Gmane
From: Thatcher Ulrich <tu <at> tulrich.com>
Subject: Re: Ditching C++
Newsgroups: gmane.games.devel.sweng
Date: 2006-07-03 16:48:06 GMT (2 years, 21 weeks, 6 days, 2 hours and 35 minutes ago)
On 6/30/06, Tom Hubina <tomh <at> vexxed.com> wrote:
> >
> > The argument you raise - needing to avoid good style to get speed - is
> > one that I raised myself against Java about nine years ago.  It isn't
> > true anymore and hasn't been true for a very long time.
> >
>
> Unless you're working on a crappy VM (like with J2ME phones, or any
> number of other environments).
>
> Having a garbage collected system is great when it works for you. It
> sucks a lot when it doesn't.

This is a fascinating paper, "Quantifying the Performance of Garbage
Collection vs. Explicit Memory Management".  They came up with a way
to compare the performance of Doug Lea's malloc with various GCs, in
real Java programs.  From the abstract:

"These results quantify the time-space tradeoff of garbage collection:
with five times as much memory, an Appel-style generational collector
with a noncopying mature space matches the performance of
reachability-based explicit memory management. With only three times
as much memory, the collector runs on average 17% slower than explicit
memory management. However, with only twice as much memory, garbage
collection degrades performance by nearly 70%."

http://www.cs.umass.edu/~emery/pubs/gcvsmalloc.pdf

I find it sobering.  If my application is performance sensitive, and
sensitive to memory size, then using GC is a bad idea.

-Thatcher
_______________________________________________
sweng-gamedev mailing list
sweng-gamedev <at> lists.midnightryder.com
http://lists.midnightryder.com/listinfo.cgi/sweng-gamedev-midnightryder.com