I'm thinking about slow. Specifically I'm wondering what is
slow? As in, what important things written in Perl are slow, and seem to be
more slow than they should be. Perl applications that appear to be giving the
interpreter heartburn. I'm looking for volunteers for experimenting on.
Specifically, there's an item in perl
todo titled Profile Perl - am I hot or not? But it's not clear to
me what would be good candidate Perl source code to run against the interpreter
to profile it. So, dear lazyweb, useful suggestions please...
PPI
Alias on 2006-04-19T16:00:09
PPI is slow. Really slow. It handles about 1500 lines per gigacycle (on average) but when you run it against a big directory of stuff, it still feels really slow.
I've optimised (and in a few places micro-optimised) the hell out of it. I really don't know what more I can do. Maybe there's something there for you.
One clue that sticks out is that UNIVERSAL::isa (the function) shows up in some of my profiling reports for it.
Perhaps there's something that could be done to make UNIVERSAL::isa faster?
Plucene
Smylers on 2006-04-19T16:20:21
I believe that Plucene's slowness is the main thing preventing it from being viable. Or at least it was at YAPC Belfast; things may have changed since then.
Smylers
Using Phalanx to answer "what is slow?"
tomhukins on 2006-04-19T17:10:13
The Phalanx Project identifies important, commonly used modules. You could use the modules listed to discover where people might care about Perl running slowly.
