On Fri, Aug 1, 2008 at 2:30 PM, Robert Cummings <robert@xxxxxxxxxxxxx> wrote: > On Fri, 2008-08-01 at 13:24 -0400, Andrew Ballard wrote: >> On Fri, Aug 1, 2008 at 12:34 PM, Robert Cummings <robert@xxxxxxxxxxxxx> wrote: >> > On Fri, 2008-08-01 at 11:12 -0500, Boyd, Todd M. wrote: >> >> > -----Original Message----- >> >> > From: bk@xxxxxxxxxx [mailto:bk@xxxxxxxxxx] On Behalf Of Yeti >> >> > Sent: Friday, August 01, 2008 10:58 AM >> >> > To: Andrew Ballard >> >> > Cc: PHP General list >> >> > Subject: Re: Internationalisation and MB strings >> >> > >> >> > Oh right. Doing 1 measurement only is not even worth a theory. >> >> > >> >> > Well, I'm wondering how much PHP can speed that result up, since we are >> >> > calling the same function with the same parameter 10000 times. Wouldn't >> >> > it >> >> > be even more realistic if we called it with changing strings? >> >> >> >> ---8<--- snip >> >> >> >> > > I ran this script several times, and the results below are fairly >> >> > typical: >> >> > > >> >> > > MB_STRLEN took : 0.054733037948608 milliseconds >> >> > > >> >> > > STRLEN took : 0.037568092346191 milliseconds >> > >> > How did you measure these? I don't recall a time function that returns >> > milliseconds... only functions that return seconds (and parts thereof in >> > microseconds)... such as microtime(). >> > >> > Cheers, >> > Rob. >> >> If microtime(true) returns fractional seconds, wouldn't multiplying >> the result by 1000 (as he did) convert the units from seconds to >> milliseconds? >> >> 1 second = 1000 milliseconds = 1000000 microseconds, thus 0.000037568 >> seconds = 0.037568 milliseconds, correct? Am I missing something? > > I didn't see the multiplication by 1000 and the above numbers don't show > any trailing zeros or truncation of precision as I would expect if the > number had been multiplied by 1000. Although, maybe the multiplication > occurred before the division by the number of runs. It just looked > suspect to me. > > Cheers, > Rob. I've had moments like that. (One just yesterday in fact. :-) ) I did subsequently notice with the OP's code you will occasionally get a negative number, too. Apparently the code enters some sort of space/time continuum and sometimes manages to go back in time a few microseconds. (Actually, there is a much more rational explanation for it if anyone cares, but this one sounds cooler.) Andrew -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
