GTx 280 pink screen crash on first running 3D game from 2-3 hours rest.

[jincuteguy]

So recently I had "pink" screen crashings everytime I first run a 3D game when I first turn on my computer. Here is how it happened. So every night I turned off my computer and go to sleep. I woke up the next day and turned on my computer. Everything boot up fine and I logged in to my desktop. Then I open up a 3D game which is Left 4 Dead. After I got in the game for like 3 - 5 mins I got a pink screen crash. I can't alt tab out I had to hard restart the comp. Then it works fine after that not a single crash. It just happens like that everytime I turned off the c omputer for at least 2 - 3 hours then the crash happen after I turn on the comp and open up a 3D game. I tried this with other 3D game too like World of Warcraft and same thing happened. I tried with both window XP and Vista (32 and 64 bit) and the problem still happen. I dont overclock my compter everything runs at stock speed.

My comp spec:

Intel Quad core Q9450 @ 2.66ghz
4G Corsair Dominator
EVGA GTX 280 (stock)
Powersupply Enermax Galaxy DXX series 850W
Asus P5K/EPU motherboard

I have no heating issue at all. My cpu cooler is Noctua Ibought it for $70. I have 5 x 120mm fans and 1x 200mm fan. I have RealTemp, CPU-Z, GPU-z running and the temps are fine. If anyone has this same problem or have any suggestions I really appreciated.

[scubes]

i have the same prob my pc goes on @ 9 in the morning the 1st thing i run is the crysis benchmark. it crashes after a couple of secs but theres times when it is fine games crash to like the model 2 emulater im playing with virtua fighter2. mine hasnt got a pink screen though like yurs its like lego on mine i think cos the pc is cold thats my problem bcos off being off for hours. if freezing here in northeren ireland @ the mo really cold.....

[jincuteguy]

Yea the weird thing is it only crash once after the first turn on. then if I leave my c omp on it runs fine the rest of the day or until I turn it off for 2 hours at least. Yea this started to happen just a couple weeks ago and it could prob the cold weather. Winter just started here too. But I really dont know what is the problem.

[Bee_Dee_3_Dee]

Jincuteguy and Scubes,

This can probably can help you, as far as confidence that a GTX280 can ROCK.

The game can be a BEE-ITCH- as if I needed to mention it. :D But it can be conquered!

Because I had big problems with Crysis and my GTX280 when I first played while OCing (my CPU), I compared Crysis with absolutely no OCing. As a result I found the solution for me was CPU Voltage. And I just LUVED the 280 after that. I had my GTX280 for three months. (Traded Up for a 295 anyway- IT JUST ARRIVED! :D)

All I did was add a little Voltage to my CPU. No Joke! Instead of 1.26250V @3.7 Mhz I used 1.26875V.

Here's how I figured out my solution:

The entire first month I had a new GTX280, I ran my Intel C2D E8400 Wolfdale System with all default settings except for the +.3 volts on Corsair Dominator RAM- to take it from 1.8V to 2.1V on my MOBO. It's required for this RAM to run at its intended 1,066 Mhz... Corsair's Tech support ROCKS!

Yes, if I don't overclock (especially in Crysis) I do loose performance in many games (20% slower in Crysis). But even though the Game ran much slower (as I would witness a month later while OCing the CPU @ 3.7 Ghz it ran 20% faster :)) not OCing was necessary for me in order to road test a new video card. What better way to proove that the hundreds of Dollars I spent on a video card resulted in a good quality product?

So, after a whole month of no crashing, while the CPU was @ factory 3.0Ghz defaults, I overclocked and gradually had some problems especially if playing for long periods in most games other than Crysis, and almost immediately in Crysis. Crysis crashed and crashed and I couldn't put two and two together.

I was like so, "%$*&$^@#|*" mad. So I googoogoogoogoogled.

But then I remembered that without OCing, the System Rocked. No crashes whatsoever. So I got out my OCing Spread Sheets and put on my Detective hat.

I asked a key question. What is different now? (I've never had problems with running anything with the CPU @3.7GHz.)

The answer was that in addition to a new GTX280 video card, I had gotten the best 620 Watt PSU several months earlier. And all my Spreadsheet data was aquired from a crappy PSU and my original 8800GTX. Yes, it was the same MOBO, RAM, and CPU, but changing a Video Card and PSU is too drastic of a change to continue relying 100% on the same data.

So I ran Crysis again without OCing, and it didn't crash. The game was obviously slower, but it didn't crash. Then I put the CPU back to 3.7Ghz and used the Voltage that use to always work for me, but it crashed. So in addition to the CPU being back to 3.7Ghz, I added more Voltage than what use to always work for me. And with the Voltage increased from 1.26250V to 1.26875V it was like a miracle because zero crashes occurred after that. AND CRYSIS ROCKED TOO. That extra 20%+ boost in performance (3.7GHz vs. 3.0GHz) is a must. And the PC, was all of a sudden, indestructible again from all games including Crysis.

