time() and $^T will be giving you elapsed real time, but in whole seconds.

To get something with a finer grain, you should look at Time::HiRes. That provides:

• Time::HiRes::time() which you can use instead of time() which will give whatever resolution your system will support -- it's reasonable to expect at least milli-seconds, and probably micro-seconds. You will have to do your own $Start_Time = Time::HiRes::time(), early on in the script, though. (Sadly, it does not seem to be possible to $^T = Time::HiRes::time().

• you can also (Update: but not under Winders):

    use Time::HiRes qw(clock_gettime CLOCK_REALTIME) ;
  
    my $t = clock_gettime(CLOCK_REALTIME) ;  # same as: my $t = Time::Hi
  +Res::time()
  [download]

• and if you want the cpu time, rather than the elapsed time, then:

    use Time::HiRes qw(clock_gettime CLOCK_PROCESS_CPUTIME_ID) ;
  
    my $cpu = clock_gettime(CLOCK_PROCESS_CPUTIME_ID) ;
  [download]

  should give you cpu time since the process started. CLOCK_PROCESS_CPUTIME_ID is a POSIX thing, which is probably implemented on your system (Update: but not under Winders).

Update: Time::HiRes::time() is supported on Winders. Unfortunately Time::HiRes::clock_gettime() isn't... Time::HiRes::clock() I find is also not implemented on Winders :-(

However, testing on Linux and Windows XP, suggests that the following is supported by POSIX:

  use POSIX qw(CLOCKS_PER_SEC) ;

  my ($realtime, $user, $system, $cuser, $csystem) = POSIX::times() ;
[download]

where this is retrning: the elapsed realtime since some point in the past (such as system startup); user and system times for this process; and user and system times used by child processes. They are in units of CLOCKS_PER_SEC, which on my XP is 1000 (so ms).