• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Artifacting AMD 6950 (Updated)

thilanliyan

Lifer
Jun 21, 2005
11,332
1,178
126
Damn it seems I'm not having much luck with 6950s lol.

My 2nd 6950 seems to be artifacting at anything above 1.2v whereas it wasn't doing that before. This card was NOT flashed at all, just OCed using Afterburner. The thing is it will artifact even if all I do is up the voltage without touching the clocks. Temps seem to be fine...core doesn't go above ~50C (watercooled) and what may be the VRM temps don't go much higher than that either.

Any suggestions?

EDIT: Artifacts seem to go away after letting it run for a bit:
http://forums.anandtech.com/showpost.php?p=31164318&postcount=19


What the heck could be the cause of that? VRMs running too cool? Lol. :)
 
Last edited:

Elfear

Diamond Member
May 30, 2004
7,035
557
126
What applications is it artifacting in? Games, ATI Tool, OCCT? You probably know this but ATI Tool has been shown to artifact at stock clocks with the 6900 series cards. It needs to be updated to work correctly. OCCT seems to work well though.

What are you using to measure the VRM temps? What kind of airflow do you have on your cards?
 

thilanliyan

Lifer
Jun 21, 2005
11,332
1,178
126
What applications is it artifacting in? Games, ATI Tool, OCCT? You probably know this but ATI Tool has been shown to artifact at stock clocks with the 6900 series cards. It needs to be updated to work correctly. OCCT seems to work well though.

What are you using to measure the VRM temps? What kind of airflow do you have on your cards?
I've tried Unigine and Metro 2033 and they both artifact. Even if I run stock clocks but up the voltage to 1.2v it will start artifacting. I'm going by GPUz for VRM temp, although that may not infact be reporting VRM temps. The GPU is watercooled and everything else is sinked with a fan blowing right onto the components.

@happy, I'm using a Corsair HX620.

EDIT: I just tried UNDERCLOCKING both core and mem but upping volts and same thing...artifacts.
 
Last edited:

happy medium

Lifer
Jun 8, 2003
14,387
475
126
I find it hard to believe that you got 2 high quality XFX cards going bad in a matter of weeks.

Do you overclock your cpu? IF not mabe you should try, just to see if mabe it is your power supply. Your cpu overclock should not be stable with you psu if your psu is going bad.
If it is overclocked try to run it stock speeds.

Something else is killing your cards I think?
 

thilanliyan

Lifer
Jun 21, 2005
11,332
1,178
126
^The first card I killed on my own as I stated in my other thread.

This card has not been molested at all (stock BIOS), just OCed using AB.

My CPU is OCed to 3.8GHz and has been that way for quite a while.
 

thilanliyan

Lifer
Jun 21, 2005
11,332
1,178
126
^Yeah exactly, I've read of people taking there's to 1.3v with no problems. And I have very good cooling on it actually.
 

badb0y

Diamond Member
Feb 22, 2010
4,012
27
91
Oh man that's terrible. I am at 1.3 voltage right now everything going good. I think I can push it to 1.4 too lol but I am not feeling it right now.

Hmm, how do you check the VRM temps? GPU-z doesn't show them :( .
 

happy medium

Lifer
Jun 8, 2003
14,387
475
126
My CPU is OCed to 3.8GHz and has been that way for quite a while
The first thing I would do is lower the cpu overclock and recheck video card stability.
a overclocked 6 core cpu (150+ watts) and overvolted overclocked 6950 gpu (200 watts) could potentially wear down a psu. That is about 350 watts on the 12v rail not including you case fans , motherboard and hard drives. That might bring it to 400 watts.

If you downclock and it still acting up, looks like its rma time ,again.

Edit your psu has a 600 watt 12v rail, that should be ok.
 
Last edited:

badb0y

Diamond Member
Feb 22, 2010
4,012
27
91
The first thing I would do is lower the cpu overclock and recheck video card stability.
a overclocked 6 core cpu (150+ watts) and overvolted overclocked 6950 gpu (200 watts) could potentially wear down a psu. That is about 350 watts on the 12v rail not including you case fans , motherboard and hard drives. That might bring it to 400 watts.

If you downclock and it still acting up, looks like its rma time ,again.
He never RMA'd the other one, he broke it so he had to buy another one.
 

OVerLoRDI

Diamond Member
Jan 22, 2006
5,491
3
81
I agree with happy medium, try lowering or removing your CPU overclock and then see what happens. Is your ram overclocked as well?

What brand is your PSU?
 

Elfear

Diamond Member
May 30, 2004
7,035
557
126
How are you checking the VRM temps?
From post #4

I'm going by GPUz for VRM temp, although that may not infact be reporting VRM temps.
That is a head-scratcher OP. PSU seems perfectly capable. My bet would still be cooling issues with the VRM. Those things get pretty hot even with heatsinks and a fan.

