real = sum( for all processes, p ){ user(p) + sys(p) };
[download]

Modern computer systems are "time share" systems so the one process you timed isn't the only process getting CPU time. In fact, in the "for all processes", you need to include "the idle 'process'".

If you sum over all processes, you'll have to divide the result by the number of CPUs :)

Oh... i think that i understand now. You want to say, that if my application is sleeping on some kind of I/O system calls then Real Time will be Sys+Real+Sleep_Time?

What is the application? What time resolution do you need? What is the code you have tried?

My application is cgi script. I need at least milli-second. My code is very simple. At case of times(). I use:

$start = [times()];
...somecode...
$end = [times()];


$time_diff =($start->[0] + $start->[1]) - ($end->[0] + $end->[1]);
[download]

at case of Time::HiRes

$start=[gettimeofday];
...somecode...
$end = [gettimeofday];

$time_diff = tv_interval($start, $end);
[download]

But if real <=> sys+user, my first case will be always lower that at case 2

Your implementation context is CGI script, but that is not your application. Your application may be timing a database query or some such. So, what is the application?

You need to decide what it is that you want to report - wall clock time or process time.

With most operating systems that are not designed for real time environments reporting wall clock times of less than a few milliseconds is fairly meaningless because time slices for tasks are of that order and will render any real time reporting at that level meaningless.

---

Perl reduces RSI - it saves typing

The most obvious example where real isn't the sum of user and sys:

$ time sleep 3

real    0m3.002s
user    0m0.004s
sys     0m0.000s
[download]