[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PowerPC vs. MCL

In article <199402241438.AA22778@vipunen.hut.fi>, matti_karjalainen@hut.fi

> Finally a technical question. To me it seems that migrating to a native
> PowerPC compiler could be done in steps. As the first step, what about
> retargeting the existing LAP assembler to reasonably efficient PowerPC
> code. What I mean is using a M68 instruction model with PowerPC macro
> instruction replacements. I have undestood that LAP is a relatively small
> piece of software to be changed this way. What is the problem with this
> thinking ?
I think the problem with mixing emulation and native code is the time taken
in the context switch between the two.  I have the impression that the MCL
kernal is written in assembler.  If the kernal was left emulated it
wouldn't matter that the compiler spitted out Power code, everything would
slow down with the switch to the emulated kernal.

I have complete faith in the Apple Engineers to come up with a viable plan.
 I don't have such faith in Apple management paying for it.  Dylan is not
the answer.  As others have remarked it will take years for it to catch up
to MCL (and of course Dylan could be canceled by Apple management because
it isn't making Apple enough money ;- ).