I remember Igor made a comment not so long ago about fixing something in the idle process (the relinquishBlaBla primitive). Maybe it's related...


On Fri, Apr 25, 2014 at 2:35 PM, Norbert Hartl <norbert@hartl.name> wrote:
Here is mine. And it used to be around 2.

Norbert

��
top - 12:33:02 up 16 days, 15:45, ��1 user, ��load average: 5,03, 6,69, 7,02
Tasks: ��30 total, �� 1 running, ��29 sleeping, �� 0 stopped, �� 0 zombie
%Cpu(s): 13,8 us, ��0,6 sy, ��0,5 ni, 82,9 id, ��2,2 wa, ��0,0 hi, ��0,0 si, ��0,0 st
KiB Mem: ��32834868 total, 31193000 used, ��1641868 free, ��1449816 buffers
KiB Swap: 33553332 total, �� 672508 used, 32880824 free. 21207288 cached Mem

�� PID USER �� �� ��PR ��NI �� ��VIRT �� ��RES �� ��SHR S ��%CPU %MEM �� �� TIME+ COMMAND
16667 mobility ��20 �� 0 �� 72856 ��47096 �� 2188 S �� 1,0 ��0,1 ��65:07.04 pharo-vm
22820 inbox �� �� 20 �� 0 �� 70500 ��43904 �� 1344 S �� 1,0 ��0,1 ��26:37.38 pharo-vm
26745 events �� ��20 �� 0 �� 70500 ��40692 �� 1264 S �� 1,0 ��0,1 145:37.59 pharo-vm
32093 geo �� �� �� 20 �� 0 ��136308 ��47268 �� 1436 S �� 1,0 ��0,1 ��92:17.80 pharo-vm
10523 gcm �� �� �� 20 �� 0 �� 82528 ��53152 �� 2284 S �� 0,7 ��0,2 ��94:52.39 pharo-vm
11531 sms �� �� �� 20 �� 0 �� 70628 ��42096 �� 1380 S �� 0,7 ��0,1 ��91:49.13 pharo-vm
18739 stats �� �� 20 �� 0 �� 70496 ��41980 �� 1384 S �� 0,7 ��0,1 ��45:30.60 pharo-vm

Am 25.04.2014 um 07:30 schrieb Sven Van Caekenberghe <sven@stfx.eu>:


On 25 Apr 2014, at 01:19, Esteban A. Maringolo <emaringolo@gmail.com> wrote:

What % of CPU is it using?

My VM on idle state consumes about ~4-5% CPU.

That is about what I am used to, here is a real view of an Amazon AWS EC2 Small Instance running 4 Pharo VMs:

top - 05:25:09 up 413 days, 15:27, ��1 user, ��load average: 0.88, 0.66, 0.49
Tasks: ��73 total, �� 3 running, ��70 sleeping, �� 0 stopped, �� 0 zombie
Cpu(s): ��0.6%us, ��0.4%sy, ��0.0%ni, 96.3%id, ��0.2%wa, ��0.0%hi, ��0.0%si, ��2.5%st
Mem: �� 1706660k total, ��1654928k used, �� ��51732k free, �� 136692k buffers
Swap: �� �� �� ��0k total, �� �� �� ��0k used, �� �� �� ��0k free, ��1006556k cached

�� PID USER �� �� ��PR ��NI ��VIRT ��RES ��SHR S %CPU %MEM �� ��TIME+ ��COMMAND�� �� �� �� �� �� �� �� �� �� �� �� �� ����
31639 ubuntu �� ��20 �� 0 1030m ��99m 1444 R ��4.0 ��5.9 ��61:47.82 pharo�� �� �� �� �� �� �� �� �� �� �� �� �� �� ����
��4314 ubuntu �� ��20 �� 0 1030m 162m 1504 S ��2.0 ��9.7 140:28.55 pharo�� �� �� �� �� �� �� �� �� �� �� �� �� �� ����
��8014 ubuntu �� ��20 �� 0 1030m ��77m 1512 S ��2.0 ��4.6 561:06.30 pharo�� �� �� �� �� �� �� �� �� �� �� �� �� �� ����
31647 ubuntu �� ��20 �� 0 1030m ��92m 1444 R ��2.0 ��5.6 ��63:31.80 pharo�� �� �� �� �� �� �� �� �� �� �� �� �� �� ����
�� �� 1 root �� �� ��20 �� 0 ��3540 1392 ��732 S ��0.0 ��0.1 �� 0:58.61 init����

Sven

Esteban A. Maringolo


2014-04-24 15:43 GMT-03:00 Norbert Hartl <norbert@hartl.name>:
I have a strange effect on my machines. I have a dedicated server and a lot of virtual instances using��LXC. The last few days I migrated most projects of our projects from pharo2.0 to pharo3.0. And on the��same time the images moved from an ubuntu 12.04. system to a ubuntu 14.04 system. Looking at the top��command it appears the images/vms need noticable less CPU or better the tools report less. The monitored��overall CPU usage also is much lesser than before.

This is not proved by any measurement. Has anyone an idea why this could happen or appear differently?

Norbert