why only one in eight cpu (or cpu threads) works?

[chuanlun]

hi, there are eight cpus (or say cpu treads) in my computer, I have installed fedora 17 (i686), but only one cpu works, why?
see top output below:
Tasks: 248 total, 2 running, 246 sleeping, 0 stopped, 0 zombie
Cpu(s): 16.4%us, 1.3%sy, 0.0%ni, 81.4%id, 0.8%wa, 0.0%hi, 0.0%si, 0.
Mem: 3312468k total, 3114264k used, 198204k free, 3288k buffers
Swap: 5406716k total, 1168360k used, 4238356k free, 2381424k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
21145 huwx 20 0 79464 10m 1048 D 21.6 0.3 0:03.38 bowtie
21147 huwx 20 0 79468 10m 1048 R 21.6 0.3 0:02.87 bowtie
21149 huwx 20 0 78944 9.8m 1048 D 15.7 0.3 0:00.16 bowtie
21128 huwx 20 0 79464 9.9m 1048 D 11.8 0.3 0:48.65 bowtie
21130 huwx 20 0 79476 10m 1048 D 11.8 0.3 0:41.45 bowtie
1 root 20 0 17108 2876 1424 S 0.0 0.1 0:15.26 systemd


The system details Installed on my computer
$ uname -a
Linux localhost.localdomain 3.4.2-4.fc17.i686 #1 SMP Thu Jun 14 22:19:00 UTC 2012 i686 i686 i386 GNU/Linux

[DBelton]

Your top output doesn't separate it by each core.

You can type 1 while in top and it will give you the cpu usage per core.

Something like this:

Tasks: 147 total,   1 running, 146 sleeping,   0 stopped,   0 zombie
%Cpu0  :  2.0 us, 18.0 sy,  0.0 ni, 76.0 id,  0.0 wa,  2.0 hi,  2.0 si,  0.0 st
%Cpu1  :  0.0 us,  2.0 sy,  0.0 ni, 98.0 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:   4020004 total,  1693240 used,  2326764 free,    72788 buffers
KiB Swap:  2047996 total,        0 used,  2047996 free,   719436 cached

  PID USER      PR  NI  VIRT  RES  SHR S  %CPU %MEM    TIME+  COMMAND           
  855 root      20   0  129m  16m 7228 S  10.2  0.4   2:24.36 X                 
 2219 root      20   0  115m 1444 1044 R   6.1  0.0   0:00.55 top               
    1 root      20   0 51232 7492 2256 S   0.0  0.2   0:12.96 systemd           
    2 root      20   0     0    0    0 S   0.0  0.0   0:00.02 kthreadd          
    3 root      20   0     0    0    0 S   0.0  0.0   0:00.04 ksoftirqd/0       
    4 root      20   0     0    0    0 S   0.0  0.0   0:00.17 kworker/0:0       
    5 root       0 -20     0    0    0 S   0.0  0.0   0:00.00 kworker/0:0H      
    6 root      20   0     0    0    0 S   0.0  0.0   0:00.24 kworker/u:0       
    7 root       0 -20     0    0    0 S   0.0  0.0   0:00.00 kworker/u:0H      
    8 root      rt   0     0    0    0 S   0.0  0.0   0:00.86 migration/0       
    9 root      rt   0     0    0    0 S   0.0  0.0   0:00.00 watchdog/0        
   10 root      rt   0     0    0    0 S   0.0  0.0   0:00.71 migration/1       
   12 root       0 -20     0    0    0 S   0.0  0.0   0:00.00 kworker/1:0H      
   13 root      20   0     0    0    0 S   0.0  0.0   0:00.04 ksoftirqd/1       
   14 root      rt   0     0    0    0 S   0.0  0.0   0:00.01 watchdog/1        
   15 root       0 -20     0    0    0 S   0.0  0.0   0:00.00 cpuset

[george_toolan]

Try a different system monitor like "Gnome System Monitor".

To find out if all the cores where recognized by the operating system try

cat /proc/cpuinfo

[chuanlun]

Follow DBelton's direction,
I am clear that the os can recognize eight cores and all of them are working, but there exists bottlenecks somewhere else, and make the total cpu usage like only one cpu works.

top - 12:49:45 up 48 days, 3:36, 12 users, load average: 8.07, 8.25, 8.4
Tasks: 247 total, 4 running, 243 sleeping, 0 stopped, 0 zombie
Cpu0 : 56.5%us, 2.3%sy, 0.0%ni, 19.1%id, 21.7%wa, 0.3%hi, 0.0%si, 0.
Cpu1 : 35.7%us, 1.7%sy, 0.0%ni, 20.3%id, 42.3%wa, 0.0%hi, 0.0%si, 0.
Cpu2 : 51.7%us, 2.3%sy, 0.0%ni, 5.7%id, 39.9%wa, 0.3%hi, 0.0%si, 0.
Cpu3 : 46.3%us, 5.0%sy, 0.0%ni, 4.3%id, 44.3%wa, 0.0%hi, 0.0%si, 0.
Cpu4 : 47.8%us, 2.0%sy, 0.0%ni, 22.7%id, 27.4%wa, 0.0%hi, 0.0%si, 0.
Cpu5 : 58.0%us, 2.0%sy, 0.0%ni, 26.7%id, 13.0%wa, 0.3%hi, 0.0%si, 0.
Cpu6 : 66.3%us, 7.7%sy, 0.0%ni, 1.3%id, 24.0%wa, 0.3%hi, 0.3%si, 0.
Cpu7 : 2.7%us, 1.0%sy, 0.0%ni, 90.7%id, 5.6%wa, 0.0%hi, 0.0%si, 0.
Mem: 3312468k total, 3112700k used, 199768k free, 10488k buffers
Swap: 5406716k total, 1136156k used, 4270560k free, 2280016k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
9867 huwx 20 0 89968 12m 1064 R 293.5 0.4 50:30.71 bowtie
10101 huwx 20 0 79468 10m 1044 R 26.9 0.3 0:00.81 bowtie
9904 huwx 20 0 79448 10m 1048 R 23.9 0.3 0:58.30 bowtie
9897 huwx 20 0 79304 9.9m 1048 D 23.6 0.3 1:04.59 bowtie
10032 huwx 20 0 79300 9.9m 1048 D 13.3 0.3 0:31.28 bowtie

[george_toolan]

Where do you get that impression?

load average: 8.07, 8.25, 8.4

That's pretty much eight cores working ;-)

One bottleneck is memory. You really should install a PAE kernel or even better a 64 bit operating system.

If you use a system monitor like gkrellm you might realize that the cores stop "working" when the system starts swapping.

[DBelton]

I agree.

You have over 1GB swapped out to disk, and that is very slow.

You are killing what benefits your processor gives you with the memory limitations. You need more than 4GB of RAM to run what you are currently running, and as george toolan states, you will have to change to either the PAE kernel or the x86_64 kernel to utilize more than 4GB.

Your other post about it not seeing all of the memory, where I suggested installing the PAE kernel (you said you had 16GB of RAM (4 4GB sticks), if you do that, then this problem should resolve itself.

Edit:

This reminds me.. Did you upgrade your memory after you installed Fedora? The installer is usually pretty good about installing the PAE kernel if it detects more than 4GB of RAM when it's run.

[chuanlun]

Thanks a lot, all friends. I have installed PAE, and it works fine in last month.
But I should say sorry to all, I have not visited my post, since my workload:-(((