Millisecond time

[Sirius Black]

Hi all,
I've to realize a shell script which have to take the time and milliseco nd to analyze how many time the script took for its execution..
Have you any idea ??
Fabrizio

[Pedro Cirne]

Hi,

Use "time <script>" or "timex <script>"

Enjoy :-)

[harry d brown jr]

timex

man timex

live free or die
harry d brown jr

[Yogeeraj_1]

hi,

have a look at the following threads:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=13885
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=122073
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=402167

hope this helps!
regards
yogeeraj

[A. Clay Stephenson]

The timex command (which will give you .01 second resolution) is about as good as you are going to get from the shell. It would be possible to write a small C program or Perl script to give better resolution but the overhead of doing a fork and exec to spawn the child timer process from the shell would completely overwhelm any millisecond accuracy that you were trying to achieve.

[Sirius Black]

Hi,
I try with time ( not good) and timex ( so and so ) I've to take the millisecond time in order to evaluate how many time the process takes !!!
So there is a command, perl or sh which I can call at the start and the end of the script ??
Fabrizio

[A. Clay Stephenson]

Perl's Time::HiRes module would do this easily BUT it would only make sense to use it within the same Perl script. As I tried to explain to you before, the overhead of spawning an additional process completely masks any millisecond accuracy that you hoped to achieve. From the shell, timex is your least evil choice although it would be quite trivial to write a Perl script that would return millisecond timestamps. The problem is that with the overhead of forking and execing the Perl script, you would really need to throw away at least 50 msec of accuracy.