Re: Three more shoot outs

[Joshua Isom]

I do tend to use the latest revision, rarely more than a day old. I've made up a quick script to make parrot, mainly because I have gmp and gdbm installed by fink. I have --optimized enabled now. I'm primarily using and 800Mhz PPC, 512k cache. I rarely get any difference in speeds between -j and -C that would be out of a statistical norm, and unless I'm testing with -p or -t, I use -C. For a lot of the benchmarks, my installed perl is actually faster than theirs, which is fairly surprising to me.

Anyway, I've got it working uses all the regexes. I stuck to the p6 rules, and kept the hash to print out the regex they want to see. It's been running now for an hour now and it hasn't even reached the main matching yet for the full benchmark. This is my first use of PGE so if anyone has any ideas of how I can improve removing/replacing sections of text I'm welcome to hear it. Currently I just go through and find all the matches, storing them in an array, and then, starting at the end of the string, remove/replace the text so the alignments still are accurate. I made a file(fasta N=100000), and tested it, which is about one meg, one fifth the benchmark, and it took me 13 minutes. Perl takes about six and a half seconds for the same file... The only caveat now from the benchmark is that substitution isn't done by the regex.

regexdna.pir
Description: Binary data