With regard to the change from PPC to Intel on the Mac platform (see
http://www.apple.com/pr/library/2005/jun/06intel.html), will OpenMCL develop
to generate native code for there Macs, or will OpenMCL programs rely on
compability modes (emulation) in the future?
--
Oyvin
Am 10.10.2005 12:32 Uhr schrieb "Oyvin Halfdan Thuv" unter
<·········@orakel.ntnu.no> in ······················@apollo.orakel.ntnu.no:
> With regard to the change from PPC to Intel on the Mac platform (see
> http://www.apple.com/pr/library/2005/jun/06intel.html), will OpenMCL develop
> to generate native code for there Macs, or will OpenMCL programs rely on
> compability modes (emulation) in the future?
It would be best to ask that question on the OpenMCL mailing list or
directly to 'Clozure Associates'.
a) does OpenMCL run in the emulator? I haven't heard anything about that
yet.
b) will there be a port to native x86 code generation? 32bit and 64bit?
Clozure ( http://www.clozure.com/ ) says the following on their web site:
"Intel Support
Apple Computer recently announced its intention to move from the PowerPC to
the Intel x86 architecture. Clozure Associates remains committed to the
development of OpenMCL, and will move it forward through this transition and
beyond. We are actively exploring how best to support Apple's new
architecture and how to take advantage of new opportunities it affords users
of OpenMCL."
So, I guess financial support for the transition would be welcome...
Regards,
Rainer Joswig
Hi!,
I wrote to Gary Byers regarding this a coupla years ago. His reply was
that its a pain to port OpenMCL to a register starved arch like the
x86.
cheers,
-Krishna
In article <························@g49g2000cwa.googlegroups.com>, Krishna wrote:
> I wrote to Gary Byers regarding this a coupla years ago. His reply was
> that its a pain to port OpenMCL to a register starved arch like the
> x86.
I can imagine, yes. The situation is somehow different now, however, it's more
like a migrate-or-die situation.
The comments from R. Joswig seems promising though, and I hope they have the
funds for continuing developement of the product. Personally I'd happily pay a
shareware fee for a software package like OpenMCL.
--
Oyvin
Krishna wrote:
> Hi!,
>
> I wrote to Gary Byers regarding this a coupla years ago. His reply was
> that its a pain to port OpenMCL to a register starved arch like the
> x86.
Maybe, but most languages (even directly compiled ones) have
x86-backends. It's mostly a question of spilling registers, changing
the calling-conventions to use the stack (goodbye cheap-n-easy tail
calls), and dealing with weird instructions (like div and mul only
accepting arguments in certain registers).
Unfortunately I never built a real compiler myself, and talk is cheap ;)
--
State, the new religion from the friendly guys who brought you fascism.