I suspect this is a GPU problem rather than Power Supply issue, but here goes.
I'm running mainly a gaming PC, on Vista Home,
CPU is an Athlon 64X2 Dual 4200+ proc
PS is a Corsair HX750
GPU is a Diamond ATI Radeon HD 4670
Just over the last week, I've had 6 total shutdowns while playing one particular game, Mount and Blade.
The first 4 times, I could reboot easily, no problems, but after a period into the game, the system shuts down again, each time after a shorter period.
Now there is a long wait before I can even get the GPU to respond.
I'll explain this in more detail:
The system starts normally, the opening prompt screen comes up, but almost immediately the monitor LED goes from green to red, and it's just black, nothing.
The system seems to be running normally, all fans are running, but no video, not even the "NO SIGNAL" box on the monitor.
After waiting a very considerable time (around 5 or 6 hours anyway) I can start up the system normally, and use it for internet (as I'm doing right now), but if I start the same game, almost instant system shut down happens now.
This is not just a crash, the system shuts off instantly and completely as if the power was shut off at the supply.
I thought this must be an overheating problem with the card, so I purchased an Arctic Cooling GU cooling fan/heatsing for this card (Accelero L2 Pro), and installed on the card, it's running, but still I get shutdowns. Twice since the installation.
Now, for anyone familiar with this particular game, I fairly recently (roughly last 8-9 months) experienced a severe drop in frame rate in heavily wooded areas, so much so as to have to avoid these maps completely, as the game is effectively unplayable, this was not the case at first when I began playing easrly last year.
Both the card and the power supply are approximately the same age, purchased around March of 2009.
Is the card damaged, or would this be a PS issue?
Could the card have been gradually worn out/overheated to the point where it will now no longer function properly?
And I have yet to try a different game, since the peroid of rest needed to restart my system effectively is so long, then I have to use my wife;s laptop, which is quite slow.
I will attempt to run another game shortly, but expect a similar system failure.
Thanks for the suggestions.
I forgot to mention that it's the 1 GB card, though I doubt that matters.
After about 2 1/2 hours cooling I did try a couple of different games, first a low intensity graphics game (Pirates!- an old favorite), then after about 30 problem free minutes, I tried Empire:Total War.
I got as far as the opening movie, then instantly got a reboot and black screen.
I downloaded those apps and tried testing, but seconds after Furmark started testing, I got the black screen and red monitor LED again.
The idle temps were 46C and 52C before I tried Furmark.
I'm on the wife's laptop at the moment.
I don't have access to another PCI-E card or another similarly equipped PC, so it's just guesswork here.
I strongly suspect the card, but I really would like to be sure before spending more money (which I can't do until the end of the week anyway).
Corsair has such a good rep, I wouldn't like to think it's the supply, but it's puzzling that the card works for normal tasks and lightweight games, but suddenly quits when tapped for more strenous operations.
Weird that the card could shut down the system, but then again, I'm certainly no expert in these matters, just self taught with some common sense and a fair amount of research.
I fully appreciate that no-one can really accurately diagnose a distant problem without any access, just looking for supporting opinions, hoping that more experienced folks can offer likely causes and either confirm or refute my suspicions.
More info I forgot to mention:
The stock fan on the card was starting to act up, getting pretty noisy, and clacking against the plastic shroud over the heat sink.
Screen resolution was a mild 1024X768.
When I took it apart, I found that there is quite a lot of play in the shaft, and spinning the fan manually produces some vibration noise, so it has obviously been going bad for a while.
The card or CPU have not been overclocked by me in any way.
I am planning on getting a new card as I said, at the end of the week when I can afford it.
I'm just looking at replacing the same card, since a major upgrade would render the new cooling fan/heatsink useless, it only fits a certain series of cards, and the highest ATI card it is built for is an HD 4850, which doesn't actually seem to offer any performance benfits over the 4670 at all, possibly less performance.
I also am wondering if overclocking the new card would be advisable: would the bigger cooling apparatus be effective enough to allow some extra performance boosting?
edit: Hmmmm....just checked the GPU temp in Catalyst Control centre, 91C at 0% activity. Seems kinda overly warm to me, maybe the contact between heatsink and GPU isn't what it should be. It's just through the pre-applied paste. I should try to re apply some new paste when I can get some.
^^ When/if you are to power up again, check that the fans are running. Do not stress anything. let it boot and idle and download CPUID's Hardware Monitor. Watch your core and gpu temps carefully. If the machine keeps running, try and stress it a little - just a little - and keep watching the monitor. Keep increasing the stress slowly, and see if your temps tell you where the problem is, cpu or gpu.
NO NO NOT A HEAT RELATED PROBLEM
i had this too a while back it's a problem with VRAM
if your graphics card is still in warranty then go to diamond for warranty coverage
if you want to be sure then pop in a friend's graphics card and see if you still have the same problem
When it shuts down, is it an immediate shut down or a software driven shutdown?
If you disable the "Automatically restart" on system failures (Google it) do you get a Blue Screen of Death instead? (by default BSoDs are not shown, they just cause a restart).
Once you know that it'll be much easier.
If it is an immediate shutdown then chances are that it is the Power Supply Unit, they have a higher failure rate than graphics cards that have never been overclocked. (The 2D vs 3D power load is normal too, the fan should spin up during heavier 3D workloads and the draw on the card should rise, past a certain point or after long or 'hot' enough your PSU might have just had enough).
ED: "91C at 0% activity", OK it sounds like your video card is stuffed.
All fans are running fine, the Arctic Cooling fan is very quiet compared to the stock card fan, but I slip the case side off and check the fans every time, they're all running fine.
The card's over 1 1/2 years old, no warranty.
It's a complete system shutdown, just as if you hit the power switch on the power supply or hit the switch on the surge protector.
I loaded the latest CCC from ATI last night, and now the darn system won't boot at all, just sits in limbo with drives running and red monitor LED.
F11 system restore doesn't even respond upon booting.
GRRRR!!!! Where's my hammer!!!! X(
Definitely the card.
I'm running (limping) on a 10 year old Voodoo card until I can update.
The new Catalyst drivers seemed to really mess up my system, however.
I had to install XP as Vista would no longer work after loading the new CCC (10.7) and having the card finally die completely.