Re: [PATCH] Canonicalize (vec_duplicate (not A)) to (not (vec_duplicate A)).

[Segher Boessenkool <segher@kernel.crashing.org>]

On Thu, Jun 03, 2021 at 11:03:43AM +0000, Liu, Hongtao wrote:
> >A very typical example is how UMIN is optimised:
> >
> >   case UMIN:
> >      if (trueop1 == CONST0_RTX (mode) && ! side_effects_p (op0))
> >	return op1;
> >      if (rtx_equal_p (trueop0, trueop1) && ! side_effects_p (op0))
> >	return op0;
> >      tem = simplify_associative_operation (code, mode, op0, op1);
> >      if (tem)
> >	return tem;
> >      break;
> >
> >(the stuff using "tem").
> >
> >Hongtao, can we do something similar here?  Does that work well?  Please try
> >it out :-)
> 
> In simplify_rtx, no simplication occurs, there is just the difference between
>  (vec_duplicate (not REG)) and (not (vec_duplicate (REG)). So here tem will only be 0.

simplify-rtx is used by combine.  When you do and+not+splat for example
my suggestion should kick in.  Try it out, don't just dismiss it?

> Basically we don't know it's a simplication until combine successfully split the
> 3->2 instructions (not + broadcast + and to andnot + broadcast), but it's pretty awkward
> to do this in combine.

But you need to do this *before* it is split.  That is the whole point.

> Consider andnot is existed for many backends, I think a canonicalization is needed here.

Please do note that that is not as easy as yoou may think: you need to
make sure nothing ever creates non-canonical code.

> Maybe we can add insn canonicalization for transforming (and (vect_duplicate (not A)) B) to 
> (and (not (duplicate (not A)) B) instead of (vec_duplicate (not A)) to (not (vec_duplicate A))?

I don't understand what this means?


Segher