• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

AMD driver papers over VRM issue?

E

eloj

2[H]4U
Joined
Dec 31, 2000
Messages
3,612
So, I was researching this Powercolor 4870 1GB PCS when I stumbled on this article which details how the card is completely unable to survive Furmark, even on stock frequencies. A Powercolor rep. called it a 'software compatibility issue'. Yeah, right.

Let's see what happened a bit later:

"If you run furmark on catalyst 8.8, try rename executable of furmark to f.exe and run it. You see much more temp on VRM an core also VDDC up from 56 to 74 A. Fps rise in 2 time. VRM temp rise to 115 C or more." (comment 12)

"Yeah, FurMark is nerfed in the 8.8 Catalyst driver on purpose. So beware. If you want to really push your card you either need an older Catalyst driver (e.g., 8.7) or maybe the renaming executable trick to be done." (comment 14), http://en.expreview.com/2008/08/23/powercolor-hd-4870-pcs-stability-issues-check/4/

Seems to me AMDs partners have been complaining about having to make the necessary investments to make their VRMs survive Furmark (after all, it would probably cost a dollar or two), and here's the saddening result. Fix hardware? Nope. Paper over problem with driver hacks? Why, yes!

Sickening.
 
I've had one 4850 fail on me because of the VRM. This was with a Accelero S1V2 and no cooling on the VRM (people where still confused at the time to either cool the VRM or not).

Replaced that 4850 and installed the Accelero again (temps without a fan are 51-52c idle and 70c load) but with Zalman memory cooler on the VRM.

The card started to show artifacts (like the first card did) and noticed when I took it out that one of the memory heatsinks fell off from the VRM. Now I have another dead card.

I'm thinking of just going with a Nvidia 280 and calling it day. :mad:
 
I have a 4850 @ 850/1125 on water with the end of the stock heatsink cut to still cover up the VRM's and so far so good. Also have a 120mm fan giving it some air just in case. I get through future mark with no problems or artifacts. I think putting ramsinks is risky with my past experience with them. I have swiftech's copper ones, but can't seem to be able to put full trust in them, which is why I chopped up the end of the heatsink for the VRM's, plus it still has a firm hold with the screws holding in on tight. Max Temperature I ever see is 40Celcius on the core even at 1.4V load.
 
My 4870 never passed furmark. There is a good amount of ppl out there having the same problem. Oh well.. never fails me in a single game.
 
Pixelated Taco said:
I've had one 4850 fail on me because of the VRM. This was with a Accelero S1V2 and no cooling on the VRM (people where still confused at the time to either cool the VRM or not).

Replaced that 4850 and installed the Accelero again (temps without a fan are 51-52c idle and 70c load) but with Zalman memory cooler on the VRM.

The card started to show artifacts (like the first card did) and noticed when I took it out that one of the memory heatsinks fell off from the VRM. Now I have another dead card.

I'm thinking of just going with a Nvidia 280 and calling it day. :mad:
Click to expand...

Quit right now. You just RMA'ed 2 video cards because you don't know what you're doing? Please do switch to Nvidia so ATI doesn't have to pay for your mistakes. I want my second 4870x2 to be as cheap as possible.
 
Furmark...if only it ever really tested for gpu clocks stability instead of burning holes through the core and memory.
 
You must log in or register to reply here.
Back
Top