[libre-riscv-dev] Overall strategy RE transcendentals

[Luke Kenneth Casson Leighton]

On Tue, Sep 10, 2019 at 3:51 PM Hendrik Boom <hendrik at topoi.pooq.com> wrote:
>
> On Tue, Sep 10, 2019 at 04:28:45AM -0700, lkcl wrote:
>
> >
> > Thus the best that can be done is to use Quantitative Analysis to work
> > out which "subsets" - sub-Extensions - to include, and be as "inclusive"
> > as possible, and thus allow implementors to decide what to add to their
> > implementation, and how best to optimise them.
>
> Let's see if I understand.
>
> So there are multiple ways of implementing different instructions,
> ranging all the way from doing it entirely in software (which takes
> very little CPU space) through a series of operations (serial) within
> the CPU all the way to relly high speed parallelism.  Which of these
> are even feasible depends entirely on the operation itself.  Which are
> practical depends on time/area tradeoffs and those tradeoffs have to be
> decided based on the ultimate application domain.

yes.  different markets have different ideas on what constitutes "important"

> So we are trying to clarify all this by collecting and creating
> multiple implementations of useful operations so there is effectively a
> library of proven techniques that a CPU designer can pick and choose
> between.

not multiple implementations to be created (yet), just roughly noting
/ recording their existence, cost, capabilities, area, power usage,
flexibility

> Some of these we actually implement ourselves, some we gather from
> elsewhere and some we leave undone.

we implement the ones *we* want, and if people want a Reference Design
they pay us to add it or find it.

> And then finally (but in practice concurrently) we design a particular
> RISC-V design so that it can be realised in hardware.

alongside others also interested in this same area, because without
multiple implementors they can't really say "hey err we know we're
late to the party but we needed feature X and Y"

> This would be
> the practical proof-of-concept that gives others something to use, and
> shows still others that this project has produced useful components.

yep.

>
> -- hendrik
>
> Apologies for saying "we" and including myself, because I have done
> nothing tangible yet towards the project.

aside from ask really insightful questions, you mean.

> But I find it less estranging than saying "you".

:)