[MLton] Re: MLton

[John Skaller]

--=-ArnoJgPV0qc9Tcbplnl6
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Wed, 2005-06-29 at 12:17 -0700, Stephen Weeks wrote:

[re MLton on amd64, transfering mail to MLton list from private]

> > Some serious bugs with casts (as usual), plus unknown arch.
>=20
> A couple of x86_64 patches have been submitted since 20041109.  You
> might try the latest CVS.

Ok, I shall do so.

> > Is there work on amd64 port? I am willing to help,=20
>=20
> Thanks for the offer.  Matthew and I do plan to do a 64-bit port this
> year, starting any day now. =20

Good. I am benchmarking various systems myself, since
I no longer trust the Shootout.

So far my system Felix is faster than all other languages
on the two tests (ackermann and takfp) that I am running.
This is somewhat surprising to me since it generates C/C++
code .. but that code still runs faster with gcc than gcc
does for the C version .. for ackermann it is twice
as fast as Ocaml native code compiler .. I don't really know
why, I'm still checking the test methodology.

The bottom line for MLton is that a bytecode version
is no use -- I want something FASTER than Felix :)

I could run the x86 version of course (it should run
on an AMD64 I think) and I have an old x86 as well.

FYI: Felix is an ML like language dressed up a bit
like C/C++ to make it palatable to C++ programmers.
It emphasises C/C++ binding, and also does whole
program analysis.

> BTW, feel free to direct emails such as yours to the MLton@mlton.org
> list.

Done .. :)

--=20
John Skaller <skaller at users dot sourceforge dot net>
Download Felix: http://felix.sf.net

--=-ArnoJgPV0qc9Tcbplnl6
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQBCwxf5w7vY25tjqN4RAjJgAJ41/l37aHVDnV81VaU0zrcmOaiIGQCfXP14
EnagseKY3KcQSPK3WiFqhyk=
=k1TK
-----END PGP SIGNATURE-----

--=-ArnoJgPV0qc9Tcbplnl6--