public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: "Li, Pan2" <pan2.li@intel.com>
To: Kito Cheng <kito.cheng@gmail.com>
Cc: "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>,
	"juzhe.zhong@rivai.ai" <juzhe.zhong@rivai.ai>,
	"Wang, Yanzhang" <yanzhang.wang@intel.com>
Subject: RE: [PATCH v1] RISC-V: Support rounding mode for VFMADD/VFMACC autovec
Date: Thu, 31 Aug 2023 13:29:47 +0000	[thread overview]
Message-ID: <MW5PR11MB590896EB5C5EF9DB797A008FA9E5A@MW5PR11MB5908.namprd11.prod.outlook.com> (raw)
In-Reply-To: <CA+yXCZAXgRXKpa_dT9-X=+kJ5A1S9G+1Z9x6B=iU-9giF0otXA@mail.gmail.com>

Committed, thanks Kito.

Pan

-----Original Message-----
From: Kito Cheng <kito.cheng@gmail.com> 
Sent: Thursday, August 31, 2023 9:10 PM
To: Li, Pan2 <pan2.li@intel.com>
Cc: gcc-patches@gcc.gnu.org; juzhe.zhong@rivai.ai; Wang, Yanzhang <yanzhang.wang@intel.com>
Subject: Re: [PATCH v1] RISC-V: Support rounding mode for VFMADD/VFMACC autovec

