To: bsd-dk@hotel.prosa.dk Subject: Re: FreeBSD og danske tegn From: Anton Berezin <none@tobez--plab.ku.dk.lh.bsd-dk.dk> Date: 17 Mar 1999 14:14:00 +0100
Philippe Regnauld <none@regnauld--deepo.prosa.dk.lh.bsd-dk.dk> writes:
>> Eventually I will get around to write my own with the functionality
>> I want, in Perl.
> So you can use 25 MB of RAM, and 100% CPU for 1 minute to read a
> 3000 mail mailbox ?
Well, one day the following appeared in comp.lang.perl.moderated:
From: Tom Christiansen <none@tchrist--mox.perl.com.lh.bsd-dk.dk>
Subject: Debunking Speed Myths
Newsgroups: comp.lang.perl.moderated,comp.lang.perl.misc
Date: 7 Mar 1999 10:12:05 -0700
The new book from legendary authors Brian Kernighan and Rob
Pike, _The Practice_of_Programming_
(http://cm.bell-labs.com/cm/cs/tpop/index.html) discusses Perl
a bit. One of the interesting points is some timings. They
wrote a Markov chain program in various langauges, and tested
it on various platforms. Here are some results (pp 80-81).
PLATFORM 250mhz Irix 400mhx pII M/S line count
C 0.36 sec 0.30 sec 150
Java 4.9 9.2 105
C++/STL/deque 2.6 11.2 70
C++/STL/list 1.7 1.5 70
Awk 2.2 2.1 20
Perl 1.8 1.0 18
The code is in the book, including analyses of what's going
on.
<snip>
Interesting, how would you back up your claims in regard to Perl
speed/memory usage?
-- Anton Berezin <none@tobez--plab.ku.dk.lh.bsd-dk.dk> The Protein Laboratory, University of Copenhagen
This archive was generated by hypermail 2b30 : Wed 15 Nov 2006 - 18:24:00 CET