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

Cole Poirier colepoirier at gmail.com
Fri Dec 11 00:01:05 GMT 2020

On Thu, Dec 10, 2020 at 3:53 PM Luke Kenneth Casson Leighton
<lkcl at lkcl.net> wrote:
> indeed.
> the fundamental assumption is that we do not need to go the full route
> of hundreds of insanely specialised dedicated units.
> the more specialised the less flexibility.  the less flexibility the
> slimmer the chance they will be useful in the future.
> look at that 32 8888 to 16 1555 pixel instruction in VSX.  great in
> 1998, f*****g useless in 2021.

Right. Which comes back to the core of why we are using nmigen, being
able to change the quantity of each FU by simply changing one line of
a python dictionary instead of writing 1000s of lines of HDL everytime
we want to experiment with FU allocation. Right?


More information about the Libre-soc-dev mailing list