LGTM

On Thu, Aug 24, 2023 at 12:49 PM Pan Li via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> From: Pan Li <pan2.li@intel.com>
>
> There will be a case like below for intrinsic and autovec combination
>
> vfadd RTZ   <- intrinisc static rounding
> vfmadd      <- autovec/autovec-opt
>
> The autovec generated vfmadd should take DYN mode, and the
> frm must be restored before the vfmadd insn. This patch
> would like to fix this issue by:
>
> * Add the frm operand to the vfmadd/vfmacc autovec/autovec-opt pattern.
> * Set the frm_mode attr to DYN.
>
> Thus, the frm flow when combine autovec and intrinsic should be.
>
> +------------
> | frrm  a5
> | ...
> | fsrmi 4
> | vfadd       <- intrinsic static rounding.
> | ...
> | fsrm  a5
> | vfmadd      <- autovec/autovec-opt
> | ...
> +------------
>
> However, we leverage unspec instead of use to consume the FRM register
> because there are some restrictions from the combine pass. Some code
> path of try_combine may require the XVECLEN(pat, 0) == 2 for the
> recog_for_combine, and add new use will make the XVECLEN(pat, 0) == 3
> and result in the vfwmacc optimization failure. For example, in the
> test  widen-complicate-5.c and widen-8.c
>
> Finally, there will be other fma cases and they will be covered in
> the underlying patches.
>
> Signed-off-by: Pan Li <pan2.li@intel.com>
> Co-Authored-By: Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
>
> gcc/ChangeLog:
>
>         * config/riscv/autovec-opt.md: Add FRM_REGNUM to vfmadd/vfmacc.
>         * config/riscv/autovec.md: Ditto.
>         * config/riscv/vector-iterators.md: Add UNSPEC_VFFMA.
>
> gcc/testsuite/ChangeLog:
>
>         * gcc.target/riscv/rvv/base/float-point-frm-autovec-1.c: New test.
> ---
>  gcc/config/riscv/autovec-opt.md               | 32 ++++---
>  gcc/config/riscv/autovec.md                   | 26 +++---
>  gcc/config/riscv/vector-iterators.md          |  2 +
>  .../rvv/base/float-point-frm-autovec-1.c      | 88 +++++++++++++++++++
>  4 files changed, 125 insertions(+), 23 deletions(-)
>  create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/base/float-point-frm-autovec-1.c
>
> diff --git a/gcc/config/riscv/autovec-opt.md b/gcc/config/riscv/autovec-opt.md
> index 99b609a99d9..4b07e80ad95 100644
> --- a/gcc/config/riscv/autovec-opt.md
> +++ b/gcc/config/riscv/autovec-opt.md
> @@ -459,12 +459,14 @@ (define_insn_and_split "*pred_single_widen_mul<mode>"
>  ;; vect__13.182_33 = .FMA (vect__11.180_35, vect__8.176_40, vect__4.172_45);
>  (define_insn_and_split "*double_widen_fma<mode>"
>    [(set (match_operand:VWEXTF 0 "register_operand")
> -       (fma:VWEXTF
> -         (float_extend:VWEXTF
> -           (match_operand:<V_DOUBLE_TRUNC> 2 "register_operand"))
> -         (float_extend:VWEXTF
> -           (match_operand:<V_DOUBLE_TRUNC> 3 "register_operand"))
> -         (match_operand:VWEXTF 1 "register_operand")))]
> +       (unspec:VWEXTF
> +         [(fma:VWEXTF
> +           (float_extend:VWEXTF
> +             (match_operand:<V_DOUBLE_TRUNC> 2 "register_operand"))
> +           (float_extend:VWEXTF
> +             (match_operand:<V_DOUBLE_TRUNC> 3 "register_operand"))
> +           (match_operand:VWEXTF 1 "register_operand"))
> +          (reg:SI FRM_REGNUM)] UNSPEC_VFFMA))]
>    "TARGET_VECTOR && can_create_pseudo_p ()"
>    "#"
>    "&& 1"
> @@ -475,16 +477,19 @@ (define_insn_and_split "*double_widen_fma<mode>"
>      DONE;
>    }
>    [(set_attr "type" "vfwmuladd")
> -   (set_attr "mode" "<V_DOUBLE_TRUNC>")])
> +   (set_attr "mode" "<V_DOUBLE_TRUNC>")
> +   (set (attr "frm_mode") (symbol_ref "riscv_vector::FRM_DYN"))])
>
>  ;; This helps to match ext + fma.
>  (define_insn_and_split "*single_widen_fma<mode>"
>    [(set (match_operand:VWEXTF 0 "register_operand")
> -       (fma:VWEXTF
> -         (float_extend:VWEXTF
> -           (match_operand:<V_DOUBLE_TRUNC> 2 "register_operand"))
> -         (match_operand:VWEXTF 3 "register_operand")
> -         (match_operand:VWEXTF 1 "register_operand")))]
> +       (unspec:VWEXTF
> +         [(fma:VWEXTF
> +           (float_extend:VWEXTF
> +             (match_operand:<V_DOUBLE_TRUNC> 2 "register_operand"))
> +           (match_operand:VWEXTF 3 "register_operand")
> +           (match_operand:VWEXTF 1 "register_operand"))
> +          (reg:SI FRM_REGNUM)] UNSPEC_VFFMA))]
>    "TARGET_VECTOR && can_create_pseudo_p ()"
>    "#"
>    "&& 1"
> @@ -501,7 +506,8 @@ (define_insn_and_split "*single_widen_fma<mode>"
>      DONE;
>    }
>    [(set_attr "type" "vfwmuladd")
> -   (set_attr "mode" "<V_DOUBLE_TRUNC>")])
> +   (set_attr "mode" "<V_DOUBLE_TRUNC>")
> +   (set (attr "frm_mode") (symbol_ref "riscv_vector::FRM_DYN"))])
>
>  ;; -------------------------------------------------------------------------
>  ;; ---- [FP] VFWNMSAC
> diff --git a/gcc/config/riscv/autovec.md b/gcc/config/riscv/autovec.md
> index acca4c22b90..4894986d2a5 100644
> --- a/gcc/config/riscv/autovec.md
> +++ b/gcc/config/riscv/autovec.md
> @@ -1126,22 +1126,27 @@ (define_insn_and_split "*fnma<VI:mode><P:mode>"
>  (define_expand "fma<mode>4"
>    [(parallel
>      [(set (match_operand:VF 0 "register_operand")
