[Openmcl-devel] M1 port

[Bharat Shetty]

Sure, sounds interesting. Can you send the link to shannons trove again ?

On Sun, Jan 7, 2024 at 7:06 AM Tim McNerney <mc at media.mit.edu> wrote:

> Sure Bharat!
>
> Would you be interested in studying Shannon's trove of PowerPC
> architecture documents (I sent the link again to this mailing list a few
> days ago) and comparing the architecture and CCL's PPC register usage
> conventions to the ARMv8A (aka "ARM64") and its register usage
> conventions?
>
> At the very least look through the collection and tell us which documents
> are most useful for someone writing a compiler.
>
> Do you want to write an assembler and disassembler for the ARMv8A in the
> CCL style (i.e. compatible with the CCL compiler)?
> I particularly like "LAP" (Lisp assembly language) notation over typical
> assembler notations.
> This is useful for reading assembly language code into simulator written
> in Lisp.
>
> --Tim
>
> On 1/6/24 4:30 AM, Bharat Shetty wrote:
>
> Good to hear that :) If we need to understand ccl sources this needs an
> understanding of a lot of areas like assembly, instruction sets, linkers,
> compilers, memory management, number theory ?? etc. etc.. in depth.
>
> If we get a list of such topics/books/websites etc. I'd be glad to start
> off now. Any one pls ?
>
> Regards,
> Bharat
>
> On Sat, Jan 6, 2024 at 1:07 PM Tim McNerney <mc at media.mit.edu> wrote:
>
>> We hear you Bharat,
>>
>> One of our priorities is to write more internals documentation *while*
>> we work on the M1 port.
>> Part of the discipline will include everyone keeping (and checking in)
>> their contemporary notes/*journals*, which we will distill into a
>> document “how to port CCL to (yet) another processor” on an ongoing basis.
>> Another thing that is in the immediate plan is to write “slow but
>> correct” versions of all the *subprims* (e.g. bignum arithmetic) *in
>> Lisp*, which are all currently written in assembly language for every
>> CCL target architecture.
>>
>> --Tim
>>
>> On Jan 6, 2024, at 01:28, Bharat Shetty <bshetty at gmail.com> wrote:
>>
>> 
>> Agree people should contribute. However some good introduction to
>> internals and updating documentation would be very helpful. More important
>> than M1 port is we set this right. We might get M1 running after some
>> effort but in a couple of years we will end up in a similar situation and
>> talk about funding someone again.
>>
>> Most of us also have day jobs.
>>
>> Regards,
>> Bharat
>>
>> On Sat, 6 Jan 2024, 03:57 Andrew Shalit, <alms at clozure.com> wrote:
>>
>>> Hi Nicolas -
>>>
>>> Just to be 100% clear about this: Clozure as an entity does not exist
>>> anymore, nor does it exist as a loosely organized cabal. At this point
>>> Matthew Emerson manages the web domains and owns Github account where CCL
>>> is hosted, but that’s it.  I’m sure he would welcome as much help as anyone
>>> wants to provide and would happily give commit privileges to anyone who
>>> shows they can work on the code. If someone would rather fork and go wild,
>>> that’d be fine too.  But really, no one should let themselves be slowed
>>> down by thinking they need Clozure’s permission to do something.
>>>
>>> > On Jan 5, 2024, at 3:42 PM, Nicolas Martyanoff <nicolas at n16f.net>
>>> wrote:
>>> >>
>>> > I'm not disagreeing, but none of this is happening unless either
>>> Clozure
>>> > gives project admin right to someone really invested, or this someone
>>> > does the job of forking the project.
>>>
>>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clozure.com/pipermail/openmcl-devel/attachments/20240107/d44cb605/attachment.htm>