GeForce 7900 Inferno – Burn Baby Burn @ [H]

foofighter06....probably because when you registered they can tell by your s/n you have a card that will likely fail. I hope when I get mine I can find out if it's the new or old revision.. Glad I went with eVGA because they seem to be doing an outstanding job of taking responsibility for something that may not be their fault. :)
 
joemama said:
foofighter06....probably because when you registered they can tell by your s/n you have a card that will likely fail.

Well...both cards have very similiar s/n...

Idk i guess I'll send them both back.
 
joemama said:
I should be getting my eVGA 7900GT KO superclock in a couple days...what should I look for to see if I have one of the older problem cards or the new revision? I would rather just do an exchange with eVGA before I even install the thing if I can..

I haven't seen a sure way of telling yet. It isn't a case of a 100% 'bad batch,' so you shouldn't count your card out until you've tried it. Watch for lines to start appearing sticking out of characters or objects in games and 3dmark; that is an early symptom a lot of people have seen before their cards hit rock bottom. If you see that, underclock by several hundred mhz and RMA.

maomatt said:
Am I being needlesly pessimistic, or should I limit how hard I push my 7900GT's? Or even be proactive and RMA them back to XFX.

I wouldn't push a 7900gt past its factory settings, knowing about all this. It's not that the problem is brought on by overclocking (that's disproven by reference card owners running stock and still having trouble); it just seems like tempting fate to push a 7900.

However you choose to run them, if they are working without displaying the artifacting/shearing/tearing/screen blanking that seems to characterize the 7900gt Curse, then DON'T RMA them. If they start doing that, then return them, but if by "proactive" you mean you're worried they might go bad later.. don't. You could be trading a perfectly good pair of cards for a new set that's effectively DOA.


foofighter06 said:
Now that's all fine and everything, but here's the deal. I didnt start the RMA process.
(...)
Why would they automatically enroll me in an RMA process. When I didnt sign up for one?
And why did they only send the email to one account?

I see it was Jacob from EVGA who emailed you. Email him back and ask him these questions. He answered me when I replied to his RMA email and he has an excellent reputation on the EVGA forums for being active and responsive.
 
Filled out RMA Monday. Received email from Jacob Tuesday. Got my new 7900GT frome EVGA Thursday. Pretty damn good service if you ask me. This one goes to 565/882 when I do Detect Optimal Settings in CP.

My bad one did 542/832 Detect Optimal Settings and started artifacting after 2 weeks then I turned it down to 530/800 and that lasted about a week. Went to default clocks of 520/770 and they went downhill a couple of days later. We'll see how the new card holds up.

I've owned a 6800 Ultra/7800GT/7900GT all EVGA and I'll keep buying from them. Great warranty and RMA service.
 
I have just bought the Sparkle Calibre 7900GT 512MB card (del next week) that uses the reference PCB so I am hoping there are no problems though I'll report back if any

My MSI 7900GT was flawless and OC'd better than most and worked perfectly even with the 1.4V mod with the two hot spots sinked :)
 
I just found out from someone at eVGA that many of the 'fixed' cards that were sent out as replacements are actually 'Signature Series' cards that have had the BIOS re-flashed to the speeds of the model it's replacing before they ship it out as a replacement.

If you've revieced a 'fixed' replacement card for your defective eVGA card and it has a barcode sticker on the back that says 'Made in china' and has half of the number marked out with permanent marker, then odds are you have actually recieved one of these rebadged signature series boards.



So yesterday I decided to start some testing to see how this card runs at SS speeds (600/1600). I ran some 3DMark 03,05,06 for several hours as well as some F.E.A.R. and had great results at these speeds. And then before bed last night I left the PC running a loop of 3DMark 06's Shader Model 3.0 tests. (The water monster one and deep freeze one) and this morning it was still runnign them perfectly with no artifacting or glitching whatsover. Wow... :)
 
Got my BFG Tech 7900 GTX OC 512 on the 31'st
Loaded up Trainz2006 the next day and ran into the blinking screen of death
Tried doing 3dmark2006 oh boy what a mistake.
The card had only been out of the box less than 12 hours and it was already garbage.
Just got finished talking with the tech at BFG.
He apparently he has no clue that these problems have been going on.
First he suggested it was my power supply
then he suggested it was the video drivers didnt support the card
/rollseyes

almost $700 for a pile of junk
will not be buying another card from BFG
 
You are lucky!

My XFX 7900GTX (STOCK CLOCK) only lasted 5 (FIVE) minutes!!!

