Computer restarting

[Brac20]

I've been having an issue with my comp for a while. When under loads (usually gaming) the computer will do a full restart, no blue screen or warning.

At first I thought it might be the GPUs overheating (although I thought this would result in a blue screen) and downloaded Rivatuner. I used it to turn the fans up while gaming (although for some reason it wouldn't let me set an automated threshold limit to change the speeds automatically). This seemed to help the issue although it didn't completely fix it.

I thought maybe it was a PSU issue, but on checking the model (OCZ780MXS) it seems that it should handle the duel gfx cards (2 geforce GTX 260 cards) with no problem.

I've downloaded coretemp to check the CPU temp now (Intel Core 2 Quad Q6600). While writing this it is showing;

Core1 - 44
Core2 - 38
Core3 - 40
Core4 - 44

Are these usual for just using web browsing?

The case has 3 fans as well as the CPU fan and GPU fans.

After the last restart I checked this core temps logged prior to the restart. 60, 52, 54, 57. This was while playing Starcraft 2. They seem alright, the TjMax is 100 for the processor.

Any help would be greatly appreciated.

[smack323]

Sounds like a RAM issue.- Try a memory test and see if it resets.

[Major_A]

I agree with the RAM assessment. At one time I thought I had my computer completely stable. Until I fired up Company of Heroes and it would crash randomly. Turns out the RAM was causing the restart.

Things to try:
1) Run said memory test. If you're running Vista or 7 there is a built-in test at the Safe Mode selection screen. If the RAM fails it doesn't necessarily mean that it's bad. You might just need to do one of the following below.
2) Loosen the timings.
3) Run a divider.
4) Add more voltage to the RAM.

Download CPU-Z and post a screenshot of the Memory and SPD tabs. CPU-Z v1.55 Setup
Examples:

[Brac20]

Did a memory test and it didn't pick up any errors.

Downloaded CPU-Z

[Major_A]

RAM seems to be running with the correct profile and timings. How much voltage are you supplying to the RAM in the BIOS?

My brother in law's machine kept rebooting during gaming. Turns out it was a BIOS issue and an update fixed the problem. Look for a BIOS update apply it if there is one available.

[Brac20]

In the BIOS it said:

Vcore 1.266V
DDR2 1.8V 1.952V
+3.3V 3.264V
+12V 12.048V

All the voltage controls are set to automatic, and seem to be set to "normal".

I've not looked into BIOS updates yet, is it a simple process?

There are two versions of the BIOS on the website. A first release and then another one. They are both dated before we actually got the computer. Would the second version have been applied to motherboards made after that date? How can I check to see which version I have?

Edit: Just took the GFX cards and memory out to make sure they were all seated properly (which I think they were). The label on the RAM says something about 2.1V. don't know if this is significant.

[Major_A]

What computer/motherboard do you have? You can use HWiNFO32 to see what BIOS you have installed.
http://www.hwinfo.com/beta/hw32_357_894.zip



Flashing the BIOS is really dependent on what board you have. Older boards you needed to flash with a floppy and command line switches. Most newer boards can flash the BIOS with a utility through Windows.

[Brac20]

The motherboard is a Gigabyte Technology, GA-N650SLI-DS4L. It's got the Nvidia nForce 650i SLI SPP chipset.

According to the program you linked it is used BIOS F2. This isn't actually listed on their website, they show an F1 and an F3. Although the date of F2 confirms it is an earlier version than F3.

The BIOS manufacturer being "Award Modular BIOS v6.00PG"

[smack323]

Edit: Just took the GFX cards and memory out to make sure they were all seated properly (which I think they were). The label on the RAM says something about 2.1V. don't know if this is significant.

I noticed your CPU-z shows your memory at 1.8v- and you said the label on the RAM shows 2.1v, In your BIOS up the voltage to the RAM to 2.1V- that would be my first step before flashing the BIOS.

[Brac20]

I'll try that out tonight. I did read somewhere earlier that this motherboard doesn't cope well with RAM voltages above 1.8V. I guess we'll see when try it out tonight.

[Major_A]

Go ahead and update your BIOS to F3 with Gigabyte's @BIOS utility while you're at it.

http://www.gigabyte.com/MicroSite/121/tech_a_bios.htm

To minimize a chance of something going wrong download the BIOS first and update it with that file. Do not choose to update through the internet. You could get a bad packet and potentially screw the board up.

*EDIT*
Is all the RAM in your computer the same brand/model? Looking at CPU-Z it shows that it can run at 4-4-4 and I would assume that is where you would supply the 2.1V. Otherwise just raise it to 2.0V and let it detect the clocks.

[Brac20]

Yeah all 4 RAM sticks are exactly the same

[kenc51]

A simple test would be to run the PC with 1 stick of RAM and see if it reboots.
If it does, Run Memtes86+ with that config, if passes; its something else than RAM (proly PSU)
If it gives errors in memtest, then try another slot on the board and or stick of RAM.

To eliminate software as the issue, you can try booting from a Linux (live) CD
Something link Ubuntu would do
If you dont want to waste a blank CD, then use a USB key and this program http://unetbootin.sourceforge.net/


Ken

[Brac20]

I've just looked at the system temps in a bit more detail. The majority of sensors are showing between 30-40. One sensor is between 55-60 degrees, its for the nVidia SPP. Any chance that it could be part of the problem? I just ask because it appears to be running alot hotter than the other parts.

This problem has persisted from Windows XP to Windows 7. Is it still worth booting in Linux?

[Brac20]

Update: Tried turning the RAM voltage up. Put it up to 2.0V from 1.8V initially as any higher than that was highlighted in orange (which I took as a warning). The computer kept failing to startup after that, took 4-5 times before I could get as far as the BIOS screen to turn the voltage back down. I did read somewhere that this mobo doesn't like having RAM voltages over 1.975V or something.

[Brac20]

Someone on another forum suggested that the mobo might not be able to handle the voltage of 4 ram sticks. Currently running with just 2 sticks and has lasted alot longer without crashing. Will try the the other two sticks out tomorrow and maybe try 3. Feeling hopeful that I might finally gave a solution though. Will keep you posted and thanks for all the help up to this point, I feel I know alot more about the inner works of the computer now

[Major_A]

Once you raised the voltage to the RAM it probably tried to run the the 4-4-4 profile and it didn't work. Manually set the speed to your RAM, just use what it's defaulting to at 1.8V, but raise the voltage. It isn't going to hurt anything, just another option to try is all.