On Wed, 16 Mar 2005, voguemaster wrote:

> > Just my $0.02: IBM's implementation of the JVM is *not* optimized for
> > Itanium (and probably won't be). It is optimized, however, for other 64bit
> > processors such as AMD64 and Power4/5.
> >
>
> That's a different kind of optimization. In any case, if you write your
> code properly and you run your program in the correctenvironment, Java
> math speed is comparable to native code. To all those people who believe
> that Java is a performance hogging monster, let me burst your bubble
> right here and now.

unfortunately, java indeed is a performance hog in various areas - and
just by saying "it is not", you're not going to burst anyone's bubble ;)

it _can_ be very close to the speed of C or C++ in various aspects, but in
other aspects it is much slower. i've seen a very simple algorithm that
just by translating its inner loops to C and using jni, got a performance
increase of X3 - and i think this all came due to java's arrays bounds
checking.

i won't even mention java GUI programs. it could be that you can write a
GUI program in java that'll not appear slagish (i've seen some
minimalistic java applets that run quite fast) - but looking at the
various GUI programs i have met so far, it looks like it is much harder to
achieve this with java, then with C++ (or C).

sorry - you did not burst "my bubble" just yet :P~~

-- 
guy

"For world domination - press 1,
 or dial 0, and please hold, for the creator." -- nob o. dy

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to