Overheating GX card?

Zaij

Vault Senior Citizen
Orderite
Lately, my computer has been crashing when I play any game made within the last 4 years. This has only started recently and it's kind of annoying.

I'm assuming that it's the graphics card overheating or something because it only happens while playing games, and if I leave a game when the little warning signs appear (like strange colours in the game, etc) and then wait for 10 minutes before joining it again I can play longer than if I ran it to death.

I've got a radeon 9550 with the latest drivers.

Can anyone help me?
 
I've had a similar thing happen to me after most of my old computer got fried. I thought that my Gfx card wasn't damaged but it turned out it was. (I believe that the ram suffered the damage). I ended up replacing it. Don't know if that helps.

Anyways, what you could do is try cleaning the cooler if you have an active one or getting an active one if you have a passive one. But it seems to me that your card is simply giving up the ghost...
 
Specialist said:
Heat... Sink...?
the aluminum (or copper) thingy on your gfx card that i supposed to cool the chip?

CBR2171H.jpg

(the silvery thingy? although it can come in many colors ofc)
 
Oh, right.

Yeah, it's not dusty or anything and it is there. I think DDD is right and it's just carked it, but I've only had this computer for a year or two.
 
well not every component reaches their expected MTBF. there is luck involved too.

i remember throwing an old (discarted) 386 in a case that mustve weighed a ton out the first story window out onto the lawn and we plugged it back in just for laughs.
no laughs, no puffy smoke cloud, it just worked... :)

some stuff seems indestructable, while on the other hand just looking at something seems able to break it.


anyhow, it's important you correctly diagnose the problem. i'd be stupid to buy the wrong part, aye?
 
It could also be power draw on your power supply, or overheating in the power supply as well, resulting in lower overall system wattage. Then components stop receiving juice, stop working, then the programs and likely system in turn crash as they wait for the device to respond. Depending upon where in the system they are, they may lose power at different seeming "priorities", making it seem like a fault in another location.

What wattage is your P/S rated at now?
 
Now that I think about it, it all started when I installed Cossacks 2 (and hence, starforce).
 
Commie Mutant Troubleshooters are crashing the Computer!

Hrmmm ... I think I've got pretty much the same problem.
Well ... it sounds like the same problem.

System works fine using "plain" Windows applications, but crashes/reboots alot when playing recent games. Got it (new) around late 2003, and I'm not sure if it ever did work 100% properly.

It's been crashing since the beginning, but at that time the only 3d game played on it was Morrowind. (Boo! Hiss! I know, but it came free with the Gfx card.) And it's not really known for being bug-free. So I'm not sure if it was Morrowind or the system. It was playable, sometimes lasting an hour or two before a crash, often less. I think it got worse over time, though. My brother wanted to play C&C:Generals, and that was almost unplayable. He also wanted to play some Lord of the Ring game, and that was completely unplayable (max. of 5 minute in-game time before a system reboot). Troika's games aren't much good, either. Just about any newish game ...

I tried Half-Life and Deus Ex and those worked almost flawlessly. Over time I gave up on newer games altogether, and considering that the computer is mostly used as a family media/internet PC, it didn't matter much.

But now I'm fed up with it. Computer's filled up with lots crappy soft-, spy- and malware beyond my ability to remove properly. (As it is wont to do if you let young irresponsible sibling mess around with a computer too long.) A clean install is long overdue.
Backed up most of the files already, and it's going down in the next few days. But while I'm at it, I'd like to fix the "can't play new games" problem too.

So... That's the end of this boring narrative. Now rejoice in these utterly exciting computer stats.

-Windows XP Professional SP2
-ASUS A7V8X-X Motherboard ACPI BIOS revision 1006
-AMD Athlon MP 2.0 Ghz processor. (Could have sworn it used to be an Athlon XP 2.4Ghz, but I guess I was wrong.)
-512 MB of RAM
-ASUS V9560 Video Suite (Nvidia Geforce FX5600 128MB)

And then there's a network card, a 80GB Hard Disk, a DVD-ROM drive, a CD-RW drive, oh and a 3½ inch floppy. Sound is on the motherboard.

Oh, and since Rosh asked for the Power Suppy:
Code:
350 Watt

+3.3V	+5V	+12V	-5V	 -12V	 +5Vsb
 14A	 30A	 12A	 0.5A	0.8A	 2A
  \______/
      |
  max 200 Watt

Previous attempts at fixing the bugger include (but not limited to):

-Cleaning up mal- and spyware (this was earlier, now it's hopeless)
-Thorough virus scans
-Running as few as possible background programs.
-Latest Video drivers
-Latest DirectX (and some tests)
-Replacing Gfx card with old Geforce2 400MX/MX. (Ran once - slow. Game crashed again, computer decided it didn't want to boot up anymore after that until the regular card got back in. Geforce2 didn't break - still functional in other PC.)
-Cleaning out dust from the case and fans
-Running MEMtest86
-Running SpeedFan

MEMtest86 gave the OK on my RAM.
I'm not really sure what to do with SpeedFan. Not sure what temperature corresponds to what, and some of the readings are wonky. (might point at a problem!) So here's a screenshot.

speedfan21Sept2006.png

Taken during regular/non-intensive computer use.

