On Thursday 22 May 2003 15:07, Miguel wrote:
> mea culpa, mea culpa
> I found the problem ... an "extra" return statement was short-circuiting
> much of the calculation in the .c version ... oops!
> The java version is, in fact, a bit *faster* than the C version.
> So, my faith in Java is restored :)
> I now feel motivated again to work on the 3D engine since I know there is
> no fundamental reason for us to be substantially slower than a C
BTW, I've just printed an article at IBM's website about the combination of
GCJ and their SWT toolkit (native Swing, so to speak). This enables one to
compile binary, machine specific code, removing the need for
interpretation... though you message shows that it actually no really longer
is a real problem. (Java seems to have matured enough).
Read it at:
It features a 7Mb download with Ant scripts ready to go...