Here's the proof:

What are some of the craziest things to do in Crysis, if you want to avoid crashing?

Answer:

Repeatedly, and at free will, Quick Save and Quick Load. (Plus Alt+Tabbing, over and over.)

For three days, I tried so hard to crash Crysis after this PC was properly configured, that I was dumbfounded. It saved and reloaded every time. w00t! I treated it as if it was the original Half-Life game running on an nVidia 6, 7, or 8 Series Video card.

You know, like when you say 'that was cool.. i got to do that again'. And then you Quick Load to your hearts desire. And no crashing... IN CRYSIS... w00t!!!

Or, like when you say 'that was cool.. but let me try saving a fraction before'. And then you Quick Load again and again until your content enough to move on towards other challenges. IN CRYSIS... ON VERY HIGH... w00t!!!

It can be done, on Very High settings now, with a Single Video Card.

My system with a GTX280, and a CPU properly OCed cannot be stopped by Crysis anymore! NOT ONE Lockup or super Lag, no Crash to Desktop... nothing!

Did I say? 'It's alive'! Muahaha.

And now I've got a GTX295- I've not installed it yet because in order to be safe, I need to wake up well rested, and in a good mood to install it, along with a new 800Watt PSU.

Yes the GTX280, is just barely fast enough for Crysis, but it will run it on Very High or Enthusiast in Warhead; at 1680x1050 it'll keep dipping just below 20 FPS minimal, and as high as mid 30 FPS. Most people consider this the minimal playable Frame Rates. Looks so sweet! Barely dipping below 20, while all the craziest stuff in the game happening, is just fine.

http://www.pcgameshardware.com/aid,672323/Reviews/Geforce_GTX_295_-_Review_of_Nvidias_final_dual_graphics_card/?page=7

And anyone that has not Tweaked their OS and Problem Solved a troublesome FSP Game with help available at TweakGuides.com (http://www.tweakguides.com/), has Missed The Boat. You must check it out. (My inspiration that OCing was maybe to blame came as a result of revisiting TweakGuides.com (http://www.tweakguides.com/) for the first time in a long time and re-reading everything about Crysis there.)

GL and stay confidence because can be done. :)

[jincuteguy]

So your suggestion was to increase the CPU voltage a little? But Im not OCing my system. Im running it at default speed and everything and it still crash. But it only crash "once" everytime I first turn on my system for the day. For example, I woke up and turn on my computer. Then fire up a 3D game such as World of warcraft or Left 4 Dead. then after being in the game for 2 or 3mins it crashes. Then I have to hard restart. Then I open up a 3D game again and play . And it works fine after that the rest of the day as long as I dont turn off my comp for like 2 or 3 hours.

[jincuteguy]

Here is how the pink screen crash look like
http://38.99.76.18/img218/1195/dsc00763qh0.th.jpg

[xorbe]

The evga forums are full of pink/purple screen camera shots. RMA the card I think. PSOD syndrome. Hopefully you registered it so they don't look the other way.

BD3D, step away from the Mountain Dew...

[Bee_Dee_3_Dee]

The evga forums are full of pink/purple screen camera shots. RMA the card I think. PSOD syndrome. Hopefully you registered it so they don't look the other way.

BD3D, step away from the Mountain Dew...

:angel: Yesterday I had to crash before installing my new 295 and PSU. But I didn't want to. :cry: So I procrastinated! :type:

I Just now finished installing the new monster PSU. :D IT's ROCKING! :super:
(BFG Tech ES SERIES ES-800 800W) :bounce:

Next, the 295! w00t!!!! :dance:

:jumping:

MY CRYSIS! :box: MUAHAHAHAHAHAHA!!!!! (nana2)(nana2)(nana2)(nana2)(nana2)

[jincuteguy]

The evga forums are full of pink/purple screen camera shots. RMA the card I think. PSOD syndrome. Hopefully you registered it so they don't look the other way.

BD3D, step away from the Mountain Dew...

True but my problem is not the same as theirs. My problem is it only crash once everytime I first turn on my c omputer and first turn on a 3D game after an hour rest.

[xorbe]

That's because it heats up and "fixes" itself via thermal stress. It won't last forever, probably.

[jincuteguy]

What do u mean by it heats itself up and fixes itself?

[xorbe]

When something heats up it expands. You probably have a broken part. But when it heats up, it expands, finally pressing the slightly disconnected parts together. Then when it cools down and contracts, the broken area stops touching, and the problem repeats.

As you say, you have to turn the machine off for a couple of hours for it to happen again.

[Bee_Dee_3_Dee]

Jincuteguy,

Based on the symptoms, and the correct explanation that xorbe is giving you, the following is as common as being told to reboot. ;)

Eliminate individual hardware as the cause. There are two ways to approach it.

(1) The first one is to disconnected all unnecessary hardware, and see if the problem goes away. If it goes away, then reconnect one by one and when the problem reoccurs then that last reconnected hardawre is the culprit.