If it had lasted 12 hours at least I would have been able to play a game or two.
 
So if some of these replacements are just SS flashed with a different bios is it possible to flash these with a SS bios and get the same speeds etc or does it not matter.
 
JonnyRial said:
So if some of these replacements are just SS flashed with a different bios is it possible to flash these with a SS bios and get the same speeds etc or does it not matter.

It's very possible but you would definately be voiding your warranty in doing so.

Considering the stability problems with the entire 7900 line right now, I'm just content to let mine run at GT CO SC speeds. :p
 
MerlinsApprentice said:
Got my BFG Tech 7900 GTX OC 512 on the 31'st
Loaded up Trainz2006 the next day and ran into the blinking screen of death
Tried doing 3dmark2006 oh boy what a mistake.
The card had only been out of the box less than 12 hours and it was already garbage.
Just got finished talking with the tech at BFG.
He apparently he has no clue that these problems have been going on.
First he suggested it was my power supply
then he suggested it was the video drivers didnt support the card
/rollseyes

almost $700 for a pile of junk
will not be buying another card from BFG

I agree. Their RMA service is good as they have RMA'd two cards in a row for me.

But the person on the phone would not even respond to my asking about the issue with the slightest knowledge of there even being an issue. They also tried to push the power supply thing with me.
 
I have had 3 7900GTX's from E-VGA.
The only one I have kept has worked flawlessly and I think I am lucky for that. The other two I bought to complete an SLI setup(Bought 1, worked like crap, returned it and bought another one a month later, which was also returned). I noticed problems as soon as I attempted SLI and when I tested the cards individually, it was the new card I had purchased both times causing all of the problems. So the problems, like others have said aren't just limited to the overclocked cards. Its the stock clock ones as well.
 
Blue Falcon said:
I just found out from someone at eVGA that many of the 'fixed' cards that were sent out as replacements are actually 'Signature Series' cards that have had the BIOS re-flashed to the speeds of the model it's replacing before they ship it out as a replacement.

If you've revieced a 'fixed' replacement card for your defective eVGA card and it has a barcode sticker on the back that says 'Made in china' and has half of the number marked out with permanent marker, then odds are you have actually recieved one of these rebadged signature series boards.



So yesterday I decided to start some testing to see how this card runs at SS speeds (600/1600). I ran some 3DMark 03,05,06 for several hours as well as some F.E.A.R. and had great results at these speeds. And then before bed last night I left the PC running a loop of 3DMark 06's Shader Model 3.0 tests. (The water monster one and deep freeze one) and this morning it was still runnign them perfectly with no artifacting or glitching whatsover. Wow... :)

Wow, how did you get that BIOS? cant find 49 anywhere.
 
DangerIsGo said:
Wow, how did you get that BIOS? cant find 49 anywhere.

It's not .49 and as far as I know a .49 doesn't actually exist, it's a misprinted label apparently. The BIOS that came on the card was .45. :)
 
Got my replacement non oc'd 7900 gtx after 5 days. Guess what still doesn't work properly? :( Same crap random artifacting. In wow, if I alt-tab out of the game and click it again it fixes the problem for a random amount of time. What the heck is going on!! It doesn't sound like anyone in these company's nor nvidia knows what is wrong or if they do they don't care. Customer Service can only go so far in retaining customers.
 
I'm in the middle of deciding between a X1800XT 512MB or the 7900GT....all this makes me want to go with a louder but reliable ATI unit.....not sure what to think here. Certainly don't want to end up dead in the water (need my computer for work as well).... :(

Too bad that there seems to be a hit and miss when you order. Who can guarantee which card you get?!

This is not good for nVidia at this point.
 
ScYcS said:
I'm in the middle of deciding between a X1800XT 512MB or the 7900GT....all this makes me want to go with a louder but reliable ATI unit.....not sure what to think here. Certainly don't want to end up dead in the water (need my computer for work as well).... :(

Too bad that there seems to be a hit and miss when you order. Who can guarantee which card you get?!

This is not good for nVidia at this point.

Very true... Fortunately my eVGA 7900GT CO that I got via their Step Up plan has been trouble free so far.

I'm rather disappointed that they haven't seen fit to upgrade the cooling solution - mine shipped with the ever-so-dinky reference cooler.

As far as ensuring which card you get (working or not), I suspect that it's not totally brand "X"'s or Nvidia's problem as I think that all of the 7900's are manufactured/assembled by an outfit called Flextronics... They're the ones that might be having the QC probs.

