Processor Utilisation zero on AMD64

[jrs]

Installed AMD64 on Dell Poweredge R300 with dual corp Intel Xeon E3113 processor.

Top utitlity reports system load but 0% utilisation for both cores and all processes.

Does Processor utilisation and process cpu utilisation require AMD processor?

[jggimi]

No, 64-bit Intel CPUs run the same instruction set, though there are differences in implementation of PAE NX on some Intel processors -- see the opening paragraph of www.openbsd.org/amd64.html for details.

For help with diagnosing your problem, please post your dmesg and a sample of the incorrect top output.

[jggimi]

Ah, your dmesg is here: http://www.daemonforums.org/showthre...3867#post23867

So just post the top output.

[jrs]

load averages: 0.06, 0.08, 0.08 13:49:48
37 processes: 36 idle, 1 on processor
CPU0 states: 0.0% user, 0.0% nice, 0.0% system, 0.0% interrupt, 100% idle
CPU1 states: 0.0% user, 0.0% nice, 0.0% system, 0.0% interrupt, 100% idle
Memory: Real: 180M/409M act/tot Free: 1570M Swap: 0K/50G used/tot

PID USERNAME PRI NICE SIZE RES STATE WAIT TIME CPU COMMAND
22734 root 2 0 47M 43M sleep/1 select 0:00 0.00% perl
18568 root 2 0 3428K 3056K sleep/0 select 0:00 0.00% sshd
3226 root 2 0 10M 5036K sleep/1 select 0:00 0.00% perl
27811 root 18 0 608K 528K sleep/0 pause 0:00 0.00% ksh
24146 root 10 0 336K 708K sleep/0 nanosle 0:00 0.00% svscan
7957 root 2 0 109M 1416K idle poll 0:00 0.00% clamd
30244 qmails 2 0 448K 836K idle select 0:00 0.00% qmail-send
16765 root 2 0 47M 5372K sleep/1 select 0:00 0.00% perl
12464 root 2 0 47M 5404K sleep/0 select 0:00 0.00% perl
6315 _syslogd 2 0 484K 780K sleep/0 poll 0:00 0.00% syslogd
2108 root 2 0 468K 728K idle select 0:00 0.00% qmail-lspawn
31978 _clamav 2 0 684K 1352K idle poll 0:00 0.00% freshclam
10420 qmaill -6 0 312K 744K idle piperd 0:00 0.00% splogger
28185 root 2 0 564K 892K idle select 0:00 0.00% cron
15068 qmaild 2 0 444K 824K idle netcon 0:00 0.00% tcpserver
21983 qmailq -6 0 388K 728K idle piperd 0:00 0.00% qmail-clean
2967 root 2 0 376K 572K idle poll 0:00 0.00% supervise
22934 root 2 0 288K 580K idle poll 0:00 0.00% supervise

[jrs]

Top now appears to be working properly. Apart from a reboot , nothing relevant was changed.

[jggimi]

Odd.

The load values looks reasonable, but there were no active/running processes other than top. If it happens again, just run a looping shell script, or something else that should busy out one or both CPUs, such as sysutils/stress.

[bsdplus]

not quite sure about this, but my understanding is that this is caused by TOP sampling interval. Top get CPU utilization statistics every second, but on a idle server, most processes finish their operation within, say, 200 mili seconds, so Top just reports 0 utilization.