[Libre-soc-dev] CR-result-driven predication
Luke Kenneth Casson Leighton
lkcl at lkcl.net
Sun Dec 13 02:19:27 GMT 2020
On 12/13/20, Luke Kenneth Casson Leighton <lkcl at lkcl.net> wrote:
> the hilarious thing is that all the pieces are already in place: CRs
> and predication. it is the mixing of these two that makes for some
> powerful savings.
btw jacob i feel it's important to mention this: both the 1<<r3 idea
and the above are really trivial enhancements that leverage existing
code / design.
combining CR result testing with predicate bits is trivial, as is
adding in a binary-to-unary encoder taking r3 and converting it to
single-bit mask form. a few lines of HDL.
whereas, the idea of modifying setvl, although it has merit worth
pursuing, would take several days to properly evaluate, and several
more to write up.
with time not being on our side we need to move with some haste onto
More information about the Libre-soc-dev