But if you do go 7900, make sure that you go eVGA. They will actually cross-ship you a replacement ASAP w/ no BS, if it's a dud.
 
skippy9146 said:
Very true... Fortunately my eVGA 7900GT CO that I got via their Step Up plan has been trouble free so far.

I'm rather disappointed that they haven't seen fit to upgrade the cooling solution - mine shipped with the ever-so-dinky reference cooler.

As far as ensuring which card you get (working or not), I suspect that it's not totally brand "X"'s or Nvidia's problem as I think that all of the 7900's are manufactured/assembled by an outfit called Flextronics... They're the ones that might be having the QC probs.

But if you do go 7900, make sure that you go eVGA. They will actually cross-ship you a replacement ASAP w/ no BS, if it's a dud.

Or skip the drama.. get an x1800XT... wait till G80 and R600.. upgrade to R600.. use the x1800XT on an RD600 board as a physics accelerator.

Wow... don't need to RMA a card or buy an expensive Ageia Physics accelerator down the road. Just use your current ATi card to do the work..;)
 
ElMoIsEviL said:
Or skip the drama.. get an x1800XT... wait till G80 and R600.. upgrade to R600.. use the x1800XT on an RD600 board as a physics accelerator.

Wow... don't need to RMA a card or buy an expensive Ageia Physics accelerator down the road. Just use your current ATi card to do the work..;)


Sounds almost like a reasonable thing to do....oh horror!!!! A reasonable post on the [H]....doom is near....ah well....it's 6.6.06
 
Blacklash said:
Props to Evga.

Now when I do buy a nVidia product again I know who to buy it from. Companies that try to ignore problems or sweep them under the carpet do not impress me one bit.

Again, good job guys, trying to find a solution and showing that you care more about your customers than creating some BS illusion while they suffer.

Oh yeah and /salute to [H]ard for sounding off on the issue.


Exactly the consumer isnt stupid XFX just closing their forum was bad PR worse than people seeing they had some issues.
 
Nah, I wouldn't call it a worse issue than people having borked cards, but it sure as hell doesn't look good for XFX. I don't know about you, but I'd rather have somewhere like eVGAs board where I can try to get help/etc.. Than call some telephone # and get into an arguement w/ an overworked line-staffer because they blame the end user for the issues (I'm not saying XFX is doing this..just an example).

It's too bad the cards still aren't fixed though..Seemingly all eVGA did was add some more vregs (which upped the core to mentioned 1.45v [drastically upped load temps]), and change the cooler which has WORKING fan control via RT (( READ: shipped out re-badged N569 Signature cards w/ BIOS flashes)). But at least they've tried something, and have been *readily* accepting RMAs without question really (and paying shipping both ways), haven't shut down one of the major communications points (message boards), etc..Now if only the issue of supposidly known bad batch of vregs were confirmed and when cards with new batches could be expected.. Had I of known all the problems I'd be having getting games like Oblivion to run right, I wouldn't have fucked with the Step-Up, and kept my 7800gt/gotten a 2nd hand one to go w/ it in SLI.

I sure wish eVGA would pull an MSI (?) and go bi-GPU (err huh? :p) -- make cards for both the green side AND the red..It'd royally piss off NVIDIA I suppose, but at least people would have a choice in card types, AND the great warranty that even covers overclocking and cooler swaps (not that I'm messing around with that until I have a known GOOD card!) that eVGA gives.
 
kush said:
Nah, I wouldn't call it a worse issue than people having borked cards, but it sure as hell doesn't look good for XFX. I don't know about you, but I'd rather have somewhere like eVGAs board where I can try to get help/etc.. Than call some telephone # and get into an arguement w/ an overworked line-staffer because they blame the end user for the issues (I'm not saying XFX is doing this..just an example).

It's too bad the cards still aren't fixed though..Seemingly all eVGA did was add some more vregs (which upped the core to mentioned 1.45v [drastically upped load temps]), and change the cooler which has WORKING fan control via RT (( READ: shipped out re-badged N569 Signature cards w/ BIOS flashes)). But at least they've tried something, and have been *readily* accepting RMAs without question really (and paying shipping both ways), haven't shut down one of the major communications points (message boards), etc..Now if only the issue of supposidly known bad batch of vregs were confirmed and when cards with new batches could be expected.. Had I of known all the problems I'd be having getting games like Oblivion to run right, I wouldn't have fucked with the Step-Up, and kept my 7800gt/gotten a 2nd hand one to go w/ it in SLI.

