kernel confused about processor speed?

[erickfour]

hello--

i've been running gentoo for almost a year now on my 3.2Ghz box built on an intel desktop board. it's worked great.

however, about a week ago it started slowing down considerably, perhaps a three fold decrease in speed. so i thought i would upgrade to 2.6.10 from 2.6.1 with the logic that perhaps recompiling the kernel would do the trick. i did, no problem, now i'm running 2.6.10 but same problem.

then i thought that i would run some benchmarks. i emerged lmbench and ran mhz, with the following bizarre results:

[~]$ mhz
3165 MHz, 0.3160 nanosec clock
[~]$ mhz
6384 MHz, 0.1566 nanosec clock
[~]$ mhz
6387 MHz, 0.1566 nanosec clock

however,

[~]$ uname -a
Linux wesley 2.6.10-gentoo-r6 #1 SMP Tue Feb 1 15:21:09 EST 2005 i686 Intel(R) Pentium(R) 4 CPU 3.20GHz GenuineIntel GNU/Linux

mhz seems to be spitting out bizarre things, despite the fact that it works fine on the laptop... can someone help me out here?

thanks in advance,

e

[Boohbah]

It sounds like your cpu is overheating. Better check the temperature!
EDIT:
On second thought, maybe the Pentium 4's hyperthreading looks like two separate processors to 'mhz', though i'm not sure, just guessing.
_________________
Never try to explain computers to a layman. It's easier to explain sex to a virgin.
-- Robert Heinlein

(Note, however, that virgins tend to know a lot about computers.)

[n3mo]

Double xeon with HT enabled.

[erickfour]

boobah---

i don't think that's the problem here. i shut down the computer overnight and it did the same thing in the morning. furthermore, here's mbmon.

[~]$ mbmon

Temp.= 33.8, 42.2, 32.8; Rot.= 0, 1163, 0
Vcore = 1.48, 1.49; Volt. = 3.26, 5.04, 12.00, 0.00, 0.00

34 is certainly within normal range.

thanks for the quick reply tho!

e

[JuNix]

To find your current CPU clock rate :-

[erickfour]

bizarre... i built my own machine so i know it's single processor (ha ha) but here is /proc/cpuinfo

processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3192.757
cache size : 512 KB
physical id : 0
siblings : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
bogomips : 6307.84

processor : 1
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.20GHz
stepping : 9
cpu MHz : 3192.757
cache size : 512 KB
physical id : 0
siblings : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
bogomips : 6373.37

hmm...

did i do something funny in make.conf?

CFLAGS="-O3 -march=pentium4 -funroll-loops -fprefetch-loop-arrays -pipe"


also, by the way, i don't have frequency scaling turned on:

[/usr/src/linux-2.6.10-gentoo-r6]$ grep CONFIG_CPU_FREQ .config
# CONFIG_CPU_FREQ is not set

and locate governor only finds some documentation .txt files.

i also don't have anything suprising going on with top. i didn't see anything exceptional in /var/log/messages, but i don't read this too much so perhaps i don't know normal.

thanks, junix

[erickfour]

hi--

i thought i would add that the root filesystem has gotten corrupted twice today, and i imagine that the source of both problems must be the same.

before it got mounted read-only but for some reason i thought it was a separate problem.

so (computer thinking it has two cpu's) + (root filesystem corruption) = ????

thanks,

e

[JuNix]

You have a 3.2GHz Pentium 4, with Hyperthreading. It should show up as two CPU's and that's normal.

As for the root fs corruption... dunno! Boot from a recovery CD and fsck -y /dev/hdX whatever the device name you are using for root fs.

Soaktest your box overnight using Memtest86 maybe?

There may be a kernel config problem that isn't obvious. I've head of ppl having problems with MTRR's and APIC on dual core CPU's. Try a couple of boot's without them. I think you can pass a "noapic" line into grub or lilo at boot time to disable.

There's lots of stuff about this at the SMP Linux Howto

Also http://kerneltrap.org/node/3683

And, if you do an "uptime", what is the load average reported?

[erickfour]

well, i ran memtest86 and it crashed, even when i mix-and-matched the DIMM's.

so i'm guessing it was the motherboard. i'm going to get it checked out tomorrow.

e

[erickfour]

OK--

so i replaced the motherboard, and so memtest86 no longer crashes.

but the performance is still mighty sluggish!

i tried turning off hyperthreading but no improvement. i have also tried various combinations of APIC and MTRR, still no luck. here is some uptime-related output...


top - 16:00:29 up 1:09, 5 users, load average: 0.69, 0.60, 0.50
Tasks: 64 total, 3 running, 61 sleeping, 0 stopped, 0 zombie
top - 16:01:47 up 1:10, 5 users, load average: 0.82, 0.68, 0.54
Tasks: 56 total, 1 running, 55 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.3% us, 0.3% sy, 0.0% ni, 99.3% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 0.3% us, 0.3% sy, 0.0% ni, 99.3% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 904680k total, 610012k used, 294668k free, 77568k buffers
Swap: 499960k total, 0k used, 499960k free, 286708k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
7500 root 16 0 158m 12m 2516 S 1.0 1.4 0:12.70 X
8788 matsen 16 0 5372 2512 1996 S 0.7 0.3 0:00.07 xterm
1 root 16 0 1344 488 432 S 0.0 0.1 0:00.23 init
2 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0
4 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/1
5 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/1
6 root 5 -10 0 0 0 S 0.0 0.0 0:00.01 events/0

[~]$ uptime
16:02:33 up 1:11, 5 users, load average: 0.38, 0.58, 0.51
[~]$ w
16:02:37 up 1:11, 5 users, load average: 0.35, 0.57, 0.51
USER TTY LOGIN@ IDLE JCPU PCPU WHAT
matsen vc/1 14:53 1:09m 3.16s 0.00s /bin/sh /bin/sx
matsen pts/0 14:53 1:09m 0.08s 0.08s bash
matsen pts/1 14:57 3:00 0.18s 0.33s xterm +sb -fn 9x15 -geometry 80
matsen pts/2 16:00 0.00s 0.09s 0.00s w
matsen pts/3 15:33 18:08 0.09s 0.09s bash


nothing terribly suprising there...

any more suggestions?

thanks,

erick

[JuNix]

Well.... I wonder how slower your box is running that a known working 3.2Ghz Pentium. you could try the following to find out:

[erickfour]

ok-- it's

[JuNix]

I don't suspect your CPU or memory any more.

Take a look at the My System is Slower than it Should Be thread. It contains useful stuff.

[erickfour]

well, i repartitioned my second HD and installed the system on that, and now it's back working perfectly. i don't really know why, which is a bit scary, but i'm happy for the time being.

thanks, everyone.