Re: [PATCH] RISC-V: Allow RVV VMS{Compare}(V1, V1) simplify to VMSET

[Andrew Waterman <andrew@sifive.com>]

[-- Attachment #1: Type: text/plain, Size: 1375 bytes --]

On Sat, Apr 29, 2023 at 8:06 AM Jeff Law via Gcc-patches <
gcc-patches@gcc.gnu.org> wrote:
>
>
>
> On 4/28/23 20:55, Li, Pan2 wrote:
> > Thanks Jeff for comments.
> >
> > It makes sense to me. For the EQ operator we should have CONSTM1.
> That's not the way I interpret the RVV documentation.  Of course it's
> not terribly clear.    I guess one could do some experiments with qemu
> or try to dig into the sail code and figure out the intent from those.
>
>
>
> Does this mean s390 parts has similar issue here? Then for instructions
> like VMSEQ, we need to adjust the simplify_rtx up to a point.
> You'd have to refer to the s390 instruction set reference to understand
> precisely how the vector compares work.
>
> But as it stands this really isn't a simplify-rtx question, but a
> question of the semantics of risc-v.   What happens with the high bits
> in the destination mask register is critical -- and if risc-v doesn't
> set them to all ones in this case, then that would mean that defining
> that macro is simply wrong for risc-v.

The relevant statement in the spec is that "the tail elements are always
updated with a tail-agnostic policy".  The vmset.m instruction will cause
mask register bits [0, vl-1] to be set to 1; elements [vl, VLMAX-1] will
either be undisturbed or set to 1, i.e., effectively unspecified.

>
> jeff