[plt-scheme] Scheme performance

From: Scott McLoughlin (scott at adrenaline.com)
Date: Sat Nov 14 12:57:03 EST 2009

Date: Sun, 08 Nov 2009 10:15:09 -0500
From: Neil Van Dyke <neil at neilvandyke.org>
Subject: [plt-scheme] Scheme productivity data point


I am glad along with you that PLT did the trick in better than adequate
time, but I do wonder how much time was spent in RTS routines written in
C/C++/asm and thus optimized by non-scheme compilers?

[...] The output files take less than 7% the space of the input files.
>> And, for people who think Scheme is slow... Wall clock time for the 
>> conversion (which was doing lots of parsing, XPath, simple 
>> pattern-matching, and a little regexp-ing) was only a little more 
>> than twice the wall clock time of "dd" doing comparable file I/O [to 
>> local disk on a fast machine].

I think the "ultimate" test is a set of "real world" oriented tasks/benchmarks
that use only routines written and compiled in Scheme, skipping all RTS routines 
excepting things like arithmetic, moves, branches, os interaction like opening
files and so forth.

So either an NFA or DFA based regex processor can be easily written in scheme,
so in my theoretical "benchmark"," the regex routines would have to be written
in Scheme, and so on and so forth.

But I ramble at this point. Sorry. Congrats on you positive performance
experience again.


Posted on the users mailing list.