(2) The second is to disconnect one unnecessary piece of hardware, and see if the problem goes away. And then another, and see if the problem goes away. Then another, ... etc. Until the problem does goes away. If the problem goes away then the last piece of unnecessary hardware disconnected was the culprit. (Hopefully a $29 DVD-RW Drive. ;) Or better yet a $3.99 SATA cable. :))

Which approach to take depends on the situation.

And cables not being properly seated is also a very common cause- with the symptoms you've described and the expansion/ contraction from temperature changes. But disconnecting unnecessary hardware can lead to exposing the culprit. ;) Checking the connections of the necessary hardware is a must.

GL

[jincuteguy]

I only have a hard drive and a DVD burner in my system. But why is this have to related to the burner and hard drive? isn't it suppose to be just the video card? or maybe not the video card?

[xorbe]

Sometimes the root cause failure, say, over in your system memory, could result in symptoms of a crashing video driver.

Very doubtful it's the HDD or DVD. Most likely the GPU. However, you could try reseating the memory, the cpu, the gpu card, the data cables, the power connectors. If you have 2 sticks of ram, run with just one, and then with just the other. Try relaxing ram timings. Try an extra notch of cpu and ram voltage (safely of course). Probably none of this will help -- but -- you never know.

I bet if you plugged in an identical video card of the one you have, the problem would go away.

[Bee_Dee_3_Dee]

Sometimes the root cause failure, say, over in your system memory, could result in symptoms of a crashing video driver.

Very doubtful it's the HDD or DVD. Most likely the GPU. However, you could try reseating the memory, the cpu, the gpu card, the data cables, the power connectors. If you have 2 sticks of ram, run with just one, and then with just the other. Try relaxing ram timings. Try an extra notch of cpu and ram voltage (safely of course). Probably none of this will help -- but -- you never know.

I bet if you plugged in an identical video card of the one you have, the problem would go away.

Definitely try everything xorbe just said jincuteguy.

And BIOS mite be corrupted too. Load Fail Safe or Optimal. Even Clear CMOS too.

Plus, Does the newest bios update address any issues related to stability?

And be sure not to be using the wrong UPS. I Just had to throw that one in- but most all older UPSs are not compatible with the big PSUs nowadays.

Just make a list of everything you've tried jincuteguy.

Almost forgot, the 4G Corsair Dominator RAM can be tricky but Corsair's support ppl rocked every time i've ever spoke with 'em. And now that i mention it, my latest BIOS updates for my MOBO made the 4G Corsair Dominator I have fully compatible. (I just noticed we both have the same RAM jincuteguy.... well i got the 1066.)

GL

PS
And after it gets fixed perfect, learn Norton Ghost to help rule out the OS in the future.

[jincuteguy]

I rma the card today, waiting for the replacement one to come next week. right now im using a 9600GT card and it's working fine, no crash whatsoever.

[Martin Benca]

Hi.

RMA the card if you want to get rid of this. I have exactly the same problem but since I have 3SLI config I could isolate it. Only one of my GTX 280 cards causes this. It is _not_ overheating, neither power supply or anything else. It is the card. Generally it is not totally broken but right at start cannot run at full frequencies (so quick fix for people having this problem is simply underclock it a little).

Mine after the start when it is cold (was off for a longer time) works ok at 602/1204/1000 clocks but crashes even at standart (602/1296/1107). To make it work stable it is not sufficient or even necessary to (fully) warm it up. You have to let it run for about 3-4 minutes to slightly warm/whatever up (it doesn't even have to run some heavy 3D application, however it does need some warming/time), then turn off and on PC for just a slight moment (sleep on/off is enough). After this all is _perfectly stable_ and I can even easilly overclock it to 675/1350/1242 with absolutely no problems at all.

Just warming it up won't solve the problem. Even if you'll run full 3D for hours underclocked getting the whole system to high temperatures, once you increase the clocks you'll crash. You _have to_ do this on/off thingy to get rid of the problem and you can do it after just a little warming. You don't have to wait long. After that all will be ok.

Either way since only one of my cards causes this it is _definitely_ some hardware problem. My wild guess is some defected part needs to charge and reset (some tyristor or whatever to provide additional line of correct power supply to the chip or something so it can run stable even at higher frequencies).

I didn't RMA it myself since I'm quite happy with my cards overclocking and temperature-keeping abilities and at "standart" clocks this happens only in certain games to me (typicaly WoW title screen reliably crashes immediatelly if this problem is there) so I might have problems to prove my case and I'm not in the mood being without one of my cards for some longer time. I wait until it dies for good then I'll RMA it. If not I can live with it, GT300 is commng soon anyway. :-D However the problem I have now for three-four months and it doesn't seem to get worse.

Usually after PC start I just read some web first then I simply do sleep on/off and can run at even 675/1350/1242 without any troubles. Or if I want to play immediatelly I have Riva Tuner shortcut to underclock it temporarily to stable 602/1200/1000.