Sudden reboot when using 3d

[colossus73]

Hi,

this my first message. When I play with Nexuiz or World of Padman the PC suddenly reboots after some minutes. This happens ONLY when I use 3d. It DOESN'T happen when I watch movies in fullscreen. I have an ATI HD 4830 with catalist 9.2. I checked the temperature of the card before playing with 3d games and it's 32 celsius degree. This is my xorg:

Section "ServerLayout"
	Identifier     "Default Layout"
	Screen      0  "aticonfig-Screen[0]-0" 0 0
EndSection

Section "Files"
EndSection

Section "Module"
	Load  "glx"
	Load  "extmod"
	Load  "xtrap"
	Load  "record"
	Load  "GLcore"
	Load  "DRI"
	Load  "type1"
	Load  "freetype"
EndSection

Section "InputDevice"
	Identifier  "Generic Keyboard"
	Driver      "kbd"
	Option	    "XkbRules" "xorg"
	Option	    "XkbModel" "pc105"
	Option	    "XkbLayout" "it"
EndSection

Section "InputDevice"
	Identifier  "Configured Mouse"
	Driver      "mouse"
	Option	    "CorePointer"
	Option	    "Emulate3Buttons" "true"
EndSection

Section "Monitor"
        Identifier      "Generic Monitor"
        Option          "DPMS"
        HorizSync       30-71
        VertRefresh     56-75
EndSection

Section "Device"
	Identifier  "aticonfig-Device[0]-0"
	Driver      "fglrx"
	BusID       "PCI:4:0:0"
EndSection

Section "Screen"
	Identifier "aticonfig-Screen[0]-0"
	Device     "aticonfig-Device[0]-0"
	Monitor    "aticonfig-Monitor[0]-0"
	DefaultDepth     24
	SubSection "Display"
		Viewport   0 0
		Depth     24
	EndSubSection
EndSection

The X version is:
X.Org X Server 1.4.0.90
Release Date: 5 September 2007
X Protocol Version 11, Revision 0

I have nothing relevant in /var/log/messages, just the syslogd saying: Restart, the same in Xorg.0.log.

Do you have any clue?

[rbmorse]

Spontaneous reboots can be caused by many things, including video drivers, but usually indicate a power supply issue.

Power supply capacity diminishes over time as capacitors degrade. Perhaps yours is at a threshold where it can supply adequate current for normal stuff but not when the 3D engine on the GPU really gets going. Another possibility is one of the internal components in the power supply or the motherboard voltage regulators overheats under heavy load.

Finally, if the motherboard is a couple of years old or more, use a bright light and hand held magnifier to examine the capacitors on the motherboard (they look like miniature beer cans) for signs of leakage or internal overpressure. Any brownish or whiteish "goop" or white "feathers" on the caps indicate internal failure. Also, the tops of the caps should be flat to slightly concave...if any of them are bowed out as if there is to much pressure inside...they are defective as well.

If this is a laptop, it could indicate that the battery has an internal problem, especially if it is a couple of years old (they don't last forever).

[colossus73]

Power supply capacity diminishes over time as capacitors degrade. Perhaps yours is at a threshold where it can supply adequate current for normal stuff but not when the 3D engine on the GPU really gets going. Another possibility is one of the internal components in the power supply or the motherboard voltage regulators overheats under heavy load.

The PC power supply outputs 500 W. I had to connect the power connector on the HD4830 card to it to have the video card switching on.

Finally, if the motherboard is a couple of years old or more

It's a Asrock G31M-S. It should be recent enough.

If this is a laptop, it could indicate that the battery has an internal problem, especially if it is a couple of years old (they don't last forever).

No, it's not. Which power supply is adequate to my video card then?

Thanks for replying,
Colossus73

[rbmorse]

500W should be more than enough, but like most other consumer goods, the quality of PC Power supplies is extremely variable.

I don't have any other specific ideas. From the description you posted the problem is almost certainly thermal (time related but not regular, only when system under heavy load). Maybe the video card has a defective component, but that seems unlikely.

