[ltt-dev] LTTV questions

Mathieu Desnoyers compudj at krystal.dyndns.org
Thu Feb 26 16:40:44 EST 2009


* Ashwin Tanugula (ashwin.tanugula at broadcom.com) wrote:
> Hi,
> 
> Can somebody please answer these questions.
> 
> 1. I see the following statistics listed for process with PID 3905
> 
> Statistic for  '3905-257.288103298' :
> 
> cpu time :           0.114179414
> events count :  2091
> elapsed time (includes per process waiting time) :           8.745760749
> cumulative cpu time (includes nested routines and modes) :           1.669448492
> 
> Shouldn't the cumulative cpu time be greater than the sum of cpu time and elapsed time.
> 
> Can somebody explain what these values mean.
> 

elapsed time and (especially) cumulative cpu time are parts of LTTV that
would need to be reviewed. But about your comment, elapsed time counts
the per process waiting time (scheduled out), which is not part of
neither cpu time nor cumulative cpu time. Therefore it's normal that
elapsed time is bigger, because it includes the time during which the
process has been scheduled out.

> 2.  I see the Birth Sec for init to be 257 on my trace data ( I probably started the trace at 256th sec)
> 
> Can somebody tell me what the Birth sec for each process means here.
> 

The time at which the process has been first active (either created or
scheduled for the first time) since the startup of the machine.

It helps differenciating between PIDs which are re-used over time.

Mathieu

> Thanks,
> Ashwin
> 
> _______________________________________________
> ltt-dev mailing list
> ltt-dev at lists.casi.polymtl.ca
> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
> 

-- 
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68




More information about the lttng-dev mailing list