Clear questions and runnable code get the best and fastest answer |
|
PerlMonks |
Re: Re: what's faster than .=by pg (Canon) |
on Mar 08, 2003 at 06:04 UTC ( [id://241346]=note: print w/replies, xml ) | Need Help?? |
Wow, just to add a little bit more. Generally speaking, realloc costs a lot, and is slow. That's why lots of c programmers, when they call realloc, they always reallocate more than they need at the moment, could be couple humdreds times of what they need, so they can largely reduce the frequency they calling realloc. For example, if one needs to allocate 4 more bytes each time, and knows that there is a chance that he would come back and repeat this again and again, why not simply allocate 400 more bytes each time, so he can reduce the frequency of calling realloc by 99%. However even with this optimization, realloc might still cost you a lot. For example, you need to realloc 1,000,000 4 bytes, if you realloc 400 bytes a time, you still need to call realloc 10,000 times. HOWEVER, Perl string does not work in this way, Perl only allocates what you want at the moment, FORTUNATELY, you still can pre-allocate the space by yourself. Just do something like:
In Section
Seekers of Perl Wisdom
|
|