SHUDO Kazuyuki wrote:
>
> Nathan Meyers <[EMAIL PROTECTED]> wrote:
>
> > I don't mean to be heretical, but using an unoptimized sample
> > implementation is not always the best solution for mission-critical
> > server applications.
>
> Sure. But,
You're right, I was painting with too broad a brush. IBM's results, as
shown in the benchmarks, have been very impressive. I'm looking forward
to seeing their JDK1.3.
Nathan
>
> > The Blackdown ports, and the more recent IBM and
> > Inprise ports, are a huge boon to Java developers and users on Linux -
> > but they're sample implementations that have ported a lot of inefficient
> > code directly from Sun.
>
> IBM has been modifying large part of Sun's classic
> VM. Especially in memory system. It's not only a ported
> version.
>
> For example, handle is removed. To be exact, With
> classic VM, double memory references is needed to reach
> object data from a handle, but IBM's mixed mode
> interpreter (MMI), Sun's HotSpot VM and ResearchVM
> require only single reference.
>
> The source code for IBM JDK is not released publicly
> yet, but we can see some part of it in some header files
> located at JDK_HOME/include/, oobj.h and interpreter.h.
>
> Kazuyuki SHUDO Happy Hacking!
> Muraoka Lab., Grad. School of Sci. & Eng., Waseda Univ.
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
