fans

[galaxyAbstractor]

My GPU gets really hot when I'm playing, and I think it is that what causes my BSoDs.

I had speedfan and flyff running for 30 min to check the temp, the GPU went to 56C and I think the System got too hot or something, as it showed -1C on them :/.

How would I best fix this without spending too much money? Currently I got 2 case fans (one at the top and one at the back), 1 CPU fan and 1 fan on the GPU. There isn't really anywhere else to put a fan, is there? What should I do? Buy a case with more fans?

 

[Sharky]

http://www.quietpc.com/gb-en-gbp/products/vgacoolers/ac-acc-l2pro

£11.30, so less than €13,- .

Set the currency to Euros at the top right, and you'll be well on your way.

I have the GX810 instead of the standard nVidia cooler on mine, and it's make a significant difference to noise levels. No more nVidia fans spinning at maximum RPM until the driver is properly loaded! The next noisiest thing is my 'silent PSU'...

EDIT: http://www.quietpc.com/se-en-eur/products/vgacoolers/ac-acc-l2pro < Sweden & Euros.

 

[Smith6612]

What GPU do you have? I've seen a bunch of 7 and 8 series cards from nVidia go past 90 Celsius. Heck, on my old 8800GTs (I still have one laying around as a spare or for selling), until I modified the cooling of them, the card would literally run at 95 Celsius with the fan speed at 45% and no artifact, however the fan was controlled by the card's BIOS so if the card saw it was getting too warm (98 Celsius and higher, thermal shutdown is apparently 108 Celsius), it would crank up the fan speed regardless of what the state of my PC was or what was loaded up, until I overrode the ran with RIVATuner until Core temp started to drop. Also, I've seen a GeForce 7300LE (OEM card) break the 110 Celsius barrier while running OCCT on it. Needless to say, that card I doubt will last long due to the temps it runs at. Next time I see the PC that card is in, I'll modify the cooling on that.

My GTX280s run at 48-55 Celcius under full load in my box with a little overclocking (which is usually when running Crysis or Folding@Home). Never given me issues. I applied some IC Diamond to the cards when I got them as the nVidia thermal paste wasn't the greatest on there.

 

[galaxyAbstractor]

What GPU do you have? I've seen a bunch of 7 and 8 series cards from nVidia go past 90 Celsius. Heck, on my old 8800GTs (I still have one laying around as a spare or for selling), until I modified the cooling of them, the card would literally run at 95 Celsius with the fan speed at 45% and no artifact, however the fan was controlled by the card's BIOS so if the card saw it was getting too warm (98 Celsius and higher, thermal shutdown is apparently 108 Celsius), it would crank up the fan speed regardless of what the state of my PC was or what was loaded up, until I overrode the ran with RIVATuner until Core temp started to drop. Also, I've seen a GeForce 7300LE (OEM card) break the 110 Celsius barrier while running OCCT on it. Needless to say, that card I doubt will last long due to the temps it runs at. Next time I see the PC that card is in, I'll modify the cooling on that.

My GTX280s run at 48-55 Celcius under full load in my box with a little overclocking (which is usually when running Crysis or Folding@Home). Never given me issues. I applied some IC Diamond to the cards when I got them as the nVidia thermal paste wasn't the greatest on there.

nVidia Geforce 8600GTS

 

[Livewire]

nVidia Geforce 8600GTS

8800gt in mine, runs 50-60c at idle, 80c at max. That's not what's causing your BSOD's, unless you mean the heat is causing problems for other hardware in the box.

I mentioned it on IRC but I dont know if you saw it or not; go snag Rivatuner ( http://www.guru3d.com/index.php?page=rivatuner ) and override the fan settings so it's not automatic - crank it as high as you can until the whine gets on your nerves (although on mine, I find it easy to tolerate even at 100% fan speed).

The problem is nvidia's cards apparently can stand some high temps, so the fans don't start running until it gets up there. It's easier to just override the fan and speed it up so it doesn't get there Works wonders here



Fair warning, Rivatuner could also do some over/under clocking if you want it to; I'd just stick with the fan speed.

 

[xPlozion]

what kind of bsod are you getting (irq_not_less_or_equal_to)?

Locking up would be a processor (if oc'ing), bsod could be ram/vidcard/os, depending on error

what vid drivers are u running?

 

[galaxyAbstractor]

what kind of bsod are you getting (irq_not_less_or_equal_to)?

Locking up would be a processor (if oc'ing), bsod could be ram/vidcard/os, depending on error

what vid drivers are u running?

yes I get irq_not_less_or_equal.

Driver version is 191.07

 

[Sharky]

191.56 is out

 

[galaxyAbstractor]

191.56 is out

no I am using the latest version, according to nvidia

http://www.nvidia.com/object/win7_winvista_64bit_191.07_whql.html

 

[Smith6612]

Vigge, irq_not_less_or_equal is usually an indication of Memory, power or driver issues. It's a very vague error message. What you should try doing is downloading a program called OCCT and run the Power Supply test and see if it BSODs. OCCT uses the CPU-Z engine so it will read processor and video card temps.

FYI: OCCT will stress out your box, so monitor your box as you run that test, as there is no safety feature for that test. The CPU + GPU test does have safety shut off on it, but it isn't as through.

While you're at it run Memtest on your box. May I also ask, have you ever seen corrupted graphics or artifacting while playing games on your PC? If it's artifacting, there might be a problem with the video memory or power supply, in which case it's best to replace the card first. If your Power Supply is an el-cheapo kind that gives irregular power to the card, that will damage the VRAM.

 

[galaxyAbstractor]

Vigge, irq_not_less_or_equal is usually an indication of Memory, power or driver issues. It's a very vague error message. What you should try doing is downloading a program called OCCT and run the Power Supply test and see if it BSODs. OCCT uses the CPU-Z engine so it will read processor and video card temps.

FYI: OCCT will stress out your box, so monitor your box as you run that test, as there is no safety feature for that test. The CPU + GPU test does have safety shut off on it, but it isn't as through.

While you're at it run Memtest on your box. May I also ask, have you ever seen corrupted graphics or artifacting while playing games on your PC? If it's artifacting, there might be a problem with the video memory or power supply, in which case it's best to replace the card first. If your Power Supply is an el-cheapo kind that gives irregular power to the card, that will damage the VRAM.

I already ran memtest. At pass 7 I got enough of not being able to use my computer. By then there was around 1600 errors.

I never seen any graphics issues in games.

 

[Smith6612]

I already ran memtest. At pass 7 I got enough of not being able to use my computer. By then there was around 1600 errors.

I never seen any graphics issues in games.

A memtest should not have errors. It's a RAM or power supply issue if you're getting errors in that case then. The GPU issue of the temps disappearing may just be a fluke, as nVidia's chips are made to stand up against high temperatures.

Does HWMonitor report Power Supply Voltages? If so can you post up a screenshot?

 

[galaxyAbstractor]

 

[Smith6612]

Those voltages look to be in spec. Run OCCT's Power Supply test and post up the voltages from SpeedFan again.