Prime95/mprime Causes Kernel Panic, but Win7 is Stable

Convergence · 2012-06-20 17:23:30

I'm not overclocking. I just thought that I would test my cpu temps when under load (think my cpu is getting noisy, but is fine). So I downloaded mprime, ran the "blend mode" and I got a kernel panic within 2 minutes. (Also, if I was running kde, the USB device notifier menu would quickly popup and disappear. This makes sense as the stdout says something about io errors with my usb hd)

Then, I disabled my unlocked core, reset my bios, and tried again. Same results.

Trying to pinpoint the problem, I ran in small fft's so as not to stress the ram. It ran fine for half an hour before I manually stopped it (that is acceptable to me.)

Then I tried the "in place fft's" test and it froze my computer instantly (keyboard unresponsive, this was worse than the "blend" test). After a few minutes I got a kernel panic.

At this point, I was thinking that something was wrong with my memory controller. (I guess I can't really rule this out) However, I decided to try prime95 under windows. I ran all three tests for half an hour, and got no errors.

(by the way, the temps seemed to max out at about 62c on all tests. This is too high for an AMD chip to run for an extended period of time, but it shouldn't cause a crash)

So, this leaves me confused as to whether or not I have a hardware problem. Perhaps Windows 7 is good at dealing with the io and memory errors and is kind of sweeping the problems under the rug and therefore hiding a real problem, or there is something wrong with mprime or the linux kernel that is causing a kernel panic under windows.

cat /var/log/errors.log | grep -i error | grep -i kernel
Jun 18 21:27:35 myhost kernel: [20092.184696] usb 6-1: device not accepting address 3, error -62
Jun 18 21:49:03 myhost kernel: [ 1058.350193] end_request: I/O error, dev sdd, sector 3907029096
Jun 18 21:49:03 myhost kernel: [ 1058.350241] Buffer I/O error on device sdd, logical block 976757274
Jun 18 21:49:03 myhost kernel: [ 1058.350281] Buffer I/O error on device sdd, logical block 976757275
Jun 18 21:49:03 myhost kernel: [ 1058.350448] Buffer I/O error on device sdd, logical block 2
Jun 18 21:49:03 myhost kernel: [ 1058.350493] Buffer I/O error on device sdd, logical block 3
Jun 18 21:49:03 myhost kernel: [ 1058.350527] Buffer I/O error on device sdd, logical block 2
Jun 18 21:49:03 myhost kernel: [ 1058.350561] Buffer I/O error on device sdd, logical block 3
Jun 18 21:49:03 myhost kernel: [ 1058.350596] Buffer I/O error on device sdd, logical block 2
Jun 18 21:49:03 myhost kernel: [ 1058.350629] Buffer I/O error on device sdd, logical block 3
Jun 18 21:49:03 myhost kernel: [ 1058.350675] Buffer I/O error on device sdd, logical block 0
Jun 18 21:49:03 myhost kernel: [ 1058.350708] Buffer I/O error on device sdd, logical block 1
Jun 18 21:49:06 myhost kernel: [ 1061.654087] usb 6-1: string descriptor 0 read error: -71
Jun 19 12:00:53 myhost kernel: [ 1063.355305] PM: Device 7:0:0:0 failed to suspend async: error -5
Jun 19 12:01:11 myhost kernel: [ 1081.892322] PM: Device 7:0:0:0 failed to suspend async: error -5
Jun 20 06:51:43 myhost kernel: [  819.663121] end_request: I/O error, dev sdd, sector 800
Jun 20 06:51:43 myhost kernel: [  819.663798] Buffer I/O error on device sdd, logical block 200
Jun 20 06:51:43 myhost kernel: [  819.664492] Buffer I/O error on device sdd, logical block 201
Jun 20 06:51:43 myhost kernel: [  819.665361] Buffer I/O error on device sdd, logical block 1024
Jun 20 06:51:43 myhost kernel: [  819.666050] Buffer I/O error on device sdd, logical block 1025
Jun 20 08:12:00 myhost kernel: [ 2169.655770] PM: Device 6:0:0:0 failed to suspend async: error -5

graysky · 2012-06-20 17:43:31

Like your sig. Dethklok rules. To your problem, first minimize variables, in the case run in mode 3 (telinit 3) to rule out a kde problem.
If still unstable, I dunno. You can also try linpack in the AUR and see if that causes issues. It is a more more robust than mprime.

Convergence · 2012-06-20 18:06:45

Maybe I should have mentioned this but I've already tried that. I "rc.d stop kdm" then ran mprime from tty1 (I still use kdm in the rc.conf method). My results were the same.

About Dethklok: Yeah, Dethklok rules. As soon as I heard Murderface say that, I knew it had to be my arch sig.

graysky · 2012-06-20 18:07:59

...linpack results?

Convergence · 2012-06-20 18:11:40

I haven't run it. I am using an AMD processor though, and my understanding is that it's only used for Intel.

Edit:

By the way, I've also run memtest+ for a couple of cycles and got no errors.

Last edited by Convergence (2012-06-20 18:22:10)

graysky · 2012-06-20 18:35:31

Actually I think newer and CPUs can work with linpack.

Convergence · 2012-06-20 19:07:34

cat lin_xeon64.txt
Wed Jun 20 12:02:36 PDT 2012
This binary version of the SMP LINPACK benchmark is optimized for and runs on only genuine Intel processors

It's a no-go I'm afraid.

By the way, I was debating custom compiling my own ck kernel, then I read your sig. I'm going to try it out next week.

Arch Linux

#1 2012-06-20 17:23:30

Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#2 2012-06-20 17:43:31

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#3 2012-06-20 18:06:45

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#4 2012-06-20 18:07:59

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#5 2012-06-20 18:11:40

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#6 2012-06-20 18:35:31

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

#7 2012-06-20 19:07:34

Re: Prime95/mprime Causes Kernel Panic, but Win7 is Stable

Board footer