What are your Power Control settings at?
 

dulemk

Junior Member
Jan 29, 2011
14
0
0
I have the same problem with my sapphire 6950.

I found that my voltage is actually the vid3 in RBE but for some reason sapphire bios can't increase the voltage at all.
So now I use assus bios.


One strange thing i notice is that if i use 1.15v on vid3 i have small amount of artifact for the first seconds in 3d, and after that it is all ok.

Than i'v tried 1.175, and i notice that i had more artifact and lasts more seconds(about 10-15), and than everything is clear again.

But if i try 1.2v i have lots of artifact and i did not waited long enough to clear, but my prediction is that they will to clear eventually.

So is this normal?? Does voltage need to lie down somewhat, so i wait for the artifact to run out, or my core voltage just wont work(1.2v is not a problem for all others cases that i know).

I have coolermaster real power 700W psu, i7 920 at 4GHz.


I use 11.1 hotfix.
 

Castiel

Golden Member
Dec 31, 2010
1,772
0
0
Damn it seems I'm not having much luck with 6950s lol.

My 2nd 6950 seems to be artifacting at anything above 1.2v whereas it wasn't doing that before. This card was NOT flashed at all, just OCed using Afterburner. The thing is it will artifact even if all I do is up the voltage without touching the clocks. Temps seem to be fine...core doesn't go above ~50C (watercooled) and what may be the VRM temps don't go much higher than that either.

Any suggestions?
Whats the highest stable core you've gotten?
 

thilanliyan

Lifer
Jun 21, 2005
11,332
1,178
126
I have the same problem with my sapphire 6950.

One strange thing i notice is that if i use 1.15v on vid3 i have small amount of artifact for the first seconds in 3d, and after that it is all ok.

Than i'v tried 1.175, and i notice that i had more artifact and lasts more seconds(about 10-15), and than everything is clear again.

But if i try 1.2v i have lots of artifact and i did not waited long enough to clear, but my prediction is that they will to clear eventually.

So is this normal?? Does voltage need to lie down somewhat, so i wait for the artifact to run out, or my core voltage just wont work(1.2v is not a problem for all others cases that i know).

I have coolermaster real power 700W psu, i7 920 at 4GHz.


I use 11.1 hotfix.
Wow okay you are right...I left it for a bit and the artifacts went away...1.2v works, 1.25v works. Very strange.

Whats the highest stable core you've gotten?
I've tried as high as 975 at 1.25v...not tried any higher.
 
Last edited:

littlezipp

Golden Member
Nov 7, 2001
1,860
0
76
Wow okay you are right...I left it for a bit and the artifacts went away...1.2v works, 1.25v works. Very strange.


I've tried as high as 975 at 1.25v...not tried any higher.
Same issue here with my Crossfire XFX 6950's. Can play Crysis for 2+ hours and temps not exceed the mid 40's, all components cool to the touch. Fuzzy dots on game boot but none during play. Random dots sometimes during web browsing too.
How do we fix?
 

dulemk

Junior Member
Jan 29, 2011
14
0
0
Wow okay you are right...I left it for a bit and the artifacts went away...1.2v works, 1.25v works. Very strange.
Yes i know but i wont risk it, i don't want my card to die eventually.
So is there someone that know what is causing this technically??
No need to be just for 6950.
 

dulemk

Junior Member
Jan 29, 2011
14
0
0
Sorry for the double post, this is some updated info.

I now flash back to my original 1408 shader 6950 bios that i take from gpuz and keep it from day one.

And now apparently MSIA voltage works, so i try to raise just the voltage(clocks stays the same 800/1250).

And the same f*ing thing happen.

So now I try to find the lowest visible artifact showing with lower voltage.
Up till now i find that i have some artifact showing at 1.12v:mad:

So now I'm kind of hoping at 1.1v the artifact to appear so i could rma the card

One more thing i will try before i give app and use only unlocked bios at 840/1300 with 1.1v is to change the driver at 10.12a
I read somewhere that 11.1 or 11.1 hotfix render some errors on some cards, so i will try that and post here.


EDIT!!:

I'v tried 10.12a, and now i know that i did not have luck with this card, i just cant increase the voltage at all.

So now i'm using my bios just unlocked at 840/1325 and i'm quite happy with this setup, so for the further notice i officially give up.
 
Last edited:

betasub

Platinum Member
Mar 22, 2006
2,677
0
0
^ I don't undestand why you had to flash to test the original stock BIOS - why not just flip the switch & re-boot? Instant stock BIOS settings.
 

ASK THE COMMUNITY