RISC-V: Add rounding mode operand for floating point instructions

  From: Juzhe-Zhong <juzhe.zhong@rivai.ai>

This patch is adding rounding mode operand and FRM_REGNUM dependency
into floating-point instructions.

The floating-point instructions we added FRM and rounding mode operand:
1. vfadd/vfsub
2. vfwadd/vfwsub
3. vfmul
4. vfdiv
5. vfwmul
6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
7. vfsqrt7/vfrec7
8. floating-point conversions.
9. floating-point reductions.

The floating-point instructions we did NOT add FRM and rounding mode operand:
1. vfsqrt/vfneg
2. vfmin/vfmax
3. comparisons
4. vfclass
5. vfsgnj/vfsgnjn/vfsgnjx
6. vfmerge
7. vfmv.v.f

TODO: floating-point ternary: FRM and rounding mode operand should be added but
they are not added in this patch since it will exceed the number of operands can
be handled in optabs.cc. Will add it the next patch.

gcc/ChangeLog:

        * config/riscv/riscv-protos.h (enum frm_field_enum): New enum.
        * config/riscv/riscv-vector-builtins.cc (function_expander::use_widen_ternop_insn): Add default rounding mode.
        * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add FRM_REGNUM.
        (riscv_hard_regno_mode_ok): Ditto.
        (riscv_conditional_register_usage): Ditto.
        * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
        (FRM_REG_P): Ditto.
        (RISCV_DWARF_FRM): Ditto.
        * config/riscv/riscv.md: Ditto.
        * config/riscv/vector-iterators.md: split smax/smin and plus/mult since smax/smin doesn't need FRM.
        * config/riscv/vector.md (@pred_<optab><mode>_scalar): Splitted pattern.
        (@pred_<optab><mode>): Ditto.

---
 gcc/config/riscv/riscv-protos.h           |  10 ++
 gcc/config/riscv/riscv-vector-builtins.cc |   7 +
 gcc/config/riscv/riscv.cc                 |   7 +-
 gcc/config/riscv/riscv.h                  |   7 +-
 gcc/config/riscv/riscv.md                 |   1 +
 gcc/config/riscv/vector-iterators.md      |   6 +-
 gcc/config/riscv/vector.md                | 171 ++++++++++++++++++----
 7 files changed, 171 insertions(+), 38 deletions(-)
  