I sure wish eVGA would pull an MSI (?) and go bi-GPU (err huh? :p) -- make cards for both the green side AND the red..It'd royally piss off NVIDIA I suppose, but at least people would have a choice in card types, AND the great warranty that even covers overclocking and cooler swaps (not that I'm messing around with that until I have a known GOOD card!) that eVGA gives.

I finally have two cards running, keeping my fingers crossed til my 7950's are ready
 
My eVGA 7900GT KO started acting up today. I know it's not heat since my VF900 is keeping it below 50C under load and it's not power since my PSU supplies 35A on the 12V rail and ran my 7800GTX flawlessly for 6 months.

After playing games the past couple days, suddenly things started tearing and artifacting really bad after a few minutes of playing time.

At least eVGA has a lifetime warranty and off the card goes once my cross shipped replacement arrives.

Did I read right somewhere that the replacement cards come at 1.45V? That a lot of these problems could be due to a faulty voltage regulator?
 
skippy9146 said:
But if you do go 7900, make sure that you go eVGA. They will actually cross-ship you a replacement ASAP w/ no BS, if it's a dud.

Monday I applied for RMA. Still no new cards for me.

Not that quick of a turnaround if u ask me. :(
Pretty average compared to my 9800XT i RMA'd.

Child of Wonder said:
At least eVGA has a lifetime warranty and off the card goes once my cross shipped replacement arrives.

Let me know how long it takes to get your replacement. I've been waiting for a working set, and the anticipation is killing me. :eek:
 
After reading through all the stuff, I thought I would throw in my two bits. I was nVidia guy since the Riva 128, through the TNT, TNT2, Geforce 1 thru 4 and up until the the ATI 9700pro came along. I ran ATI cards for a few years and was running an X850XT PE when I decided I wanted to run a dual card system and I had to make a choice between SLI and Crossfire late last year. In light of the less than stellar Crossfire availability at the time I decided that SLI was the way to go. I built a system with a A8N32 and a couple of BFG 7800GT's December. Worked great and I was pretty happy with them.

A few weeks back I decided to "upgrade" to the 7900GTX's... oops. Long story short, I went through 3 eVGA 7900GTX EGS cards (which run at spec) in less then two weeks. Oblivion, FEAR, and Quake 4 all had the privilege of killing a card. They lasted between 5 and 15 hours between them. None of them were OC'ed prior to failure. Once they were screwed I took the liberty of messing around a bit, and found that in each case the core clock made no difference (usually stable to 700+ mhz) on the problem however underclocking the ram 600mhz (1200 DDR) was the only "fix". I've returned the cards to my vendor for a refund.

I've been building PC's for 20 year, and I've never seen such a wide spread manufacturing issue on "high end" hardware. Kudos to eVGA to stepping up to the plate and admitting there is trouble. That is the only thing keeping me from running off to a new CF 3200 board ATI cards today. Either way, I'm going to wait a bit and see what happens. If nVidia gets their shit together and fixes this problem, I'll give them (because of eVGA) another run. FYI... my eVGA 7600GT in my PVR has been fine... The thing that sucks the most is I only have one of my 7800GT's in my box now. :(
 
Blue Falcon said:
Taken whilst running the Real Time High Dynamic Range demo:

V.GPU:
new7900gt_vgpu.jpg


vGPU measured at 1.2v while at the desktop, 1.45v when a 3D app is running.

V.MEM:
new7900gt_vmem.jpg

The other day I downloaded 7900GTX bios from MVtech to try to see the diff in the various manufacturers. I downloaded the stock and overclocked versions of XFX, BFG, EVGA, Asus, and PNY since I had 1 (non-oced vers.) go dead already. They all had the same core voltage, except the oc'ed PNY, for a mere 20mhz o-clock they had raised the core voltage to 1.45v from 1.4v. CoolBits autodetects both cards at 675/1800, even now that one is dead. The night I had one die was cool with the window and case open, the core temp never exceeded 47c
 
Bio-Hazard said:


Hmm interesting, especially since the only one I see is for the signature series card. Not really suprising given the established fact that eVGA in several cases shipped out re-flashed signature series cards to replace bad GT CO and GT KO cards. .45 is probably whatever GT CO BIOS they had had handy to flash onto the card before they shipped it to me.