> -         (fma:VF
> -           (match_operand:VF 1 "register_operand")
> -           (match_operand:VF 2 "register_operand")
> -           (match_operand:VF 3 "register_operand")))
> +         (unspec:VF
> +           [(fma:VF
> +             (match_operand:VF 1 "register_operand")
> +             (match_operand:VF 2 "register_operand")
> +             (match_operand:VF 3 "register_operand"))
> +            (reg:SI FRM_REGNUM)] UNSPEC_VFFMA))
>       (clobber (match_dup 4))])]
>    "TARGET_VECTOR"
>    {
>      operands[4] = gen_reg_rtx (Pmode);
> -  })
> +  }
> +  [(set (attr "frm_mode") (symbol_ref "riscv_vector::FRM_DYN"))])
>
>  (define_insn_and_split "*fma<VF:mode><P:mode>"
>    [(set (match_operand:VF 0 "register_operand"   "=vr, vr, ?&vr")
> -       (fma:VF
> -         (match_operand:VF 1 "register_operand" " %0, vr,   vr")
> -         (match_operand:VF 2 "register_operand" " vr, vr,   vr")
> -         (match_operand:VF 3 "register_operand" " vr,  0,   vr")))
> +       (unspec:VF
> +         [(fma:VF
> +           (match_operand:VF 1 "register_operand" " %0, vr,   vr")
> +           (match_operand:VF 2 "register_operand" " vr, vr,   vr")
> +           (match_operand:VF 3 "register_operand" " vr,  0,   vr"))
> +          (reg:SI FRM_REGNUM)] UNSPEC_VFFMA))
>     (clobber (match_operand:P 4 "register_operand" "=r,r,r"))]
>    "TARGET_VECTOR"
>    "#"
> @@ -1155,7 +1160,8 @@ (define_insn_and_split "*fma<VF:mode><P:mode>"
>      DONE;
>    }
>    [(set_attr "type" "vfmuladd")
> -   (set_attr "mode" "<VF:MODE>")])
> +   (set_attr "mode" "<VF:MODE>")
> +   (set (attr "frm_mode") (symbol_ref "riscv_vector::FRM_DYN"))])
>
>  ;; -------------------------------------------------------------------------
>  ;; ---- [FP] VFNMSAC and VFNMSUB
> diff --git a/gcc/config/riscv/vector-iterators.md b/gcc/config/riscv/vector-iterators.md
> index 4023a038fe9..9b2fb135bdd 100644
> --- a/gcc/config/riscv/vector-iterators.md
> +++ b/gcc/config/riscv/vector-iterators.md
> @@ -81,6 +81,8 @@ (define_c_enum "unspec" [
>    UNSPEC_VCOMPRESS
>    UNSPEC_VLEFF
>    UNSPEC_MODIFY_VL
> +
> +  UNSPEC_VFFMA
>  ])
>
>  (define_c_enum "unspecv" [
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-frm-autovec-1.c b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-frm-autovec-1.c
> new file mode 100644
> index 00000000000..f4f17a306d5
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-frm-autovec-1.c
> @@ -0,0 +1,88 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv --param=riscv-autovec-preference=fixed-vlmax -ffast-math -mabi=lp64 -O3 -Wno-psabi" } */
> +/* { dg-final { check-function-bodies "**" "" } } */
> +
> +#include "riscv_vector.h"
> +
> +/*
> +**test_1:
> +**     ...
> +**     frrm\t[axt][0-9]+
> +**     ...
> +**     fsrmi\t1
> +**     ...
> +**     vfsub\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrm\t[axt][0-9]+
> +**     ...
> +**     vfmadd\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     ret
> +*/
> +void
> +test_1 (vfloat32m1_t op1, vfloat32m1_t op2, vfloat32m1_t *op_out, size_t vl,
> +       double *in1, double *in2, double *out)
> +{
> +  *op_out = __riscv_vfsub_vv_f32m1_rm (op1, op2, 1, vl);
> +
> +  for (int i = 0; i < 4; ++i)
> +    out[i] += in1[i] * in2[i];
> +}
> +
> +/*
> +**test_2:
> +**     ...
> +**     frrm\t[axt][0-9]+
> +**     ...
> +**     fsrmi\t1
> +**     ...
> +**     vfsub\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrm\t[axt][0-9]+
> +**     ...
> +**     vfmadd\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrmi\t4
> +**     ...
> +**     vfsub\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrm\t[axt][0-9]+
> +**     ...
> +**     ret
> +*/
> +void
> +test_2 (vfloat32m1_t op1, vfloat32m1_t op2, vfloat32m1_t *op_out, size_t vl,
> +       double *in1, double *in2, double *out)
> +{
> +  op2 = __riscv_vfsub_vv_f32m1_rm (op1, op2, 1, vl);
> +
> +  for (int i = 0; i < 4; ++i)
> +    out[i] = out[i] * in1[i] + in2[i];
> +
> +  *op_out = __riscv_vfsub_vv_f32m1_rm (op1, op2, 4, vl);
> +}
> +
> +/*
> +**test_3:
> +**     ...
> +**     frrm\t[axt][0-9]+
> +**     ...
> +**     vfmadd\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrmi\t4
> +**     ...
> +**     vfsub\.vv\tv[0-9]+,v[0-9]+,v[0-9]+
> +**     ...
> +**     fsrm\t[axt][0-9]+
> +**     ...
> +**     ret
> +*/
> +void
> +test_3 (vfloat32m1_t op1, vfloat32m1_t op2, vfloat32m1_t *op_out, size_t vl,
> +       double *in1, double *in2, double *in3, double *out)
> +{
> +  for (int i = 0; i < 4; ++i)
> +    out[i] = in1[i] + in2[i] * out[i];
> +
> +  *op_out = __riscv_vfsub_vv_f32m1_rm (op1, op2, 4, vl);
> +}
> --
> 2.34.1
>

      reply	other threads:[~2023-08-31 13:30 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-24  4:49 pan2.li
2023-08-31 13:09 ` Kito Cheng
2023-08-31 13:29   ` Li, Pan2 [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=MW5PR11MB590896EB5C5EF9DB797A008FA9E5A@MW5PR11MB5908.namprd11.prod.outlook.com \
    --to=pan2.li@intel.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=juzhe.zhong@rivai.ai \
    --cc=kito.cheng@gmail.com \
    --cc=yanzhang.wang@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).