Caution: Time::HiRes returns a floating point number, but its precision is not necessarily very fine. For example, I noticed that on Windows, Time::HiRes::time() works in steps of around 55ms, or 1/18 of a second. That's the frequency of the system timer interrupt. | [reply] |
I've also seen (and commented on) this, but it only happens under special circumstances.
Under normal circumstances, with reasonably up-to-date versions of Perl and Time::HiRes and an NT-based version of Windows, you should be getting much better than 1/18 second resolution. On my machine it's in the order of 60 microseconds (1/16666th of a second):
P:\test>perl -MTime::HiRes=time -wle"print time for 1 .. 20; print $Ti
+me::HiRes::VERSION, ' ',
1121243012.87537
1121243012.87555
1121243012.87562
1121243012.87569
1121243012.87576
1121243012.87582
1121243012.87588
1121243012.87594
1121243012.876
1121243012.87607
1121243012.87614
1121243012.8762
1121243012.87626
1121243012.87633
1121243012.87639
1121243012.87645
1121243012.87651
1121243012.87658
1121243012.87664
1121243012.8767
1.59 5.008004
The special circumstances where the low accuracy results show up is under the debugger and in (some) evaled code. Despite spending a lot of time looking at this on 2 separate occasions, I cannot work out why Perl should produce different results under these circumstances.
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
Lingua non convalesco, consenesco et abolesco. -- Rule 1 has a caveat! -- Who broke the cabal?
"Science is about questioning the status quo. Questioning authority".
The "good enough" maybe good enough for the now, and perfection maybe unobtainable, but that should not preclude us from striving for perfection, when time, circumstance or desire allow.
| [reply]
[d/l] |