Suitability (of Java)

John Quinn john at johnquinn.com
Sat Apr 3 10:48:32 PST 2004


On Saturday 03 April 2004 12:15, Bill's LFS Login wrote:
> On Sat, 3 Apr 2004, Don Smith wrote:
> > John Quinn wrote:
> > > Don may be surprised to see these results:
> > > http://osnews.com/story.php?news_id=5602
> >
> > The only surprise was that Java 1.4.2 was really slow at trig. Is anyone
> > surprised that compilers written by MS produce the fastest code on an OS
> > written by MS?
>
> More surprising to me was C performance. As with any tool, skill of use
> may have effects and we don't know where his strengths and weaknesses
> may have come into play.

Yeah, I don't know what's up with that trig thing.  It'd be nice if we could 
figure out what happened.  But he mentioned in the article he didn't do
a lot of tuning or have the expertise to do so.  They all had their 
weaknesses but if you take out the trig, even I was surprised Java did 
so well.  I was also pleasantly surprised but Tushar's test.  That JIT stuff
really helps.  And that was with starting the JVM too.  I wonder what GCJ
would do to that.

-- 
/**
 * @author	John Quinn <john at johnquinn dot com>  LFS 2454
 */



More information about the hlfs-dev mailing list