I ran it around 15 times and the winner changes radically, both yield random numbers between 40-80 in Chrome.Non-caching seems to win most of the time but with the change rate, the numbers don't seem reliable.
On Mon, Jul 20, 2009 at 1:55 PM, John Resig <jere...@gmail.com> wrote: > You need to be careful with your testing. Executing a function which has > the regexp will result in non-deterministic results (the function call > itself will probably take longer to run than the actual regexp being > executed). > > A revised test: > http://dev.jquery.com/~john/test/regexp.html > > Results (New vs. Old): > FF 3.5 123ms 119ms > SF 4.0 8ms 11ms > IE 6.0 220ms 490ms > > As you can see, the new engines do optimize this quite well - but Internet > Explorer is definitely a different matter. > > --John > > > On Mon, Jul 20, 2009 at 10:28 AM, Ariel Flesler <afles...@gmail.com>wrote: > >> You're right, the regexps was too simple. I updated the test with longer >> and more complex expression and the time is still the same. >> >> I agree about caching regexps that are re-used in many places, obviously. >> Not so much about one-time expressions or strings. >> I'm really aware of how the different interpreters work, but it seems to >> me that immutable literals (strings, numbers, regexps) are the first thing >> I'd cache in a constant. Maybe current interpreters do that already >> internally ? >> >> On Mon, Jul 20, 2009 at 12:56 PM, John Resig <jere...@gmail.com> wrote: >> >>> It has less to do with RegExps and more to do with inline objects. You'll >>> note that I also moved inline functions out and declared them above. I'd, >>> eventually, like to do the same with inline arrays and object literals. >>> RegExp, Function, Array, and Objects shouldn't be re-declared on every >>> function call (or on every iteration in a loop!). >>> >>> While very-simple RegExp optimizations may be a wash, performance-wise, >>> (I note that your test regexp is only against a couple characters and >>> doesn't even use a capturing expression) the code simplification and >>> minification benefits are very real. Going through this process I was able >>> to remove a large number of duplicate RegExp - and that's always a good >>> thing. >>> >>> --John >>> >>> >>> On Mon, Jul 20, 2009 at 9:20 AM, Ariel Flesler <afles...@gmail.com>wrote: >>> >>>> >>>> Hi >>>> >>>> I've been seeing the multiple commits for >>>> http://dev.jquery.com/ticket/4111. >>>> Moving regexps to local vars to cache them. >>>> >>>> I was wondering if this is a perfomance measure and if so, if there's >>>> some benchmark out there to prove the gain. >>>> I personally think these changes make the code more complicated and >>>> longer. I'd rather see the regexp in place if I want to understand >>>> what it does. >>>> >>>> I went ahead and created a benchmark for this: >>>> http://benchmarker.flesler.com/regexp/run/ >>>> >>>> The conclusions were that there's absolutely no gain in speed except >>>> on IE. Even on IE, it doesn't seem to be such a critic improvement and >>>> it's slower the more you get away from the variables' scope chain (in >>>> nested functions). >>>> >>>> Any opinion ? >>>> >>>> Cheers >>>> -- >>>> Ariel Flesler >>>> >>>> >>> >>> >>> >> >> >> -- >> Ariel Flesler >> http://flesler.blogspot.com >> >> >> > > > > -- Ariel Flesler http://flesler.blogspot.com --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jQuery Development" group. To post to this group, send email to jquery-dev@googlegroups.com To unsubscribe from this group, send email to jquery-dev+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/jquery-dev?hl=en -~----------~----~----~----~------~----~------~--~---
