Hi 39er

Im also running suse 10.1 and boinc but my system is running stable with no problems. If you know the exact time and date of the freeze or close enough then go to your k menu select system then yast control centre then miscellaneous and in the right hand panel youll find view system log. Select this and then the /var/log/messages tab and it should give you an error message for the failure or freeze. hope this helps you.

Keep us posted

Slapshot

G'day 39er

Have you checked for overheating? Check the heatsink fan for dust buildup and check that the heatsink isn't loose. (I'm presuming that this is a relatively new problem).

G'day Trog Dog Chruncher,

Thanks for the hint. It makes sense since the the PC does nothing else but crunch numbers and the CPU load is nearly 100%.
The PC here is an older IBM Netvista. When I opened it up I was actually surprised by the small size of the heatsink/fan assembly.

I shall certainly investigate your suggestion further. Thanks mate.

cheers,
-39er

Hi 39er,

As Trog Dog suggested it could be temperature related, my wifes windows pc only crunches fine with no case sides on and the stock pentium fan is clean and running fine. My cpu has a layer of arctic silver thermal paste and a little enclosed evaporative gas cooler assembly on it to keep running continuously. I have also seen this before whilst browsing the suse forums and have found the thread I was looking for after some grey matter searching.

The link is here
http://www.linuxquestions.org/questions/showthread.php?t=452466

Also ther are a couple of things here to try about half way down is the cpu frequency scaling and the powersav demon. Give it a try and also suse 10.0 is doing same for some so here is a link to a thread for that but it could be mem timing power supply or cpu overheating. This could be as you are driving it harder than anyone predicted with a stock cooler when they designed it. They didn't think it would run 24/7 with 100 % load.

Here's the link
http://www.linuxquestions.org/questions/showthread.php?t=412884

Also mate there was a problem that I had myself when I installed it and I had to run a kernel specifically made by Nvidia for my 3d accelerated card an FX 5600. You dont by any chance have one of those do you.

Also on the subject of temps there is a package you can install called GKRELLM which monitors pretty much everything from processor percentage to your e-mail and makes a Penguin dance if you have any LOL, But it does monitor cpu usage temp and hard disk activity.

Dont give up and we'll get there mate
Keep us posted with any news

Slapshot

39er,

Glad to hear you're getting it sorted out. Have you checked the CPU core voltage? Some BIOS-motherboard combinations allow you to play with the core voltage. Higher core voltage makes the CPU run hotter. If the core voltage is higher than called for in the specs for that CPU then you should be able to reduce the core voltage and reduce the temperature. I was just thinking that maybe you or perhaps a previous owner tried overclocking the CPU and raised the core voltage, a common practice in overclocking. They may have found the overclock did not work and may have forgot to set the voltage back to specification. If you go into the BIOS and find it does not allow tinkering with core voltage then obviously that's not the problem. If it does allow voltage adjustments then you might want to look up the voltage spec at the CPU manufacturer's website. The specs will be in a document called a Data Sheet. The trick is to make sure you've got the exact model number for your CPU when you look for the datasheet.