[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Axiom-developer] The state of gcl/maxima/acl2
From: |
Camm Maguire |
Subject: |
[Axiom-developer] The state of gcl/maxima/acl2 |
Date: |
12 Dec 2002 12:12:06 -0500 |
Greetings! I thought this would be a good time for an update here, as
the acl2 work has basically been completed. All these results pertain
to the current GCL cvs, released as Debian package version -70/rpm
version -71, and currently found for all platforms (with the exception
of Windows, which is slightly older) at
ftp://ftp.gnu.org/pub/gnu/gcl/cvs.
Debian packages for maxima and acl2 can be found at:
http://ftp.debian.org/debian/pool/main/m/maxima/
and at
http://ftp.debian.org/debian/pool/main/a/acl2/
I encourage those with knowledge pertaining to the ???? fields below
to offer their amendments. Is there anything else being left out
here? Specifically, I haven't heard from BSD in quite a while.
When axiom is released, we'll try to add it to this table too. I'm
hoping that the automated process of producing binary packages on
these different machines as we continue to develop GCL will have the
dual effect of giving us early warning of any bugs we might introduce,
as well as provide incentive for end users to become interested in
Lisp.
The results quoted below pertain to the latest officially released
versions of maxima and ACL2. The results for maxima CVS are extremely
likely to be identical, based own my own tests on x86.
Take care,
=============================================================================
cpu OS BFD SGC compile max5.6 5.6test acl2.7 acl2.7cert Note:
(kernels)
x86 Linux yes yes pass pass pass pass pass
ppc Linux yes yes pass pass pass pass mostly (1)
sparc Linux yes yes pass pass pass pass pass
m68k Linux yes yes pass pass pass pass pass* (2)
arm Linux yes no pass pass pass pass pass (3)
s390 Linux yes yes pass pass pass pass pass (4)
mips Linux no yes pass pass pass pass pass
mipsel Linux no yes pass pass pass pass pass
alpha Linux no yes pass pass pass pass pass
ia64 Linux no yes pass pass pass pass pass
hppa Linux no no pass pass pass pass pass
sparc solaris yes yes pass pass pass pass pass (5)
x86 w32 no no pass pass ???? pass ????
x86 xBSD yes yes pass ???? ???? ???? ????
??? osX ??? ??? ???? ???? ???? ???? ????
(1) A few of the certifications currently fail because the image size
is too large for the native 24bit relocations without additional
trampoline code. This should be fixed by the -mlongcalls option
in forthcoming gcc-3.3, for which I think we should wait, unless
anyone needs all the certs on acl2 ppc immediately. Completing
the ACL2 port turned up cache flushing corner case bugs in GCL.
(2) Build will complete in a few days -- result anticipated :-).
Seriously, previous build succeeded in all steps before being
killed by the autobuilder for taking too long to generate output.
Completing the ACL2 port turned up cache flushing corner case bugs
in GCL, and one SGC related bug.
(3) Provides an example of how the -mlong-calls enables big images
like ACL2 on 24bit reloc machines like this.
(4) SGC recently enabled by s390 Linux kernel updates -- thanks
Gerhard!
(5) Maxima verified by me, acl2 shipped and certified by acl2
upstream.
--
Camm Maguire address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah
- [Axiom-developer] The state of gcl/maxima/acl2,
Camm Maguire <=