jiffies vs real time

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

jiffies vs real time

William Cohen
The number of jiffies per second varies based on architecture and
distribution. As a result of this, the systemtap.samples/pfaults.exp
times out because the test runs for too long on the various i386
platforms, e.g.  RHEL4-U2, FC4 and i386 rawhide. On i386 machines 250
jiffies/second would cause the completion of the test to occur 40
seconds after the start, longer than the 30 seconds allowed in the test.

Could the translator allow the use of time units like seconds or
millsecond with "probe timer"? This would avoid having to revise the
tests when different numbers of jiffies per second are encountered. It
would also allow people to write tests the generate a report every x
seconds in a portable manner.

-Will
Reply | Threaded
Open this post in threaded view
|

Re: jiffies vs real time

Frank Ch. Eigler-2
Hi -

wcohen wrote:

> The number of jiffies per second varies based on architecture and
> distribution. [...]  Could the translator allow the use of time
> units like seconds or millsecond with "probe timer"?

Of course.  bug #1276 was opened months ago, and includes the
suggested "timer.ms(NN)" probe point.  It was pointed out multiple
times in the interim as a good little exercise for someone who wants
to learn hacking on the translator.

- FChE