Thanks for the linkage, if I get really brave I might flash the signature series BIOS onto the card to see how it runs. :p
 
Hi everyone,

I have the same problem most of u, but wit an ASUS 7900GT card. I've been suffering from the above already mentioned flashing, freezing, artifact problems, but finally I found a solution (which is only half the battle) that worked for me (and has been mentioned above, but not in detail).

First step is to disable that stupid PEG link mode if you have new ASUS motherboard. What it did to my PC is that since I disabled that option the computer become stable under most games and 3DMark 05 and 06 (it is a huge achievement) when using factory frequencies 450/660. BUT not in F.E.A.R.

Second step is to DC the RAM frequency to 600MHz. AND IT WORKS even with F.E.A.R.! At the same time I could OC the GPU to 564Mhz and no problems at all. So I have to disagree with everyone who blames the voltage of the GPU, cause mine goes with 1,2V.

Am I happy to have to DC my RAM? Definitely NOT! But as I can NOT do an RMA because of various reasons (not US citizen is just one) I am happy, that finally F.E.A.R. does not screw up my PC anymore.
I've been testing for quite some time now, and found that OCing RAMs doesn’t do as much good as OCing the GPU. So my card is currently much faster (about 10-15% in 3DMark) with 564/600 than with the original 450/660. AND STABLE!!!

Next step I'm going to do is cleaning the back of my card. Why? Because last time I did it helped to raise the RAMs frequency. Why? I have no clue, but it worked. And I also might try to raise the voltage of the RAMs, but there is no Radio Shock within 10000km, so I have to find a place where I can get hold of a digital multimeter. But the question remains: what the heck is happening with the RAMs? Is it the residue on the back of the card, or is it the heat the RAM modules generate even on factory frequencies or some voltage moderating something on the card? Donno

Hope I could be some help (as this thread was very much helping me) until u guys wait for the RMA.
 
Just wanted to post my experiences with my XFX Geforce 7900GT
The card lasted for about 20 min. while using stock drivers.
After about 20 min. playing Q4 I would get artifacts and strange lines. I quit the game and when I get back into windows my display is going black for a few seconds everytime I change something on the screen. e.g. clicking the mouse button.
I upgraded BIOS, drivers, but after 10 - 15 min of playing any game the same issue occurs.
I contacted XFX and put in a support request but this has not been answered yet. This is 4 days ago, and I am still sitting here with a useless card and no answer from XFX. :mad: :confused:

I am not happy to say the least.
 
BFG 7900GT OC working for about 2 weeks similar problems ran 3dmark06 fine for 2 weeks and games (WOW, Oblivion, Far Cry). Now in 3dmark06 deep freeze crashes and I am getting artifacts and crashes in Oblivion. Chips on the back were way too hot to touch.

Anyone have any reccomendations on a more stable Nvidia card?
 
Hi, I read this topic for a few days, I'm impresed the way of eVGA work with the users and help for them. I have a Leadtek 7900GT and my card run great about 4 days and after that go crazy, screen start to blank when I exit the game, artifacting in Deep Freeze, crashes during gameing... becouse I live in Poland I can't even dream about such thing like fast RMA and good working card. Only hope for me is buy a new card from ATI. Meny of my friends from ouer forum have the same problem with XFX OC-ed cards, always have NVidia card on board but now its time to change ...
 
Would these same issues affect the 7800 GTX or the 7600 GT ? I'm not really that hard-core into gamming (yet) but I'm looking for a graphics card solution for a Dell XPS 600 which I bought barebones (case, psu. mobo). I'm pretty sure that it was originally configured with a 7800 GTX. Wondering how much performance difference there is between a 7800 GTX and a 7900 GT and if I'd get a moe stable video card with the 7800? I'm planning on running the Dell through a samsung 24' widescreen FP.
 
Back again, and this time with my RMA'd card testing. Luckly it's my first RMA and I have to say, after 3-4 days of some very lax testing, things seem to be keeping together nicely. Albeit, I replaced 3 of my 120mm fans with TT smartfans(up to 93.7 CFM!). Little louder, these fans, but they instantly reduced the ambiant temperatures 4-6°, with actual component temps dropping by more.

Because I started working with the RMA'd card before installing the fans, I noted an incredible decrese in heat from adding the additional air flow. Using my TT Armor case, with two icage's almost level with the card and one of the cage's stripped of HD's so the fan was pointed right at it (same set up with my previous 7900 GTX), I got idle temps around 55-56°, after game temps got to about 66-69°, Synthetic benchmarking things got up 77° and even as high and higher than 85°.

