On Wed, Nov 19, 2025 at 1:31 PM Robin Dapp <[email protected]> wrote:
>
> > Yes, it's supposed to be a start.
> >
> > Btw, after Robin now implemented "type punning" for gather/stride loads
> > to handle power-of-two size groups load/store-lanes needs similar
> > support which would esp. help ARM where there's only up to ld4(?).
>
> From what I heard so far nobody really likes segmented loads/stores
> (=load/store lanes) on the riscv hardware side anyway.  On our design
> we have experimented with just replacing them with strided loads in
> the backend but that's not better either.  So, yes, it's on my list but
> the impact on performance is not clear to me yet.

I'm aware of multiple designs that implement the unit-stride segment
accesses with performance similar to unit-stride non-segment accesses,
and implement the strided segment accesses at performance similar to
strided non-segment accesses, provided the segments don't cross
microarchitecturally significant boundaries.  Like many of the more
complex features in RVV, I'd expect more implementations to optimize
these cases over time.

>
> --
> Regards
>  Robin
>

Reply via email to