrackerJack]

the 3xx0 series seems to have no benefit, but the x1k does.
no word on the hd2k series, because apparently nobody that has crossfired hd2k cards has the balls to run this test.

I've ran this both in single and crossfire. But once ATI release crossfire drivers for my card, then i'll post more scores. But 2615 on a single 2900gt is pretty darn good

 

[Random Murderer]

I've ran this both in single and crossfire. But once ATI release crossfire drivers for my card, then i'll post more scores. But 2615 on a single 2900gt is pretty darn good

lol, you don't really count, sorry.
in order for you to qualify, your cards need to be able to run crossfire, lol.

 

[CrackerJack]

lol, you don't really count, sorry.
in order for you to qualify, your cards need to be able to run crossfire, lol.

Well i can run crossfire, but my score is only like 2400

 

[P4-630]

I decided to do some overclocking with my Gigabyte hd3870..

The Samsung gddr3 was clocked at 1170MHz..
No pencil mod, just overclocked the core and mem..


I did not go any higher because there were a few arctifacts showing up.

Arctifacts free at core 910MHz and mem 1150MHz

 

[CrackerJack]

Arctifacts free at core 910MHz and mem 1150MHz

try to raise the voltage up a bit,

 

[dark2099]

Tried the X-treme G Drivers and got a slight increase.
dark2099
3338

 

[P4-630]

I could run 3dmark06 arctifact free.

Only furmark shows a few sometimes.

 

[P4-630]

try to raise the voltage up a bit,

I never tried to raise the gpu voltage with overclocking, which tool are you using to do that?
when overclocked that high it runs idle about 52 degrees C, but full load it stays under 70.

 

[CrackerJack]

I never tried to raise the gpu voltage with overclocking, which tool are you using to do that?
when overclocked that high it runs idle about 52 degrees C, but full load it stays under 70.

RBE, it works for my card. But I don't know if it will work for yours

 

[P4-630]

RBE, it works for my card. But I don't know if it will work for yours

Ok, well could you tell me how you did it?

(By the way my card has gddr3)

Thanks

 

[CrackerJack]

Ok, well could you tell me how you did it?

(By the way my card has gddr3)

Thanks

Just change were i got the red cycles. You will have more to change though.
I know this actually works, because i used Rivatuner to monitor the volts in 2d and 3d state. And Everlast, here's the link
http://forums.techpowerup.com/showpost.php?p=771455&postcount=215

 

[P4-630]

Thanks for the info CrackerJack,
I may give it a try...But..
First I'm doing some more reading/research about this, since overvolting a gpu is more risky then just bumping up de clock speeds..

 

[CrackerJack]

Thanks for the info CrackerJack,
I may give it a try...But..
First I'm doing some more reading/research about this, since overvolting a gpu is more risky then just bumping up de clock speeds..

yeah it's a risk, everyone will tell you that. But the higher clocks you get the less stable it will be, but change the volts will allow it more stable. Just don't take big steps. I went from 1.1 to 1.2. Doesn't seem like awhole lot. But I got an extra 50mhz on the core and 200mhz on memory. I'm sure I can get higher clock speeds, but then that's where cooling comes in. I only run at 850/1200 when i bench. I keep the fans at 93% speed. I don't get no higher than 75c. If you decide to change your voltage, take the clock speeds back to normal. Then work your way back up.

 

[P4-630]

yeah it's a risk, everyone will tell you that. But the higher clocks you get the less stable it will be, but change the volts will allow it more stable. Just don't take big steps. I went from 1.1 to 1.2. Doesn't seem like awhole lot. But I got an extra 50mhz on the core and 200mhz on memory. I'm sure I can get higher clock speeds, but then that's where cooling comes in. I only run at 850/1200 when i bench. I keep the fans at 93% speed. I don't get no higher than 75c. If you decide to change your voltage, take the clock speeds back to normal. Then work your way back up.

And you don't have a 3870 but 2x hd 2900gt..

 

[JrRacinFan]

For Random Murderer

Xfire enabled

Xfire disabled

Both cards were kept at these settings:

Also it looks as though I get better performance with 8.4 cats.

 

[CrackerJack]

And you don't have a 3870 but 2x hd 2900gt..

yeah i've just got 2x 2900gt

 

[Random Murderer]

For Random Murderer

Xfire enabled

Xfire disabled

Both cards were kept at these settings:

Also it looks as though I get better performance with 8.4 cats.

so the x1k cards are the most recent to see a benefit from crossfire in opengl.

 

[dark2099]

Higher clocks, also watched the temps with everest for both cards and load with gpuz for the 2nd card and the 2nd card didnt see any usage.
dark2099
3394

 

[P4-630]

I just gave it another try.


Default core 780MHz overclocked to 918MHz
Gddr3 default speed 950MHz overclocked to 1197MHz and the chips just stay cool
(no voltage or pencil mods used)

Result: 3790 (os win XP)

Core 918
mem 1188

Who said that hd 3870's with gddr3 are crippled cards?

 

[Wile E]

Wile E
8800GT 1GB
3654
http://www.ozone3d.net/benchmarks/score.php?id=bef360706f6cae07f0acc12e1015c5ad

 

[Fahim]

XFX GeForce 7950 GX2 1GB (520,1300) 175.75

XFX GeForce 7950 GX2 1GB (550,1400) 175.75

Will post my other cards later...