I think my 7900GT is screwed up

[TheRyuu]

Maby, just maby, my 7900GT is completly screwed. Maby I got one of those bad batches for the 4th time in a fvck'n row!

CS:S Dust2 Screen Shot

Something tells my it's not suppose to look like that.

Suggestions? I kinda feel bad for Evga since this would by my 4th (yes thats right, i'm on my third 7900GT and this will be my FOURTH) RMA. For some reason, the card in my first slot is ALWAYS fine. I never find anything wrong with it. But the second slot, (which works just fine) seems to get the screwed up cards.

Guess the next step would be to try out each card one at a time?

Edit (Update): everything is working fine now, no worrys.

 

[buzzsaw13]

Does it do that for other games?

 

[TheRyuu]

Originally posted by: buzzsaw13
Does it do that for other games?

CS:S, Oblivion, yes, it does it for other games too.

It's gotten worse too. In the beginning I could play for a few hours and only see a few artifacts here and there, but now it's much worse. I'll try Oblivion again just to make sure.

 

[videopho]

It's not "Maby". It's defintely a bad card, assuming stock speed is currently used. "Maby" it's time to start a 5th RMA process. Sorry to see such bad luck with a single product line. You may set a new world record in terms of number of replacements of the same video card in such a short time span. Well eVGA just recently announces the newer P/N for the 7900 series so let's hope eVGA has finally turned the tide on this issue.

 

[TheRyuu]

Originally posted by: videopho
It's not "Maby". It's defintely a bad card, assuming stock speed is currently used. "Maby" it's time to start a 5th RMA process. Sorry to see such bad luck with a single product line. You may set a new world record in terms of number of replacements of the same video card in such a short time span. Well eVGA just recently announces the newer P/N for the 7900 series so let's hope eVGA has finally turned the tide on this issue.

It's actually my 4th RMA (well maby 3rd, lost count, lets just say too many). It was fine when I got it a month ago, but now it just showed the symptoms of a bad card.

Not going to worry about it now though. Time to just stop with messing with the computer for a while

 

[KutterMax]

Do you know how hot your card is running? Does the game initially play fine and then after a few minutes do the artifacts appear? That would suggest that heat could be an issue.

My previous videocard, a 6800GT starting doing something similar while playing World of Warcraft some months back. Although my temperatures weren't too bad, I replaced my heatsink with a new one - this helped for a while but a few months later that problems recurred and I replaced the 6800GT with a 7800GS OC and the problem went away.

 

[Muhadib]

I agree 100% with KutterMax. It certainly looks like a heat issue from the screenshot. Especially if it works fine for a while then the problem presents itself. Not to mention 4 new cards ate it in the same part of your case.

I would take the side off your case and point a "house" box fan in your case just to rule out a heat issue. Unfortunately "the damage" may have already been done and, you'll most likely RMA your 5th card to be baked in your oven yet again.

If the box fan idea doesn?t help try underclocking your memory to 650 and see if the problems go away. If so, it?s likely the damage has already been done.

P.S. If you want to find out what card is the problem you'll have to test them one at a time.

 

[Pens1566]

It looks like the common 7900 death. Not a heat issue with the case or card. Same thing happened with 3 of mine. Case was ~35C, card 48/65. Its the memory on the old batch. New EVGA cards have replaced the samsung memory with infineon.

 

[Muhadib]

Originally posted by: Pens1566
It looks like the common 7900 death. Not a heat issue with the case or card. Same thing happened with 3 of mine. Case was ~35C, card 48/65. Its the memory on the old batch. New EVGA cards have replaced the samsung memory with infineon.

I'm going to take a wild stab and say 4 times in a row is a bit too "common" to rule out the card's environment. But who knows, I suppose its possible, just not likely. Even being very generous: By saying 1 in 10 of the cards is faulty and he's had 4 cards. The chances would be 1 in 10,000 that this would happen.

It is more likely that people push the card over the limit and/or don't take proper heat precautions than a 1 in 10,000 chance.

My 2 cents...

 

[Pens1566]

No OC on the card, in a P180 that runs a 7800GTX KO(more power and warmer) perfectly well. Also ran my 7800GT CO perfectly well before I stepped up to the 7900GT. I would agree with your logic if there weren't many people with similar circumstances. Card temps hit high 60s under load. I don't have temp issues and I'm not running a low end PSU or crappy RAM. Also, Prime stable 12+ hours. Swap 7900GT out for 7800GTX, everything works. What conclusion would you draw???

 

[TheRyuu]

I'm on water cooling. The max my GPU's ever get is 37C under load on the core, and I have copper BGA ram sinks on the RAM even though you really don't even need them. I also have a FAN right next to both of the cards, just so that they get a little air flow.

I doubt that it's a heat problem. It's probably just another 7900GT death just like all the others. It most likely (as everyone says) has to do something with the VRAM and the VRAM voltage regulators. Hopefully the new 58x 7900GT's don't have this problem.

And for the record, this is my 3rd RMA, not my 4th (counted wrong). You know what they say, third time's the charm.

 

[Muhadib]

Originally posted by: Pens1566What conclusion would you draw???

