problem with GTX 295? Or PSU?

grambowk

grambowk

grambowk

Associate
Joined
18 Jan 2010
Posts
2
I've been having some serious problems when playing games that result in the entire PC freezing up, including keyboard and mouse. The only thing I can do at this point is to hard reset the machine.

There are no recorded events in the event viewer. Interestingly, any music that was playing in-game at the time of the crash keeps on playing after the crash.

The crashes occur within 5-10 minutes of starting to play.

Here's a quick run-down of my system.

Asus Rampage II Extreme motherboard
12 GB RAM (6x 2GB Corsair 1600 MHz dimms)
One Gainward GTX 295 graphics card
Antec 850 W PSU

At first I suspected the graphics card (the fan was going fast so I suspected overheating). Interestingly, when idle the gpu temperatures are around 80 degrees, which seems a bit high. However, the crashes can occur when the GPU temp is as high as 100 or as low as 85. Additionally, on the advice of systems support I ran a furbank test and even though the gpu temperatures went up to 109 degrees (which is higher than the limit of 105) I didn't see any crashes. So maybe it's not the graphics card.

So maybe it's memory. I did an overnight memtest (7 passes in all) and not a single problem. I've removed 3 of the dimms and swapped and switched them around to rule out any one dimm.

At this point I'm at a bit of a loss as to what is causing the problem. The idle temp of the graphics card seems a bit high but there's nothing conclusive as to that being the cause.

Could the PSU be a little under powered? The requirement for the GTX 295 is 680 W (as I understand) and I noticed that at times the CPU is consuming as much as 150 W. So that's potentially 830 W and I imagine other components of the system could easily consume another 20 watts to push it beyond the 850 W limit. Unfortunately CPUID hardware monitor doesn't provide any details on how much power the GTX 295 is consuming.

Anyone have any ideas and/or suggestions I could try to isolate the issue?

By the way, I've got Win 7 x64 installed and all the temps on the CPU cores and the board are no more than 50 (for the CPU) or lower.
 
Well did it work fine before putting GTX 295 in? or has it just started playing up now? I would suggest new drivers e.t.c then maybe reinstall windows
 
It is possible that you are border line with your power supply.

I run the following:
a. overclocked i7 920 currently @ 3.8ghz (it is on most of the day).
b. 6gb RAM
c. Gigabyte EX58 UD5 Mboard with AKASA Nero Cooler
d. BFG 295 GTX on 24" Benq @ 1920 x 1200
e. I Play a game called IL2 Forgotten Battles.
f. Windows 7 x64 Pro.

To keep the GPU cool I use the EVGA Precision tool that can be used on all nVidia cards. I feel this is necessary because it can throw out some heat and to keep the temps down I need to run the fan at around 80% and sometimes more.

Oh yes. I use a 1000w power supply that I was told would be ample by OCUK when I sorted the system around a year ago.

If you really want to push you PC to the limit, I would suggest downloading and using the Intel Burn Test program.
There are many places to get it, here is one of many..

http://www.start64.com/index.php?option=com_content&task=view&id=2737&Itemid=76
 
grambowk said:
The requirement for the GTX 295 is 680 W
Click to expand...
Two hundred watts:
http://www.xbitlabs.com/articles/video/display/radeon-hd5870_7.html#sect0

Try running Prime overnight.
Also that many memody modules cause more stress so make sure memory voltage is that specified for them. (actually very few DDR3s work at JEDEC standard DDR3 voltage)

GunRunner said:
It is possible that you are border line with your power supply.
Click to expand...
As possible as your car having not enough power to haul your shopping bag unless you fit this under its hood:
http://i301.photobucket.com/albums/nn56/hchristopher/c6676eb3.gif
 
Interestingly, any music that was playing in-game at the time of the crash keeps on playing after the crash.
Click to expand...

Sounds like the card is hanging up as you say. When I OC too far on my card the display driver will stop responding. You can always try underclocking core and RAM to see if they are faulty and cannot work at the full speed.
 
This problem is exactly the same one I was having the other day when I installed my 5770 card, it kept freezing while playing COD4 the screen would suddenly go black and the game sound could be heard continuing in the background but I could do nothing but a hard re-boot.

I thought it was Windows 7 64 bit + 9.12 drivers +COD4 disagreeing with each other but I RMA'd the card and asked them to test it to determine if the card was at fault or the drivers.

They said today the card had been extensively tested( how I do not know ) and found to be faulty, perhaps yours is the same but I will say this they are not worth messing around with to determine a cause just RMA it and get another one and try it but preferably a different make to rule out a bad batch.
 
Thanks for all the responses. I've just run a test with IntelBurnTest, which went fine. The only odd thing was that it only detected 2.8 GB of RAM on my system even though I have 6 GB in now. If I try to select more I get a message saying I can't select more RAM than what is available on the system.

I'll try running Prime95 overnight (I'm assuming that's the tool being referred to).

Voltage for the memory is set to 1.65 but it is often below that when looking at the monitors (between 1.60 and 1.65).

I have all the latest windows 7 x64 drivers for the relevant components (including chipset, sound, graphics, etc...). And for good measure I applied the latest bios update.

I think I've pretty much exhausted all my options so unless the prime test shows something's amiss I'll RMA the card and hope they can find a fault with it.

By the way the system came with all current components installed and the only thing I did was install Windows 7 (upgrade from Vista). Everything worked fine until a month ago and it slowly got worse (at first the freezes would happen within 2-3 hours of playing) and now they're occurring within minutes so it definately seems like some hardware component is deterioating over time.
 
You must log in or register to reply here.
Back
Top Bottom