LGTM

<juzhe.zhong@rivai.ai>於 2023年5月15日 週一，19:50寫道：

> From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
>
> This patch is adding rounding mode operand and FRM_REGNUM dependency
> into floating-point instructions.
>
> The floating-point instructions we added FRM and rounding mode operand:
> 1. vfadd/vfsub
> 2. vfwadd/vfwsub
> 3. vfmul
> 4. vfdiv
> 5. vfwmul
> 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> 7. vfsqrt7/vfrec7
> 8. floating-point conversions.
> 9. floating-point reductions.
>
> The floating-point instructions we did NOT add FRM and rounding mode
> operand:
> 1. vfsqrt/vfneg
> 2. vfmin/vfmax
> 3. comparisons
> 4. vfclass
> 5. vfsgnj/vfsgnjn/vfsgnjx
> 6. vfmerge
> 7. vfmv.v.f
>
> TODO: floating-point ternary: FRM and rounding mode operand should be
> added but
> they are not added in this patch since it will exceed the number of
> operands can
> be handled in optabs.cc. Will add it the next patch.
>
> gcc/ChangeLog:
>
>         * config/riscv/riscv-protos.h (enum frm_field_enum): New enum.
>         * config/riscv/riscv-vector-builtins.cc
> (function_expander::use_widen_ternop_insn): Add default rounding mode.
>         * config/riscv/riscv.cc (riscv_hard_regno_nregs): Add FRM_REGNUM.
>         (riscv_hard_regno_mode_ok): Ditto.
>         (riscv_conditional_register_usage): Ditto.
>         * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Ditto.
>         (FRM_REG_P): Ditto.
>         (RISCV_DWARF_FRM): Ditto.
>         * config/riscv/riscv.md: Ditto.
>         * config/riscv/vector-iterators.md: split smax/smin and plus/mult
> since smax/smin doesn't need FRM.
>         * config/riscv/vector.md (@pred_<optab><mode>_scalar): Splitted
> pattern.
>         (@pred_<optab><mode>): Ditto.
>
> ---
>  gcc/config/riscv/riscv-protos.h           |  10 ++
>  gcc/config/riscv/riscv-vector-builtins.cc |   7 +
>  gcc/config/riscv/riscv.cc                 |   7 +-
>  gcc/config/riscv/riscv.h                  |   7 +-
>  gcc/config/riscv/riscv.md                 |   1 +
>  gcc/config/riscv/vector-iterators.md      |   6 +-
>  gcc/config/riscv/vector.md                | 171 ++++++++++++++++++----
>  7 files changed, 171 insertions(+), 38 deletions(-)
>
> diff --git a/gcc/config/riscv/riscv-protos.h
> b/gcc/config/riscv/riscv-protos.h
> index 835bb802fc6..12634d0ac1a 100644
> --- a/gcc/config/riscv/riscv-protos.h
> +++ b/gcc/config/riscv/riscv-protos.h
> @@ -231,6 +231,16 @@ enum vxrm_field_enum
>    VXRM_RDN,
>    VXRM_ROD
>  };
> +/* Rounding mode bitfield for floating point FRM.  */
> +enum frm_field_enum
> +{
> +  FRM_RNE = 0b000,
> +  FRM_RTZ = 0b001,
> +  FRM_RDN = 0b010,
> +  FRM_RUP = 0b011,
> +  FRM_RMM = 0b100,
> +  DYN = 0b111
> +};
>  }
>
>  /* We classify builtin types into two classes:
> diff --git a/gcc/config/riscv/riscv-vector-builtins.cc
> b/gcc/config/riscv/riscv-vector-builtins.cc
> index 1de075fb90d..f10f38f6425 100644
> --- a/gcc/config/riscv/riscv-vector-builtins.cc
> +++ b/gcc/config/riscv/riscv-vector-builtins.cc
> @@ -3482,6 +3482,13 @@ function_expander::use_widen_ternop_insn (insn_code
> icode)
>    add_input_operand (Pmode, get_tail_policy_for_pred (pred));
>    add_input_operand (Pmode, get_mask_policy_for_pred (pred));
>    add_input_operand (Pmode, get_avl_type_rtx (avl_type::NONVLMAX));
> +
> +  /* TODO: Currently, we don't support intrinsic that is modeling
> rounding mode.
> +     We add default rounding mode for the intrinsics that didn't model
> rounding
> +     mode yet.  */
> +  if (opno != insn_data[icode].n_generator_args)
> +    add_input_operand (Pmode, const0_rtx);
> +
>    return generate_insn (icode);
>  }
>
> diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc
> index b52e613c629..de5b87b1a87 100644
> --- a/gcc/config/riscv/riscv.cc
> +++ b/gcc/config/riscv/riscv.cc
> @@ -6082,7 +6082,8 @@ riscv_hard_regno_nregs (unsigned int regno,
> machine_mode mode)
>
>    /* mode for VL or VTYPE are just a marker, not holding value,
>       so it always consume one register.  */
> -  if (VTYPE_REG_P (regno) || VL_REG_P (regno) || VXRM_REG_P (regno))
> +  if (VTYPE_REG_P (regno) || VL_REG_P (regno) || VXRM_REG_P (regno)
> +      || FRM_REG_P (regno))
>      return 1;
>
>    /* Assume every valid non-vector mode fits in one vector register.  */
> @@ -6150,7 +6151,8 @@ riscv_hard_regno_mode_ok (unsigned int regno,
> machine_mode mode)
>        if (lmul != 1)
