Re: [aarch64] Use wzr/xzr for assigning vector element to 0

[Richard Sandiford <richard.sandiford@arm.com>]

Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org> writes:
> On Tue, 17 Jan 2023 at 18:29, Richard Sandiford
> <richard.sandiford@arm.com> wrote:
>>
>> Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org> writes:
>> > Hi Richard,
>> > For the following (contrived) test:
>> >
>> > void foo(int32x4_t v)
>> > {
>> >   v[3] = 0;
>> >   return v;
>> > }
>> >
>> > -O2 code-gen:
>> > foo:
>> >         fmov    s1, wzr
>> >         ins     v0.s[3], v1.s[0]
>> >         ret
>> >
>> > I suppose we can instead emit the following code-gen ?
>> > foo:
>> >      ins v0.s[3], wzr
>> >      ret
>> >
>> > combine produces:
>> > Failed to match this instruction:
>> > (set (reg:V4SI 95 [ v ])
>> >     (vec_merge:V4SI (const_vector:V4SI [
>> >                 (const_int 0 [0]) repeated x4
>> >             ])
>> >         (reg:V4SI 97)
>> >         (const_int 8 [0x8])))
>> >
>> > So, I wrote the following pattern to match the above insn:
>> > (define_insn "aarch64_simd_vec_set_zero<mode>"
>> >   [(set (match_operand:VALL_F16 0 "register_operand" "=w")
>> >         (vec_merge:VALL_F16
>> >             (match_operand:VALL_F16 1 "const_dup0_operand" "w")
>> >             (match_operand:VALL_F16 3 "register_operand" "0")
>> >             (match_operand:SI 2 "immediate_operand" "i")))]
>> >   "TARGET_SIMD"
>> >   {
>> >     int elt = ENDIAN_LANE_N (<nunits>, exact_log2 (INTVAL (operands[2])));
>> >     operands[2] = GEN_INT ((HOST_WIDE_INT) 1 << elt);
>> >     return "ins\\t%0.<Vetype>[%p2], wzr";
>> >   }
>> > )
>> >
>> > which now matches the above insn produced by combine.
>> > However, in reload dump, it creates a new insn for assigning
>> > register to (const_vector (const_int 0)),
>> > which results in:
>> > (insn 19 8 13 2 (set (reg:V4SI 33 v1 [99])
>> >         (const_vector:V4SI [
>> >                 (const_int 0 [0]) repeated x4
>> >             ])) "wzr-test.c":8:1 1269 {*aarch64_simd_movv4si}
>> >      (nil))
>> > (insn 13 19 14 2 (set (reg/i:V4SI 32 v0)
>> >         (vec_merge:V4SI (reg:V4SI 33 v1 [99])
>> >             (reg:V4SI 32 v0 [97])
>> >             (const_int 8 [0x8]))) "wzr-test.c":8:1 1808
>> > {aarch64_simd_vec_set_zerov4si}
>> >      (nil))
>> >
>> > and eventually the code-gen:
>> > foo:
>> >         movi    v1.4s, 0
>> >         ins     v0.s[3], wzr
>> >         ret
>> >
>> > To get rid of redundant assignment of 0 to v1, I tried to split the
>> > above pattern
>> > as in the attached patch. This works to emit code-gen:
>> > foo:
>> >         ins     v0.s[3], wzr
>> >         ret
>> >
>> > However, I am not sure if this is the right approach. Could you suggest,
>> > if it'd be possible to get rid of UNSPEC_SETZERO in the patch ?
>>
>> The problem is with the "w" constraint on operand 1, which tells LRA
>> to force the zero into an FPR.  It should work if you remove the
>> constraint.
> Ah indeed, sorry about that, changing the constrained works.

"i" isn't right though, because that's for scalar integers.
There's no need for any constraint here -- the predicate does
all of the work.

> Does the attached patch look OK after bootstrap+test ?
> Since we're in stage-4, shall it be OK to commit now, or queue it for stage-1 ?

It needs tests as well. :-)

Also:

> Thanks,
> Prathamesh
>
>
>>
>> Also, I think you'll need to use <vwcore>zr for the zero, so that
>> it uses xzr for 64-bit elements.
>>
>> I think this and the existing patterns ought to test
>> exact_log2 (INTVAL (operands[2])) >= 0 in the insn condition,
>> since there's no guarantee that RTL optimisations won't form
>> vec_merges that have other masks.
>>
>> Thanks,
>> Richard
>
> [aarch64] Use wzr/xzr for assigning 0 to vector element.
>
> gcc/ChangeLog:
> 	* config/aaarch64/aarch64-simd.md (aarch64_simd_vec_set_zero<mode>):
> 	New pattern.
> 	* config/aarch64/predicates.md (const_dup0_operand): New.
>
> diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
> index 104088f67d2..8e54ee4e886 100644
> --- a/gcc/config/aarch64/aarch64-simd.md
> +++ b/gcc/config/aarch64/aarch64-simd.md
> @@ -1083,6 +1083,20 @@
>    [(set_attr "type" "neon_ins<q>, neon_from_gp<q>, neon_load1_one_lane<q>")]
>  )
>  
> +(define_insn "aarch64_simd_vec_set_zero<mode>"
> +  [(set (match_operand:VALL_F16 0 "register_operand" "=w")
> +	(vec_merge:VALL_F16
> +	    (match_operand:VALL_F16 1 "const_dup0_operand" "i")
> +	    (match_operand:VALL_F16 3 "register_operand" "0")
> +	    (match_operand:SI 2 "immediate_operand" "i")))]
> +  "TARGET_SIMD && exact_log2 (INTVAL (operands[2])) >= 0"
> +  {
> +    int elt = ENDIAN_LANE_N (<nunits>, exact_log2 (INTVAL (operands[2])));
> +    operands[2] = GEN_INT ((HOST_WIDE_INT) 1 << elt);
> +    return "ins\\t%0.<Vetype>[%p2], <vwcore>zr";
> +  }
> +)
> +
>  (define_insn "@aarch64_simd_vec_copy_lane<mode>"
>    [(set (match_operand:VALL_F16 0 "register_operand" "=w")
>  	(vec_merge:VALL_F16
> diff --git a/gcc/config/aarch64/predicates.md b/gcc/config/aarch64/predicates.md
> index ff7f73d3f30..901fa1bd7f9 100644
> --- a/gcc/config/aarch64/predicates.md
> +++ b/gcc/config/aarch64/predicates.md
> @@ -49,6 +49,13 @@
>    return CONST_INT_P (op) && IN_RANGE (INTVAL (op), 1, 3);
>  })
>  
> +(define_predicate "const_dup0_operand"
> +  (match_code "const_vector")
> +{
> +  op = unwrap_const_vec_duplicate (op);
> +  return CONST_INT_P (op) && rtx_equal_p (op, const0_rtx);
> +})
> +

We already have aarch64_simd_imm_zero for this.  aarch64_simd_imm_zero
is actually more general, because it works for floating-point modes too.

I think the tests should cover all modes included in VALL_F16, since
that should have picked up this and the xzr thing.

Thanks,
Richard

>  (define_predicate "subreg_lowpart_operator"
>    (ior (match_code "truncate")
>         (and (match_code "subreg")