[colossus73]

500W should be more than enough, but like most other consumer goods, the quality of PC Power supplies is extremely variable.

I don't have any other specific ideas. From the description you posted the problem is almost certainly thermal (time related but not regular, only when system under heavy load). Maybe the video card has a defective component, but that seems unlikely.

I think the problem lies in the power supply. I even tweaked a bit in the BIOS but the problem persists. When I use 3d after some minutes the PC reboots. I will change the power supply.

Thanks for replying,
colossus73

[colossus73]

500W should be more than enough, but like most other consumer goods, the quality of PC Power supplies is extremely variable.

I don't have any other specific ideas. From the description you posted the problem is almost certainly thermal (time related but not regular, only when system under heavy load). Maybe the video card has a defective component, but that seems unlikely.

It just came to my mind that when I assembled the processor (a Pentium 4 630 HT) I didn't put any thermic paste between it and the heatsink. Could this be the problem? I wouldn't want to buy another power supply and then still have the problem.

[Saist]

It just came to my mind that when I assembled the processor (a Pentium 4 630 HT) I didn't put any thermic paste between it and the heatsink. Could this be the problem? I wouldn't want to buy another power supply and then still have the problem.

simple way to test: get some thermal paste and try it out.

[rbmorse]

It just came to my mind that when I assembled the processor (a Pentium 4 630 HT) I didn't put any thermic paste between it and the heatsink. Could this be the problem? I wouldn't want to buy another power supply and then still have the problem.

If you used the stock Intel heat sink/fan unit, it probably came with a thermal pad already attached (it would have been a gray square or patch on the surface where the heat sink contacts the CPU). I suspect if there was no thermal pad/paste at all you would not be able to use the machine for more than a few minutes under any circumstances. Most Intel-chipset machines have a high temperature alarm and the machine would slow way, way down as it approached the upper thermal limit. A severe CPU overtemp condition should result in a complete shutdown rather than reboot.

If you remove the heat sink for any reason you should always clean the residue of the old thermal compound from both the CPU and heatsink and renew it before reassembling. If you have removed the heat sink since initial assembly and did not renew the thermal compound that might cause the symptoms you see...but I would still expect alarms and shutdown rather than a spontaneous reboot.

Since you're here I'm assuming you run a Linux of some flavor, you can install lm-sensors and a front end like gkrellm or xsensors or something like that to tell you approximately what your CPU temp is at any moment. It's probably easier than taking the machine apart just to see if there really is a thermal pad under the heatsink.

I'd hate for you to replace the power supply and find that does not fix the problem, too, but that remains a possibility. Diagnosing hardware problems when you can't put hands on the machine is very difficult and people are always leaving out critical details. The only way to be sure the power supply is operating properly is to take it to a shop that is equipped to test power supplies under load (most aren't). Those that are will likely charge for the test because the machine you need is quite expensive.

Short of that, the only diagnostic for power supplies is to swap and see if the problem persists.

I get a bit of religious fervor about power supplies and I think that except for cable/connection problems power supplies are at the root of most PC hardware problems. I also think a lot of issues that appear to be software are really transient power supply issues.

If you decide to replace your power supply there are two meaningful questions: How much does it weigh? How long is the warranty coverage? A well-designed power supply using top quality components and heat sinks is going to be heavier than a poor quality unit. Similarly, a quality unit will carry a warranty that exceeds the legal requirements for your market.

Check the reviews that Michael posts on his main page for specific recommendations, but in any case resist the temptation to buy a cheap unit and assume it will be good enough. A good power supply is going to be relatively expensive and there is no way around that.

[bridgman]

The other thing to watch out for is total +12v current. I still see older power supplies on the market which have high wattage but a lot of the power is on the +5v side, which isn't so useful these days. Graphics cards mostly draw from the +12v line, and I think the same goes for the CPU on most motherboards these days.

[colossus73]

Thank you all for your kind suggestions, I put the thermal paste after removing the residues and now it works like a charm! I completed the Unigine Tropic benchmark with antialiasing 8x without any reboot.

Thank you all for your kind replies,