This is the mail archive of the
guile@sources.redhat.com
mailing list for the Guile project.
Re: help to get this running faster
- To: David Pirotte <david at altosw dot be>, guile-user at gnu dot org
- Subject: Re: help to get this running faster
- From: Ken Anderson <kanderso at bbn dot com>
- Date: Sun, 03 Sep 2000 09:50:38 -0400
- Cc: silk-user at lists dot sourceforge dot net
- List-Id: <guile-user.gnu.org>
At 05:24 PM 9/3/2000 , David Pirotte wrote:
>Hello,
>
>can people familiar with guile help me to get the following code optimised
>(so it runs as fast as possible, keeping the code entirely written in scheme)
>
>it's ok for 10 records or so, but little contact database (mine is 670
>contacts)
>already shows it's to slow
>
>i am an "old" cl developper, new to scheme, not knowing the "tricks" to get
>speedy code
I'm an old CL developer too, if you mean common lisp.
I'm not a guile expert. However, here are some things you can try.
T1: rewrite char-pos to not use call/cc. This is just a guess that call/cc might be expensive.
T2: rewrite load-records so it avoids reading the entire file into memory with read-line. I'd do this with a tiny tail recursive finite state machine. Since tokens are separated by #\tab and records by #\newline, you should only need a couple of states.
_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user