[Libre-soc-dev] Hardware-accelerated specialized instructions

Cole Poirier colepoirier at gmail.com
Thu Dec 10 23:09:57 GMT 2020


On Thursday, December 10, 2020, Luke Kenneth Casson Leighton <lkcl at lkcl.net>
wrote:

> On 12/10/20, Cole Poirier <colepoirier at gmail.com> wrote:
>
> > In the case of (1) and (2), the instruction may share FUs that have
> already
> > been allocated for existing instructions
>
> we have no idea which approach will work best until it is attempted.
> this is 6+ months away.
>
> at that time Jeff Bush's metric pixels/clock will be applied.
>
> if the number is too low, we evaluate and redesign.
>
> unless we have an expert who has done precisely this in exactly the
> same field and has significant world leading experience there is
> nothing that we can do except follow that approach.
>
> therefore guessing or speculating which might be better is of no
> value.  *we don't know*.
>

That is what was meant by the next sentence:
“This is where the simulation and analysis of resource contention from
simulating real-world use cases like H.264 or Vorbis encoding/decoding is
essential.”

Cole


More information about the Libre-soc-dev mailing list