home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.programmer
- Path: newshost.grace.cri.nz!maths!peterm
- From: peterm@maths.grace.cri.nz (Peter McGavin)
- Subject: Re: 680X0 -> PPC translator?
- Message-ID: <PETERM.96Feb19135541@tui.maths.irl.cri.nz>
- Date: 19 Feb 1996 00:55:41 GMT
- References: <4fstt2$old@newsstand.cit.cornell.edu>
- Organization: Industrial Research Ltd
- In-reply-to: fkrj@crux4.cit.cornell.edu's message of 14 Feb 1996 15:08:18 GMT
-
- fkrj@crux4.cit.cornell.edu (Benjamin Kenobi) writes:
- >I'm just wondering, is there any good reason AT can't write a program which
- >will tranlsate binary executables from 680X0 -> PPC, doing optimizations along
- >the way? That way we could translate most of our old software and it would
- >run efficiently. Why didn't Apple do this with the PowerMacs?
-
- Automatic translation is a big subject. The major problem to overcome
- is how to distinguish code from data reliably. Automatic methods are
- usually misled by data in the code segment (e.g, constant strings),
- switch statements, procedure variables, modification of return
- addresses on the stack, arithmetic with the program counter or buggy
- code. Other problems include self-modifying or self-relocating code,
- interrupts and banging of different hardware. On some systems,
- translation of data formats is also an issue, e.g, fp formats,
- big/little endian.
-
- It is straightforward to translate cleanly structured, bug-free code.
- For example, VAX VMS executables follow strict rules so they can be
- translated to Alpha relatively easily. On the Amiga, however,
- anything goes.
-
- Emulation is easier, but slower. There are other possibilities, such
- as incremental translation, which is somewhere inbetween.
- --
- Peter McGavin. (p.mcgavin@irl.cri.nz)
-
-