Standalone byte code

[Simon Marlow]

Nope :-)

> -----Original Message-----
> From: Reuben Thomas=20
> Sent: Tuesday, May 01, 2001 11:57 AM
> To: Simon Marlow; Julian Seward (Intl Vendor); Simon=20
> Peyton-Jones; chak@cse.unsw.edu.au; cvs-ghc@haskell.org
> Subject: RE: Standalone byte code
>=20
>=20
> Unless you consider code that you JIT at load-time to be portable.
>=20
> -----Original Message-----
> From: Simon Marlow=20
> Sent: Tuesday, May 01, 2001 11:56 AM
> To: Julian Seward (Intl Vendor); Simon Peyton-Jones;
> chak@cse.unsw.edu.au; cvs-ghc@haskell.org; Reuben Thomas
> Subject: RE: Standalone byte code
> =20
> > [As you know] one alternative we've been discussing is doing
> > dynamic code generation from a low-level, portable, virtual code.
> > The RTS would contain a JITter translating to supported processors
> > for example x86, sparc, ppc, and could just have an interpreter
> > as a transitional mechanism for not-yet-supported platforms.
>=20
> Just pointing out that this route would also be rougly 10=20
> times slower than native code for the portable version, so=20
> there would still be no acceptably fast portable code.
>=20
> Cheers,
> 	Simon
>=20