Walt Mankowski on Tue, 2 Jul 2002 11:07:22 -0400 |
On Mon, Jul 01, 2002 at 02:12:35PM -0400, Darxus@chaosreigns.com wrote: > FYI, we still don't have a speaker scheduled for this wednesday. This will > be the first time since the middle of 1999. Back then it was fairly normal > to not have a speaker. This is fine, in my opinion. We just have a larger > group than we did back then. I can give my talk entitiled "A case study in when memoization doesn't work", which i just gave at the YAPC conference last week. Here's the abstract: Memoize.pm is a great tool, but sometimes it can actually make your program run *slower*. In this talk I'll present a case study of how I attempted to use memoization to speed up a program, and in the process made it run an order of magnitude slower. As we walk through various iterations of a program we'll discuss optimization techniques, profiling, some very rudimentary Perl internals, and even some number theory. It's only about a half hour talk, mainly because I assumed my audience already knew Perl and knew what memoization was. I can add some slides explaining memoization and get to talk to about 45 minutes. It would help if people knew a little Perl, but the techniques I'll be discussing should be useful for most programming languages. Walt Attachment:
pgpk3N8IOQSdb.pgp
|
|