...

Blah. Now what do I do?

-Gerko, calling troubleshooters.
 
if temp 1 corresponds to CPU & that screen was taken during idle, then you might have a heat issue. check the bios for overheating settings, it might've been selected to go into emergency shutdown (making it look like a crash to you).

try running a program called SuperPi for an extended period of time. basically it stresses your cpu pretty hard. then look at how hot your cpu gets. (and if it possibly crashes)

of course, it's possible that the speedfan readings are no reliable. if you doubt the readings, try to see if AsusProbe works on your motherboard.

anyhow, the first possibility i'd check would be the CPU: is the heatsink clean & is the thermal paste still making enough contact. (if SuperPi doesnt cause troubles than you can probably scratch that issue)

your -12V and -5V is kinda off, but it's unlikely to cause this much trouble.
 
Rumours of the Computer crashing are Treason

Running SuperPi for 11 minutes got Temp1 up to around 59°. After which it was done, and it got back down to to 54°. No voltage fluctuations registered. (more than 0.05 anyway)

I'll run a longer test later.

BIOS ... right. I did look around there, but there's a password, which I didn't get around to removing until yesterday. Did this by removing the battery for a few seconds. Guess that also set everything back to Default, and I'm not sure what it used to be on. (Stupid, stupid. Should have run some other diagnostics and kept the results, but noo ... )

Emergency shutdown options are there, but seemingly only based on CPU, and then at 85 or 95 degrees, 95 being the default. If SuperPi is to be one of the most intensive things you can make it do, I don't think a game could suddenly heat it up by 40 degrees more in half the timespan. As I said, I'll run a longer SuperPi test later to see how far up it goes, but I don't think that's it.

Setting the CPU to 2400 Mhz makes it refuse to boot up. Oops. Setting it back to 2000 makes it happy again. Lower also works, but meh.

Tried the SmartDocter program, which kinda looks like a GPU-only version of SpeedFan, but with more understandable (to me) readings and settings. Setting GPU Fan speed to High takes it up to ~6000 RPM, makes an awful lot of noise and knocks the temperature down about 1 degree, 2 degrees tops. Not sure if it's more effective at higher temperatures, but it's staying at Medium (~4000 RPM) for now.

I'll try out the AsusProbe thing.
[EDIT #1]
Tried it. It doesn't register the Power Fan running, but then again neither does SpeedFan. CPU temperature registers about 12 degrees higher (if SpeedFan's Temp1 is indeed the CPU). Not sure what to think of that. <s>BIOS seems to agree with SpeedFan, though.</s>
[/EDIT #1]
[EDIT #2]
Scratch that. BIOS does seem to agree with Asus Probe now. Set the Emergency shutdown to 85 degrees and I'll go run SuperPi a bit longer[/EDIT #2]

[EDIT #3]
Okay, did that ...
superpitemperature22Sept2006.png

Temperature climbs up and flats out at either 61 or 76 degrees, depending on which temperature reading I should trust. Got there in about 10-15 minutes. Stayed there until SuperPi was done, then dropped back to about 54 or 65.
Not ideal temperatures, but still well out of range of OMG CRASHZ0RS!!!1111oneoneone
[/EDIT #3]

Thanks anyway!
Anything else ?

-Gerko, not a Commie Mutant Traitor.
 
i'd put more stock in the AsusProbe than in the Speedfan readouts. Speedfan can take wrong readings due to using the wrong bus settings.

Asusprobe on the other hand was developed by Asus which is the manufacturer of your board, so chances are, they should be about equal to what you'll get from BIOS readouts. (although BIOS will always show lower, since rebooting & going into BIOS gives the cpu time to cool off a bit)
anyhow, AsusProbe is far from perfect, but usually about right.

try running one of the longer SuperPi calcs with AsusProbe on. see if there is indeed such a high boost in temp.
 
Whoops. Cross-posted.
I just did. Ran a 16M digits calculation, took 24 minutes. Had both SpeedFan and AsusProbe running, temp got up to 61 or 76 degrees centigrade respectively and stayed there until the end of the calculation. Dropped down pretty quickly after that.

Hrmmmm ...

-Gerko, kind of stuck
 
Back
Top