[GE users] usage_scaling for exec hosts

goncalo goncalo at lip.pt
Wed Sep 9 15:19:04 BST 2009


Dear All...

I have a simple question regarding cpu accounting.

1./ I've started a simple job which took the following cpu time in a 
give host:

     [root at sge01 common]# qacct -j 26163 | egrep 
"(cpu|hostname|start_time|end_time)"
     hostname     wn111.ncg.ingrid.pt
     start_time   Wed Sep  9 15:00:20 2009
     end_time     Wed Sep  9 15:02:59 2009
     cpu          103.532

2./ Afterwards, I've defined to that host

     [root at sge01 common]# qconf -se wn111.ncg.ingrid.pt | grep usage_scaling
     usage_scaling         cpu=1.200687

3./ I've re-run the same job and got:

     [root at sge01 common]# qacct -j 26164 | egrep 
"(cpu|hostname|start_time|end_time)"
     hostname     wn111.ncg.ingrid.pt
     start_time   Wed Sep  9 15:06:05 2009
     end_time     Wed Sep  9 15:08:44 2009
     cpu          134.317

So, I see that the difference between start_time and end_time is 159 
seconds, and the same for both jobs (As it should since it is exactly 
the same job executing in the same machine). Nevertheless, I'm getting a 
cpu ratio of  134.317/103.532 = 1.29734768 when I initially configured 
1.200687.

So, what I'm I doing wrong here?

Cheers
Goncalo

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=216590

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list