Re: [vhdl-200x] Modular types, alternative solutions

Frankly, before we go supporting modular types, I'd like to see longer 
integers than 32 bit signed integers.  That limitation is more of a pain 
to work around than working with modular arithmetic using the language 
as-is.

On 10/23/2014 5:47 PM, Joseph M Gwinn wrote:
>
> I've been listening in to the debate on modular types et al, and have 
> a practical comment to offer, from the days when we coded directly on 
> the iron:
>
> Most or all modern computer instruction sets support multi-precision 
> integer arithmetic to some degree, especially addition and 
> subtraction, where "multi-precision" means integer multiples of a 
> 32-bit or sometimes 64-bit integer words.  Such arithmetic can execute 
> very fast, even though it's done in software, and the code is simple.
>
> To get the details, look at the CPU hardware instruction set 
> documentation.  One can always access the data and instruction using 
> assembly code.  Some kinds of C probably also have the hooks and 
> handles.  We are talking about maybe ten lines of code, so portability 
> between hardware platforms isn't really necessary.  One just provides 
> an assembly-coded C function for each CPU type.
>
> Joe
>
>
> Inactive hide details for Jim Lewis ---10/23/2014 05:34:21 PM---Hi 
> Tristan, >> The second proposes adding discriminants (anotheJim Lewis 
> ---10/23/2014 05:34:21 PM---Hi Tristan, >> The second proposes adding 
> discriminants (another concept borrowed
>
> From: Jim Lewis <jim@synthworks.com>
> To: vhdl-200x@eda.org
> Date: 10/23/2014 05:34 PM
> Subject: Re: [vhdl-200x] Modular types, alternative solutions
> Sent by: owner-vhdl-200x@eda.org
>
> ------------------------------------------------------------------------
>
>
>
> Hi Tristan,
> >> The second proposes adding discriminants (another concept borrowed
> >> from ADA).   I am actually surprised that ADA did not do modular
> >> types this way as this provides a general feature that we can use
> >> for other things, such as saturating types.
> > Your discriminant proposal doesn't allow large modular types, and
> > doesn't provide strong typing :-(
> The proposal shows an example using current VHDL types.
>
> Certainly if integers are extended, the extended integers could be 
> used.  Hard to base a proposal on something that does not exist yet. 
>  In addition, as a general feature, discriminants could be used
> to create a saturating integer type or even a real number with a 
> modulus (independent as to whether it makes sense or not).
>
> Cheers,
> Jim
> p.s.  I added notes on your thoughts about generics.  It would be 
> interesting to hear your reply.  Also note that Cliff W has been 
> experimenting with a package that uses generics to do modulo types.
>
> -- 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Jim Lewis                                  Jim@SynthWorks.com
> VHDL Training Expert http://www.SynthWorks.com
> IEEE VHDL Working Group Chair
> OSVVM, Chief Architect and Cofounder
> 1-503-590-4787
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>
> -- 
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
>
>
>
> -- 
> This message has been scanned for viruses and
> dangerous content by *MailScanner* <http://www.mailscanner.info/>, and is
> believed to be clean. 


-- 
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

   "They that give up essential liberty to obtain a little
    temporary safety deserve neither liberty nor safety."
                                            -Benjamin Franklin, 1759
  


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.