From: Eric Crampton <EricCrampton@...>
Date: 2000-12-16T07:40:01+09:00
Subject: [ruby-talk:7358] Re: Ruby performance question

Dave Thomas <Dave@Thomases.com> writes:

> Eric Crampton <EricCrampton@worldnet.att.net> writes:
> 
> > Running on a small dataset of 13 MB of data (around 81000 lines), the
> > program takes about 27 seconds on my Linux machine. An equivalent Perl
> > code to do the same thing runs in about 10 seconds.
> 
> What happens if you disable garbage collection before you start, and
> re-enable it after you read in the whole file? It could be that with
> the large number of objects you're creating the GC scan could be quite 
> expensive.

I'm (generally) reading more data than could fit in physical memory,
so that's not really an option.

Incidently, I tried disabling GC and running it on my small dataset
(13 MB) which I would think would fit in physical memory. Somehow, the
way Ruby is organizing data in memory causes it to store far, far, far
more than 13 MB. I ran out of physical memory (1 GB of RAM) and it
went into swap. :-(

-- 
"A truly lazy person will work very hard to not have to work
later." --Larry Wall