[Biojava-l] Compiling with Jikes

[Bradley A. Smith]

> > Granted jdk 1.3 is better than jdk 1.2; however, I still find that jikes
is
> > about 1.5 times faster than jdk 1.3. In addition, the bytecode produced
by
> > jikes runs faster.
>
> That's interesting.  Do you have any numbers for this?  And does
> it just apply to interpreters and naive JITs, or does it scale
> to things like Hotspot/ibmjit/Compaq Fast VM?
>

I don't have any comprehensive numbers; however, I have tested a primarily
numeric program on both Linux and Windows, and found that the best
combination on both platforms was jikes for compilation, and the IBM JRE 1.3
for runtime. I admit this result may hold for only this one program; however
if I can't compile with jikes, I can't test the performance. Hence, I'm
interested in compiling BioJava with jikes.

Thanks,
  Bradley