High CPU load more than 100%

SUSE Product Topics SLES Configure-Administer
1

Dear all,

top -d 1 and then press “1”
show me high amount (even 200%) of CPU all the time :

  PID USER      PR  NI  VIRT  RES  SHR S   %CPU %MEM    TIME+  COMMAND
 5739 ndbadm    20   0 40.6g  33g 3.8g S    110 53.0   1811:21 hdbindexserver
 5743 ndbadm    20   0 9166m 1.7g 308m S      7  2.6 115:30.74 hdbxsengine

…and why only 2 lines are while I have 2 processors each 8 cores?

Regards
GN

2

Which SLES version and patches? Which RPM version for ‘top’?

You do not show the whole top output, so not sure what’s there, but when I
press ‘1’ I see all of my virtual processors. Perhaps post the output
from the following command:

grep -c ^processor /proc/cpuinfo

–
Good luck.

If you find this post helpful and are logged into the web interface,
show your appreciation and click on the star below…

3

[QUOTE=ab;22591]Which SLES version and patches?

# cat /etc/SuSE-release SUSE Linux Enterprise Server 11 (x86_64) VERSION = 11 PATCHLEVEL = 2
[/QUOTE]

# grep -c ^processor /proc/cpuinfo 32

now my top “1” shows more CPUs but first is taking more than 100% for one process

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 5053 ndbadm 20 0 31.0g 23g 3.8g S 108 37.8 612:46.69 hdbindexserver 6148 gdm 20 0 266m 12m 9336 S 24 0.0 0:24.35 gnome-settings- 5057 ndbadm 20 0 8902m 1.6g 305m S 7 2.6 43:35.52 hdbxsengine 4872 ndbadm 20 0 8201m 2.0g 503m S 6 3.1 37:10.34 hdbnameserver 5784 ndbadm 20 0 1827m 455m 17m S 4 0.7 51:20.88 java 11 root 20 0 0 0 0 S 1 0.0 0:09.05 kworker/0:1 142 root 20 0 0 0 0 S 1 0.0 0:09.42 kworker/8:1 205 root 20 0 0 0 0 S 1 0.0 0:12.88 kworker/11:1 6202 gdm 20 0 243m 18m 12m S 1 0.0 1:08.08 gdm-simple-gree 18020 root 20 0 9052 1304 828 R 1 0.0 0:00.17 top 1 root 20 0 10540 812 680 S 0 0.0 0:02.25 init 2 root 20 0 0 0 0 S 0 0.0 0:00.01 kthreadd 3 root 20 0 0 0 0 S 0 0.0 0:01.81 ksoftirqd/0

4

Hi gniecka,

[QUOTE=gniecka;22612]now my top “1” shows more CPUs but first is taking more than 100% for one process

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 5053 ndbadm 20 0 31.0g 23g 3.8g S 108 37.8 612:46.69 hdbindexserver 6148 gdm 20 0 266m 12m 9336 S 24 0.0 0:24.35 gnome-settings- [...] [/QUOTE]

are you after the fact that a single process can take more than 100% CPU? It’s easily understandable when you remember that “top” reports “percentage of a single CPU”, rather than “percentage of all CPUs available”… and if you take into consideration that a single process may run in parallel on two CPUs (if programmed with multi-threading capabilities).

In other words: That single process runs more than a single thread, in total these threads allocate scheduler time slices that would make make 1.08 CPUs busy.

Regards,
Jens