Sign in with
Sign up | Sign in
Your question

7900GT issues at [H]

Last response: in Graphics & Displays
Share
a b U Graphics card
May 25, 2006 10:54:24 PM

[H] talks about the issues of many factory overclocked GF7's. Maybe I'm just out of the loop, but I though these things were very OC friendly. Strange the factory OC'ed cards are having such issues when enthusiasts are pushing these cards so high on their own. Is a good OC'in 7900GT common, or is it just the cherry picked ones we hear about?

http://enthusiast.hardocp.com/article.html?art=MTA2OSwx...

More about : 7900gt issues

May 25, 2006 11:37:13 PM

That is strange; one would think the cards would be good overclockers because of the 90nm process. It's a factory overclock, so at least it's got a warranty and you can RMA it.
Related resources
May 26, 2006 5:35:25 AM

Quote:


So if it's not nV's fault whose is it?


ATi's of course :roll:
May 26, 2006 3:08:14 PM

Quote:
ATi's of course :roll:


The fault lies with that 7800 GS of yours! :p 
a b U Graphics card
May 26, 2006 6:36:51 PM

Kinda like HDCP, well it's not our fault we just make the VPU/GPUs (or even cards). :twisted:

Anywhoo, I don't know what to make of it, but I'd say, be sure to have sufficient cooling that's for sure (BTW, WTF! seriously they run cooler than the ATi VPUs!), it's gotta be something on the boards themseleves, like that other xtremesys thread mentions.
May 26, 2006 6:57:11 PM

That's a good guess; I've seen no issues with ATI cards like this, even though they generally run a lot hotter.
May 26, 2006 7:04:25 PM

If it's happening on the GT's, I'd guess the manufacturers are just releasing them with too little voltage.
a b U Graphics card
May 26, 2006 7:55:08 PM

Quote:
If it's happening on the GT's, I'd guess the manufacturers are just releasing them with too little voltage.


Yeah, actually that'd make sense too. Too low voltage would cause greater chance of instability, but keep temps down compared to higher voltage required to overcome the increase noise due to the relative increase in thermals.

It'd be interesting to see what the final verdict is when it's all sorted out.
May 26, 2006 8:04:29 PM

my understanding was that b/c the 7900 runs some things @ asynch speeds the "standard" way the manufacturers were oc'ing them freaked out this "new" architecture... that is what I got from the [H] article anyway.

of course, the fact that Nv never told anyone about possible prbs w/ oc'ing the new chip until now still makes it someone elses fault right? ;) 
a b U Graphics card
May 26, 2006 8:09:36 PM

Quote:
my understanding was that b/c the 7900 runs some things @ asynch speeds the "standard" way the manufacturers were oc'ing them freaked out this "new" architecture... that is what I got from the [H] article anyway.


Well the GF7800 and 7600 also run asynchronously for pixel and Vertex clocks.
May 26, 2006 8:20:43 PM

ya, i realize that but I thought that was still Nv's reasoning for the issues in that [H] article... which is why they said that all is ok when running @ stock clocks. perhaps I read it wrong?
May 26, 2006 8:45:00 PM

Quote:
of course, the fact that Nv never told anyone about possible prbs w/ oc'ing the new chip until now still makes it someone elses fault right? ;) 


To their credit though, they release a reference that everyone is supposed to follow... 7900 GT, X amount of Mhz, and X amount of voltage. 7900 GTX, different voltage, different Mhz.

If the manufacturers are increasing Mhz but not Voltage, I don't blame Nvidia...
May 26, 2006 9:17:25 PM

agreed, but if that was the case wouldn't the makers be able to release a "quick" driver patch that bumps the voltage? This has been an issue (at least on [H]) for a while now and yet nothing has really fixed it in total.

not saying that voltage is not the issue b/c it seems a logical cause, but perhaps there is more here than what is on the surface? things that make you go hmm...
May 26, 2006 9:48:19 PM

Voltage would be controlled by hardware on an Nvidia card... you'd need to solder something to the card to properly increase the voltage.

or a pencil mod at least, to short something out.
May 26, 2006 10:09:44 PM

ya, duh... i didn't even think about that. lol

and actually when reading that forum that grapeape has above it does seem (if it is all the same issue) that it is a voltage/heat thing, which ultimately comes down to hardware. So ya, if the card functions at Nv's advertised specs then Nv is not at fault. If vendors oc the card past what Nv says is kosher then it would be the vendors fault.

of course most of us have come to demand a certain oc headroom on any hardware today, and perhaps that is not a fair demand... if ati or Nv release something that performs as the advertise but cannot oc, is it their fault at all? hmm...
a b U Graphics card
May 27, 2006 12:42:37 AM

It could be voltage like you said or something else on the card overheating like Grape said. Just odd when it's GT's and GTX's, and even supposedly reference clocked models. I understand user error and system stability always enters into these forum complaints, but it seems the [H] forums shouldn't have such a high number of those cases...more likely the majority are flawed cards with such similar issues.

Honestly I have become an EVGA fan for NV and would look toward a CO KO SC model like my 7800GT CO. With similar pricing, a high factory clocked one has looked to be the way to go....until you read horror stories. 8O

Oh well, at current prices I'm a X1800XT fan all the way right now, especially if the 7900GT's #1 appeal (OC'in) is giving them issues. Shoot, My Sapphire is $219 AR now!!!!! http://www.newegg.com/Product/Product.asp?Item=N82E1681... I can't see buying anything else in the $200-300 range.
!