[Libre-soc-dev] clamping/saturation semantics

[Hendrik Boom]

On Sat, Dec 12, 2020 at 07:10:46PM +0200, Lauri Kasanen wrote:
> On Sat, 12 Dec 2020 11:55:16 -0500
> Hendrik Boom <hendrik at topoi.pooq.com> wrote:
> 
> > > But s16-s16=s8 cannot. Consider 1-256 - if that op was done at 8-bit
> > > width, it would become 1-0=1. When the correct result is -255 (0xfeff),
> >
> > Would the correct result not be 0xff01 ?
> 
> Oh, indeed it would (checked with quick C). Then the signed sub
> question becomes whether chopping should preserve sign or not.
> 
> Perhaps we need to come up with a less controversial example. Div?

At least the size of an exact integer quotient is bounded by the size of the 
dividend.
This is not trie for integer multiplication, addition, or subtraction,

-- hendrik

> 
> - Lauri
> 
> _______________________________________________
> Libre-soc-dev mailing list
> Libre-soc-dev at lists.libre-soc.org
> http://lists.libre-soc.org/mailman/listinfo/libre-soc-dev