I would say you F?d your card somehow or it was one of the ?faulty? ones. But if you told me you had 4 cards go bad 4 times in a row, I would surely say it?s likely some other external factor coming into play or EVGA hates you and is sending you the bottom of the barrel cards out of spite. ;-p

Considering in most cases people report the cards seem to run just dandy for (~a month) and then progressively gets worse tells me that there is indeed a hardware problem and it?s something that happens over a relatively long period of time for most.

Some people report no problems and they overclock/vmod memory and core for 4+months. This tells me that only some cards are faulty and/or some external factor(s) is(are) causing the problem to arise.

EVGA released the return rates on the EVGA 7900gt?s and it wasn?t an astronomical figure. This leads me to believe that it?s an issue of perception; where only the people who have problems are talking therefore they are the majority of people being heard and the problem magnitude is skewed.

Not to mention the fact that trusting what people say is hard enough but when you throw the fact that it?s completely anonymous, (via the internet) it makes it even more difficult. I?m guessing a good percentage of the repeat RMA people are lying through there respective teeth and are knowingly nuking card after card with the reassurance and ignorance that they can just RMA for another one as if it were a diaper.

The bottom line is that there is not a definitive answer yet but one can draw conclusions never the less.

 

[Mucker]

Hey Muhadib,

I see your sig:

EVGA 7900 GT Core:678 MHz Mem:1986 MHz AS5 & VF700-Cu @ MAX temp:62c 3DMARK05:10819

and have a question for you. If your card was to quit tomorrow, honestly, would you RMA or take responsibility for overclocking the snot out of it?

 

[TheRyuu]

Well, now that it works fine, I did the memory vmod on it and honestly thought that that's what screwed my card up. But I kept saying to myself "Everyone else does it, it'll be fine". So I did it, and my card got screwed up.

Reversed the vmod and now everything is fine. WTF? How come only I seem to get problems with 7900GT's :/

Oh well. And just to put it out there, the last 2 RMA's of the bad cards were no because of a failed vmod attempt. I didn't even have time to do the vmod on the other two bad cards. They came pretty much DOA.

Thanks for the Help (if I really got any).

 

[mikek753]

looks like you know how to "kill" 7900GT

not to be rude, but looks like some steps to "kill" 7900GT:

1. get a card and it works normal - but, "slow"
2. OC to ~600/1600 - better, but not enough
3. vmod gpu - better, but memory is bottleneck
4. vmod mem - the best possible 7900gt speed

now, how many ppl went all 4 steps and got stable 7900GT for more then a month?

7900GT has or had some issue by itself, so it fails even at step 1.
step 2 may speed up issues ppl have seen with step 1.
step 3. may even more speed up 7900 GT issue and without proper cooling gpu may went bad
step 4. may kill memory right a way depends on memory chips and how wold you do it

so, if you going all those steps then chances are getting working "super OC" card shrinks
and yes - there are no "safe" OC limit - you get extra 100 3dmarks, would you stop?
another 100 then another and another until .... you know

that are all IMHO - not a rule or any kind

 

[Muhadib]

Originally posted by: Mucker
Hey Muhadib,

I see your sig:

EVGA 7900 GT Core:678 MHz Mem:1986 MHz AS5 & VF700-Cu @ MAX temp:62c 3DMARK05:10819

and have a question for you. If your card was to quit tomorrow, honestly, would you RMA or take responsibility for overclocking the snot out of it?

Personally If I were the cause of my card breaking because I did not take the proper precautions or took to great of a risk I would be damn sure it would not happen again. I feel that people taking advantage of a company is harmful to the good intentions of an ?open? RMA procedure and to properly diagnosing the real problem. In the end that sort of behavior is counterproductive for all of us.

Don?t get me wrong, I?m not saying everybody is doing it, but the numbers just don?t add up when I see people reporting RMA?s in the double digits. Keep in mind though there is no definitive answer yet so my opinions are just that.

 

[Mucker]

Originally posted by: Muhadib

Originally posted by: Mucker
Hey Muhadib,

I see your sig:

EVGA 7900 GT Core:678 MHz Mem:1986 MHz AS5 & VF700-Cu @ MAX temp:62c 3DMARK05:10819

and have a question for you. If your card was to quit tomorrow, honestly, would you RMA or take responsibility for overclocking the snot out of it?

Personally If I were the cause of my card breaking because I did not take the proper precautions or took to great of a risk I would be damn sure it would not happen again. I feel that people taking advantage of a company is harmful to the good intentions of an ?open? RMA procedure and to properly diagnosing the real problem. In the end that sort of behavior is counterproductive for all of us.

Don?t get me wrong, I?m not saying everybody is doing it, but the numbers just don?t add up when I see people reporting RMA?s in the double digits. Keep in mind though there is no definitive answer yet so my opinions are just that.

:thumbsup:

 

[nullpointerus]

Originally posted by: wizboy11
Well, now that it works fine, I did the memory vmod on it and honestly thought that that's what screwed my card up. But I kept saying to myself "Everyone else does it, it'll be fine". So I did it, and my card got screwed up.

Reversed the vmod and now everything is fine. WTF? How come only I seem to get problems with 7900GT's :/

Oh well. And just to put it out there, the last 2 RMA's of the bad cards were no because of a failed vmod attempt. I didn't even have time to do the vmod on the other two bad cards. They came pretty much DOA.

Thanks for the Help (if I really got any).

Actually, I was following your (?) 7900GT "safe vmod" and was wondering how it turned out.

Guess I won't be trying it anytime soon.