Depending on what you want to do, it might be faster to keep the data structure in C land and access it through a tied array or unpack from the Perl side. You can gain a tiny speedup by using av_make to create the array in the DIFF_MATCH case (see perlguts on working with arrays).
I would look at using "bind" variables and writing the data from e directly into these variables instead of (re)allocating new Perl data structures every time. DBI has this for the maximum speed at a cost of convenience.
In reply to Re: Faster creation of Arrays in XS?
by Corion
in thread Faster creation of Arrays in XS?
by wollmers
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |