Re: Re: [PATCH] Middle-end: Fix bug of induction variable vectorization for RVV

["juzhe.zhong@rivai.ai" <juzhe.zhong@rivai.ai>]

[-- Attachment #1: Type: text/plain, Size: 10393 bytes --]

Hi, Richard.

>> For nested_in_vect_loop we never have LOOP_VINFO_USING_SELECT_VL_P?
Could you give me an example of nested loop ?
For now, I can't produce a case.

Thanks a lot for the comments, I will try to refactor as you suggested.


juzhe.zhong@rivai.ai
 
From: Richard Biener
Date: 2023-11-10 17:46
To: 钟居哲
CC: richard.guenther; gcc-patches; richard.sandiford; kito.cheng; kito.cheng
Subject: Re: Re: [PATCH] Middle-end: Fix bug of induction variable vectorization for RVV
On Thu, 9 Nov 2023, ??? wrote:
 
> Hi, Richard.
> 
> >> I think it would be better to split out building a tree from VF from both
> >> arms and avoid using 'vf' when LOOP_VINFO_USING_SELECT_VL_P.
> 
> I am trying to split out building tree from both arms as you suggested..
> Could you take a look the following codes ?
> 
> diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
> index 8abc1937d74..24a86187d11 100644
> --- a/gcc/tree-vect-loop.cc
> +++ b/gcc/tree-vect-loop.cc
> @@ -10315,19 +10315,47 @@ vectorizable_induction (loop_vec_info loop_vinfo,
>        /* iv_loop is the loop to be vectorized. Generate:
>     vec_step = [VF*S, VF*S, VF*S, VF*S]  */
>        gimple_seq seq = NULL;
> -      if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (step_expr)))
> +      if (LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo))
>   {
> -   expr = build_int_cst (integer_type_node, vf);
> -   expr = gimple_build (&seq, FLOAT_EXPR, TREE_TYPE (step_expr), expr);
> +   /* When we're using loop_len produced by SELEC_VL, the non-final
> +      iterations are not always processing VF elements.  So vectorize
> +      induction variable instead of
> +
> +        _21 = vect_vec_iv_.6_22 + { VF, ... };
> +
> +      We should generate:
> +
> +       _35 = .SELECT_VL (ivtmp_33, VF);
> +       vect_cst__22 = [vec_duplicate_expr] _35;
> +       _21 = vect_vec_iv_.6_22 + vect_cst__22;  */
> +   vec_loop_lens *lens = &LOOP_VINFO_LENS (loop_vinfo);
> +   tree len
> +     = vect_get_loop_len (loop_vinfo, NULL, lens, 1, vectype, 0, 0);
> +   expr = force_gimple_operand (fold_convert (TREE_TYPE (step_expr),
> +      unshare_expr (len)),
> +        &seq, true, NULL_TREE);
>   }
>        else
> - expr = build_int_cst (TREE_TYPE (step_expr), vf);
> + {
> +   bool float_p = SCALAR_FLOAT_TYPE_P (TREE_TYPE (step_expr));
> +   expr = build_int_cst (float_p ? integer_type_node
> + : TREE_TYPE (step_expr),
> + vf);
> +   if (float_p)
> +     expr = gimple_build (&seq, FLOAT_EXPR, TREE_TYPE (step_expr), expr);
> + }
> +
 
I meant you keep the existing flow in the function, specifically
I think you should handle SCALAR_FLOAT_TYPE_P like it was previously
handled, just build 'vf' in the dynamic way.
 
>        new_name = gimple_build (&seq, MULT_EXPR, TREE_TYPE (step_expr),
>          expr, step_expr);
>        if (seq)
>   {
> -   new_bb = gsi_insert_seq_on_edge_immediate (pe, seq);
> -   gcc_assert (!new_bb);
> +   if (LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo))
> +     gsi_insert_seq_before (&si, seq, GSI_SAME_STMT);
> +   else
> +     {
> +       new_bb = gsi_insert_seq_on_edge_immediate (pe, seq);
> +       gcc_assert (!new_bb);
> +     }
>   }
>      }
>  
> @@ -10335,9 +10363,9 @@ vectorizable_induction (loop_vec_info loop_vinfo,
>    gcc_assert (CONSTANT_CLASS_P (new_name)
>         || TREE_CODE (new_name) == SSA_NAME);
>    new_vec = build_vector_from_val (step_vectype, t);
> -  vec_step = vect_init_vector (loop_vinfo, stmt_info,
> -        new_vec, step_vectype, NULL);
> -
> +  vec_step
> +    = vect_init_vector (loop_vinfo, stmt_info, new_vec, step_vectype,
> + LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo) ? &si : NULL);
 
again this makes the flow hard to follow.  I suppose refactoring this
overall to
 
    if (nested_in_vect_loop)
...
    else if (LOOP_VINFO_USING_SELECT_VL_P (..))
...
    else
...
 
and duplicate this tail into the cases makes it easier to follow.
 
For nested_in_vect_loop we never have LOOP_VINFO_USING_SELECT_VL_P?
 
Richard.
 
 
> Thanks.
> 
> 
> juzhe.zhong@rivai.ai
>  
> From: Richard Biener
> Date: 2023-11-09 20:16
> To: Juzhe-Zhong
> CC: gcc-patches; richard.sandiford; rguenther; kito.cheng; kito.cheng
> Subject: Re: [PATCH] Middle-end: Fix bug of induction variable vectorization for RVV
> On Wed, Nov 8, 2023 at 11:53?AM Juzhe-Zhong <juzhe.zhong@rivai.ai> wrote:
> >
> > PR: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112438
> >
> > SELECT_VL result is not necessary always VF in non-final iteration.
> >
> > Current GIMPLE IR is wrong:
> >
> > # vect_vec_iv_.21_25 = PHI <_24(4), { 0, 1, 2, ... }(3)>
> > ...
> > _24 = vect_vec_iv_.21_25 + { POLY_INT_CST [4, 4], ... };
> >
> > After this patch which is correct for SELECT_VL:
> >
> > # vect_vec_iv_.8_22 = PHI <_21(4), { 0, 1, 2, ... }(3)>
> > ...
> > _35 = .SELECT_VL (ivtmp_33, POLY_INT_CST [4, 4]);
> > _21 = vect_vec_iv_.8_22 + { POLY_INT_CST [4, 4], ... };
> >
> > kito, could you give more explanation ?
> >
> >         PR middle/112438
> >
> > gcc/ChangeLog:
> >
> >         * tree-vect-loop.cc (vectorizable_induction): Fix bug.
> >
> > gcc/testsuite/ChangeLog:
> >
> >         * gcc.target/riscv/rvv/autovec/pr112438.c: New test.
> >
> > ---
> >  .../gcc.target/riscv/rvv/autovec/pr112438.c   | 35 +++++++++++++++++
> >  gcc/tree-vect-loop.cc                         | 39 +++++++++++++++----
> >  2 files changed, 67 insertions(+), 7 deletions(-)
> >  create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/pr112438.c
> >
> > diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr112438.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr112438.c
> > new file mode 100644
> > index 00000000000..b326d56a52c
> > --- /dev/null
> > +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/pr112438.c
> > @@ -0,0 +1,35 @@
> > +/* { dg-do compile } */
> > +/* { dg-options "-march=rv64gcv -mabi=lp64d -O3 -fno-vect-cost-model -ffast-math -fdump-tree-optimized-details" } */
> > +
> > +void
> > +foo (int n, int *__restrict in, int *__restrict out)
> > +{
> > +  for (int i = 0; i < n; i += 1)
> > +    {
> > +      out[i] = in[i] + i;
> > +    }
> > +}
> > +
> > +void
> > +foo2 (int n, float * __restrict in,
> > +float * __restrict out)
> > +{
> > +  for (int i = 0; i < n; i += 1)
> > +    {
> > +      out[i] = in[i] + i;
> > +    }
> > +}
> > +
> > +void
> > +foo3 (int n, float * __restrict in,
> > +float * __restrict out, float x)
> > +{
> > +  for (int i = 0; i < n; i += 1)
> > +    {
> > +      out[i] = in[i] + i* i;
> > +    }
> > +}
> > +
> > +/* We don't want to see vect_vec_iv_.21_25 + { POLY_INT_CST [4, 4], ... }.  */
> > +/* { dg-final { scan-tree-dump-not "\\+ \{ POLY_INT_CST" "optimized" } } */
> > +
> > diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
> > index a544bc9b059..3e103946168 100644
> > --- a/gcc/tree-vect-loop.cc
> > +++ b/gcc/tree-vect-loop.cc
> > @@ -10309,10 +10309,30 @@ vectorizable_induction (loop_vec_info loop_vinfo,
> >      new_name = step_expr;
> >    else
> >      {
> > +      gimple_seq seq = NULL;
> > +      if (LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo))
> > +       {
> > +         /* When we're using loop_len produced by SELEC_VL, the non-final
> > +            iterations are not always processing VF elements.  So vectorize
> > +            induction variable instead of
> > +
> > +              _21 = vect_vec_iv_.6_22 + { VF, ... };
> > +
> > +            We should generate:
> > +
> > +              _35 = .SELECT_VL (ivtmp_33, VF);
> > +              vect_cst__22 = [vec_duplicate_expr] _35;
> > +              _21 = vect_vec_iv_.6_22 + vect_cst__22;  */
> > +         vec_loop_lens *lens = &LOOP_VINFO_LENS (loop_vinfo);
> > +         tree len
> > +           = vect_get_loop_len (loop_vinfo, NULL, lens, 1, vectype, 0, 0);
> > +         expr = force_gimple_operand (fold_convert (TREE_TYPE (step_expr),
> > +                                                    unshare_expr (len)),
> > +                                      &seq, true, NULL_TREE);
> > +       }
>  
> I think it would be better to split out building a tree from VF from both
> arms and avoid using 'vf' when LOOP_VINFO_USING_SELECT_VL_P.
>  
> Btw, you are not patching the SLP path here which I believe has the same
> problem but is currently exempt from non-constant VF at least.
>  
> Richard.
>  
> >        /* iv_loop is the loop to be vectorized. Generate:
> >           vec_step = [VF*S, VF*S, VF*S, VF*S]  */
> > -      gimple_seq seq = NULL;
> > -      if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (step_expr)))
> > +      else if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (step_expr)))
> >         {
> >           expr = build_int_cst (integer_type_node, vf);
> >           expr = gimple_build (&seq, FLOAT_EXPR, TREE_TYPE (step_expr), expr);
> > @@ -10323,8 +10343,13 @@ vectorizable_induction (loop_vec_info loop_vinfo,
> >                                expr, step_expr);
> >        if (seq)
> >         {
> > -         new_bb = gsi_insert_seq_on_edge_immediate (pe, seq);
> > -         gcc_assert (!new_bb);
> > +         if (LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo))
> > +           gsi_insert_seq_before (&si, seq, GSI_SAME_STMT);
> > +         else
> > +           {
> > +             new_bb = gsi_insert_seq_on_edge_immediate (pe, seq);
> > +             gcc_assert (!new_bb);
> > +           }
> >         }
> >      }
> >
> > @@ -10332,9 +10357,9 @@ vectorizable_induction (loop_vec_info loop_vinfo,
> >    gcc_assert (CONSTANT_CLASS_P (new_name)
> >               || TREE_CODE (new_name) == SSA_NAME);
> >    new_vec = build_vector_from_val (step_vectype, t);
> > -  vec_step = vect_init_vector (loop_vinfo, stmt_info,
> > -                              new_vec, step_vectype, NULL);
> > -
> > +  vec_step
> > +    = vect_init_vector (loop_vinfo, stmt_info, new_vec, step_vectype,
> > +                       LOOP_VINFO_USING_SELECT_VL_P (loop_vinfo) ? &si : NULL);
> >
> >    /* Create the following def-use cycle:
> >       loop prolog:
> > --
> > 2.36.3
> >
>  
> 
 
-- 
Richard Biener <rguenther@suse.de>
SUSE Software Solutions Germany GmbH,
Frankenstrasse 146, 90461 Nuernberg, Germany;
GF: Ivo Totev, Andrew McDonald, Werner Knoblich; (HRB 36809, AG Nuernberg)