Now with the new 'Wind Tunnel' in effect, all installed and blasting away, things that tend to get hot have a much harder time doing so. Idle temps on the gpu has dropped to 46-49°. Load temps after games or movies never get hotter then 55°, and I've noticed the climb towards warmer weather takes the gpu much, much longer and it can't seem to get over a 78-79° hump.

In conclusion, I've stated all of this because it is has been a big difference from a card that lasted a little around 12 hours to a card thats been pushing along for over 80+ hours without a hitch, because quite frankly, thats the only difference. Overclockers.uk spent all of 4 hours testing my old card and sent out the RMA. Nothing about the RMA card itself has changed. Even though I mentioned the Bios version of the previous card to them, I recieved the exact same Bios version with the exact same inconsistancies of it's listing in various places (Tuth be told, I expected more from Gainward). PCB layout is the same and so are the drivers and system I used before. Airflow is the only thing that has changed.

Could it be that the title of this article was spot on? Is it as simple as overly abundant air flow? Whether either are the case, it was sad seeing this article removed from a sticky with no clear resolution from manufacturers or the staff at [H]OCP and its eventual drop to the wayside. I guess that's life here on or near the 'bleeding edge', whatever problems you might be having with a current product, rest assured a newer one will replace it soon enough and there's really no need for the manufacturers to be held accountable.
 
Avat said:
... I guess that's life here on or near the 'bleeding edge', whatever problems you might be having with a current product, rest assured a newer one will replace it soon enough and there's really no need for the manufacturers to be held accountable.
The manufacturers SHOULD be held accountable if they are willingly selling broken products to consumers and they know there is a problem.
 
my guess is that all cards with the faulty voltage regulators (which may be ALL cards except the very recent ones) will eventually go bad. The best way to assplode your card is to run multiple "deep freeze" 3dmark2006 benchmarks in a row. don't choose the "loop" option in the settings (for some reason this gives the cards just enough breather to not poop their pants), just the "repeat each test X times." Select just the deep freeze test, and put the repeat on some large number. Start the test and come back in 60 minutes. If you aren't artifacting after that it is likely you are safe (for god knows how long). I am curious to see if some once "stable" cards start crapping their pants if they run this test. Report in folks!

PS - I really really really don't think this is a temp issue. Both of my bad cards have never hit abnormally high temps. I also don't think sticking heatsinks everywhere will save the day. This is all just theory based on my experiences and those of a few others on the forums, so please help me out and give me some results.
 
y0bailey said:
my guess is that all cards with the faulty voltage regulators (which may be ALL cards except the very recent ones) will eventually go bad. The best way to assplode your card is to run multiple "deep freeze" 3dmark2006 benchmarks in a row. don't choose the "loop" option in the settings (for some reason this gives the cards just enough breather to not poop their pants), just the "repeat each test X times." Select just the deep freeze test, and put the repeat on some large number. Start the test and come back in 60 minutes. If you aren't artifacting after that it is likely you are safe (for god knows how long). I am curious to see if some once "stable" cards start crapping their pants if they run this test. Report in folks!

PS - I really really really don't think this is a temp issue. Both of my bad cards have never hit abnormally high temps. I also don't think sticking heatsinks everywhere will save the day. This is all just theory based on my experiences and those of a few others on the forums, so please help me out and give me some results.
Apart from the cooling issue of one single Mosfet, which can be easily cured with a small heatsink, there is no prove whatsoever that "Faulty Voltage Regulators" are causing any problems at all.
What is absolutely increasing the problems, is Coolbits. After having removed Coolbits, all symptoms of a dying card ( flashing screens, artefacts, freezing) have gone, and my 7900GT behaves again as it should do, for some time.
I have changed the GPU/MEM frequencies directly in BIOS with Nibitor.
 
PS - I really really really don't think this is a temp issue. Both of my bad cards have never hit abnormally high temps. I also don't think sticking heatsinks everywhere will save the day. This is all just theory based on my experiences and those of a few others on the forums said:
Results: I have cleaned the back of my card (100% alcohol) and it did not have the problems any more (see my last post at page 10). It was running at 564/800 for two weeks, stable 24/7. Afterwards I had to take my PC apart and cleaned the card again (stupid), than put it again together, and I had the same problems as before, it become stable only at 564/600 again. So what I did I cleaned the card AGAIN extensively and now it works AGAIN stable at 564/800.

Tell me about it...
 
just started Cross Ship RMA, my cards Discoed up on Far Cry and began artifacting when I wanted 8x AA in ANY game.
 
Back
Top