Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

One of our workstations has begun to freeze or reset under some circunstances. We can reproduce this in both Linux and Windows so it's most likely a hardware issue.

The freezes most usually happen when using a compiler (C++, if you're curious). We believe this points to either faulty RAM or CPU overheating.

We cleaned CPU fans and replaced thermal paste. Temperature is now in the 30ºC range when idle and it does pick up when compiling (the last reading before the freeze is in the 40-50 range), but it also picks up under different circumstances which don't cause crashes. To put the temperatures in context, the processor is an Athlon XP 3500+ (which is inherently hot, I hear) and the motherboard is a K8N Neo4-F.

We tested different RAM chips in different banks and it seems to make no difference. We're currently doing an exhaustive memtest86 pass.

Any other likely causes for the freezes? At this point I fear the motherboard or the processor may be damaged. There were no hardware changes lately, but the power source did burn after a few automatic shutdowns due to heat, probably, and was replaced (the new one is more powerful so I don't think it may be the cause)

share|improve this question
1  
Could you specify the motherboard model? (sometimes there are known issues with specific brands/models and maybe someone had a similar problem and will be able to help). It'd also be a good idea to run the sensors command on Linux and post the output that you get. –  mfriedman Aug 21 '09 at 17:14
    
Added - thanks for the suggestion. sensors only reports the CPU temp, the values I cited above. –  ggambett Aug 21 '09 at 17:22
    
May be a dumb question, but have you been using different compilers for your tests? –  Psycho Bob Aug 21 '09 at 17:32
    
I'm not familiar with that MSI board in particular. CPU temp is the only value that you see when running sensors? After memtest finishes, see if you can check motherboard/internal temperature through the BIOS setup. –  mfriedman Aug 21 '09 at 17:33
    
memtest86 found nothing wrong. BIOS says system temp is at 34º, cpu temp at 42º. System fan speed around 4600 RPM, CPU fan speed 0 RPM (the CPU fan is definitely running, though) Yes, sensors doesn't show any other temperature values. It does show a couple of lines identifying the chipset. –  ggambett Aug 21 '09 at 17:45

2 Answers 2

up vote 2 down vote accepted

Here are some things that you could verify:

  • This probably isn't a software issue, since you've mentioned that it happens both on Linux and Windows. However, verify the system logs on Linux to see if there's anything unusual that gets logged when this problem happens.
  • Verify that you don't have any loose expansion cards or cables.
  • If your BIOS has a setting for AMD Cool'n'Quiet, try disabling it and run the compiler again.
  • Check the BIOS and see if the CPU voltage is set to auto (or set too high), and adjust the CPU voltage to the correct voltage in the BIOS. Here you can find the voltage settings for AMD CPUs.
share|improve this answer
    
Lowering the CPU voltage did the trick - temperature is lower so no more crashes. Thanks! –  ggambett Aug 22 '09 at 17:51
    
You're welcome! I'm glad to hear that my answer was helpful for solving your problem. –  mfriedman Aug 22 '09 at 18:29

Does it take longer to freeze/reboot when compiling if the system is warm / has crashed/rebooted before than if the system is started from cold?

Perhaps there is a minute deficiency in the motherboard components that is starting to appear when the system gets hot?

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.