Ubuntu Server Crashes when running mprime 257

[roemer2201]

Hello everybody,

i got a AMD Athlon 1200+ CPU running Ubuntu Server 8.04 with the Xfce-Desktop (i think this doesn't matter). Everytime i start mprime v25.7 the PC crashes after ~1min running.

Can somebody tell what the reason for this crash is?

[henryzz]

Originally Posted by roemer2201

Hello everybody,

i got a AMD Athlon 1200+ CPU running Ubuntu Server 8.04 with the Xfce-Desktop (i think this doesn't matter). Everytime i start mprime v25.7 the PC crashes after ~1min running.

Can somebody tell what the reason for this crash is?

probably because your computer is unstable
are you overclocking
what are your temperatures
http://www.alcpu.com/CoreTemp/

[mdettweiler]

Originally Posted by henryzz

probably because your computer is unstable
are you overclocking
what are your temperatures
http://www.alcpu.com/CoreTemp/

Core Temp won't work for Linux; lm-sensors is the widely accepted temperature measuring tool of choice for Linux. On a Ubuntu system (desktop or server, don't really matter), just follow these steps:

-Go to a terminal and run the command "sudo apt-get install lm-sensors" (or use your favorite package manager to install the lm-sensors package).

-Then run the command "sudo sensors-detect" in a terminal. It will ask you a bunch of questions; answer Yes to all of them. When it asks whether to update /etc/modules automatically, answer Yes to that too.

-Now, to check the CPU temperature, just run the command "sensors" in a terminal. Depending on your machine, it will output various amounts of sensor data; the main one to pay attention to is the one marked "core temp" or "CPU temperature". (The naming varies from motherboard to motherboard.)

Hope this helps!

[markr]

Bad memory is always a possibility. Did it crash / lockup occasionally before installing mprime? Have you tested your memory? For example, with memtest.

[roemer2201]

I think the memory isn't the point of failure. Usually this PC runs 6 Days a week and it didnt crash, no matter what i do. I also ran memtest86+ 2 Month ago ... no errors.

The temperature.... i didn't overclock this CPU. A sensors applet is allready running, but i didn't care about that while running mprime. I'll do that today! But when i am running a other programm which causes 100% cpu-load (a virtual pc or updates) the coretemp is not higher then 45°C. Anyway, i'll test it on todays evening.
Much thanks for the lm-sensors guide. I hope you dont mind if i copy this into my wiki ?

Maybe i should test the PC with an older version of mprime and whether it crashes, too.

...I willl post news tomorrow. Thanks so far!

PS.: is it possible to run mprime causing just 50% cpu-load and not 100%? Because on this machine mprime slows down all other processes.

[markr]

Originally Posted by roemer2201

PS.: is it possible to run mprime causing just 50% cpu-load and not 100%? Because on this machine mprime slows down all other processes.

From undoc.txt:

If you are running on a laptop or in a hot room and you want to slow the
program down to reduce heat, then you can add this line to prime.txt:
Throttle=n
where n is the percentage of time the program should run. For example,
Throttle=40 means the program will run 40% and be idle 60% of the time.

The program will run for a few seconds then sleep for a few, and average out at n%. Don't forget to close the program before changing prime.txt.

[henryzz]

Originally Posted by roemer2201

PS.: is it possible to run mprime causing just 50% cpu-load and not 100%? Because on this machine mprime slows down all other processes.

try different percentages
75% might not affect the running of the pc but give an extra 50% of help to GIMPS

[roemer2201]

To determine wheter ubuntu itself or the pc it installed a virtual machine and in that machine i run mprime. The same problem occured. So i think the temperature is the problem on this machine, otherwise i would not have crashed because a virtual machine doesn't interfere with the host-operating system.

Thanks for your help!

[Prime95]

The most common failure is memory, not heat. If the small FFT torture test works, then it is even more likely that memory is the problem. Try running with relaxed memory timings or with memory swapped from another computer

[roemer2201]

Well, ths PC had temperature problems month ago, then a bought a new fan and since that the pc ran without problems... now i added an extra ventilator in front of this PC. Lets se whether it crashed anyway or not.

But i'll try the FFT torture test. Thanks for that tip.