Re: New Vax - future directions :-)

[Paul Koning <paulkoning%comcast.net@localhost>]

> On Jul 6, 2021, at 10:42 AM, Mouse <mouse%Rodents-Montreal.ORG@localhost> wrote:
> 
>> So I'd say [EMUL/EMULU] definitely have value.  However, it's a fair
>> point that it might be hard to get compilers to make use of it.
> 
> Perhaps.  But it can be used under the hood by large-number support
> routines, either libc "word×word->doubleword" or large integer
> arithmetic such as is used by modern crypto.  Even if there are only
> two EMULU instructions in an executable, they can still make a
> significant difference.
> 
>> As is the case for so much of the stuff in the VAX.  It's good for
>> people writing in assembler, but difficult to make use of by
>> compilers...
> 
> Well, modern compilers, because they attempt to be multi-target.  Build
> a VAX-target compiler, with no care for non-VAX targets, and it becomes
> easier.

GCC certainly can handle single * single -> double operations, that's a standard pattern in the back end support machinery.  PDP11 and VAX both use it, as do 28 other targets GCC knows about.

>>> You don't have any modern language that want to deal with the upper
>>> part anyway.
> 
> Only antique languages (like Lisp) with bignum support, or fringe
> applications like cryptography.

I don't think Python 3 is an antique language, and it uses bignum arithmetic for all its integer arithmetic.  Python 2 do so optionally.  And cryptography is about as far from a fringe application as you can get.

	paul