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

[Cole Poirier]

On Thu, Dec 10, 2020 at 3:17 PM Luke Kenneth Casson Leighton
<lkcl at lkcl.net> wrote:
>
> indeed.
>
> unfortunately we need to get a hell of a long way down the line
> (binutils and full simulator *and* implementations of each and every
> algorithm) before such assessments can be made.
>
> which is why i put in NLnet grant requests with rather large amounts
> covering the iterative redesign process that will be necessary.

Yes, it's quite a lot. I'm investigating the correctness of my mental
model of what each stage of this design process consists in, from spec
development through to iterative, comparative testing of proposed
opcodes and the speedup they provide for each algorithm. It's nice
(read essential) that we get to use data to guide our development
instead of using a more blind approach.

Cole