>         return ((regno % lmul) == 0);
>      }
> -  else if (VTYPE_REG_P (regno) || VL_REG_P (regno) || VXRM_REG_P (regno))
> +  else if (VTYPE_REG_P (regno) || VL_REG_P (regno) || VXRM_REG_P (regno)
> +          || FRM_REG_P (regno))
>      return true;
>    else
>      return false;
> @@ -6587,6 +6589,7 @@ riscv_conditional_register_usage (void)
>        fixed_regs[VTYPE_REGNUM] = call_used_regs[VTYPE_REGNUM] = 1;
>        fixed_regs[VL_REGNUM] = call_used_regs[VL_REGNUM] = 1;
>        fixed_regs[VXRM_REGNUM] = call_used_regs[VXRM_REGNUM] = 1;
> +      fixed_regs[FRM_REGNUM] = call_used_regs[FRM_REGNUM] = 1;
>      }
>  }
>
> diff --git a/gcc/config/riscv/riscv.h b/gcc/config/riscv/riscv.h
> index f74b70de562..f55bd6112a8 100644
> --- a/gcc/config/riscv/riscv.h
> +++ b/gcc/config/riscv/riscv.h
> @@ -121,8 +121,9 @@ ASM_MISA_SPEC
>
>  /* The mapping from gcc register number to DWARF 2 CFA column number.  */
>  #define DWARF_FRAME_REGNUM(REGNO)
>       \
> -  (VXRM_REG_P (REGNO) ? RISCV_DWARF_VXRM
>      \
> -   : VL_REG_P (REGNO) ? RISCV_DWARF_VL
>      \
> +  (FRM_REG_P (REGNO)   ? RISCV_DWARF_FRM
>     \
> +   : VXRM_REG_P (REGNO) ? RISCV_DWARF_VXRM
>      \
> +   : VL_REG_P (REGNO)  ? RISCV_DWARF_VL
>      \
>     : VTYPE_REG_P (REGNO)
>      \
>       ? RISCV_DWARF_VTYPE
>      \
>       : (GP_REG_P (REGNO) || FP_REG_P (REGNO) || V_REG_P (REGNO)
>       \
> @@ -374,6 +375,7 @@ ASM_MISA_SPEC
>  #define VL_REG_P(REGNO) ((REGNO) == VL_REGNUM)
>  #define VTYPE_REG_P(REGNO) ((REGNO) == VTYPE_REGNUM)
>  #define VXRM_REG_P(REGNO) ((REGNO) == VXRM_REGNUM)
> +#define FRM_REG_P(REGNO) ((REGNO) == FRM_REGNUM)
>
>  /* True when REGNO is in SIBCALL_REGS set.  */
>  #define SIBCALL_REG_P(REGNO)   \
> @@ -392,6 +394,7 @@ ASM_MISA_SPEC
>  #define FRAME_POINTER_REGNUM 65
>
>  /* Define Dwarf for RVV.  */
> +#define RISCV_DWARF_FRM (4096 + 0x003)
>  #define RISCV_DWARF_VXRM (4096 + 0x00a)
>  #define RISCV_DWARF_VL (4096 + 0xc20)
>  #define RISCV_DWARF_VTYPE (4096 + 0xc21)
> diff --git a/gcc/config/riscv/riscv.md b/gcc/config/riscv/riscv.md
> index c5cf3af9868..91808d6bd2a 100644
> --- a/gcc/config/riscv/riscv.md
> +++ b/gcc/config/riscv/riscv.md
> @@ -136,6 +136,7 @@
>     (VL_REGNUM                  66)
>     (VTYPE_REGNUM               67)
>     (VXRM_REGNUM                        68)
> +   (FRM_REGNUM                 69)
>  ])
>
>  (include "predicates.md")
> diff --git a/gcc/config/riscv/vector-iterators.md
> b/gcc/config/riscv/vector-iterators.md
> index a282861335d..ac87ac3d114 100644
> --- a/gcc/config/riscv/vector-iterators.md
> +++ b/gcc/config/riscv/vector-iterators.md
> @@ -1436,8 +1436,10 @@
>
>  (define_code_iterator and_ior [and ior])
>
> -(define_code_iterator any_float_binop [plus mult smax smin minus div])
> -(define_code_iterator commutative_float_binop [plus mult smax smin])
> +(define_code_iterator any_float_binop [plus mult minus div])
> +(define_code_iterator any_float_binop_nofrm [smax smin])
> +(define_code_iterator commutative_float_binop [plus mult])
> +(define_code_iterator commutative_float_binop_nofrm [smax smin])
>  (define_code_iterator non_commutative_float_binop [minus div])
>  (define_code_iterator any_float_unop [neg abs sqrt])
>
> diff --git a/gcc/config/riscv/vector.md b/gcc/config/riscv/vector.md
> index e0aeeea57a4..0929d19d5ec 100644
> --- a/gcc/config/riscv/vector.md
> +++ b/gcc/config/riscv/vector.md
> @@ -5722,8 +5722,10 @@
>              (match_operand 6 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 7 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 9 "const_int_operand"        "  i,  i,  i,  i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_float_binop:VF
>             (match_operand:VF 3 "register_operand"       " vr, vr, vr, vr")
>             (match_operand:VF 4 "register_operand"       " vr, vr, vr,
> vr"))
> @@ -5733,7 +5735,7 @@
>    [(set_attr "type" "<float_insn_type>")
>     (set_attr "mode" "<MODE>")])
>
> -(define_insn "@pred_<optab><mode>_scalar"
> +(define_insn "@pred_<optab><mode>"
>    [(set (match_operand:VF 0 "register_operand"           "=vd, vd, vr,
> vr")
>         (if_then_else:VF
>           (unspec:<VM>
> @@ -5744,6 +5746,28 @@
>              (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
>              (reg:SI VL_REGNUM)
>              (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +         (any_float_binop_nofrm:VF
> +           (match_operand:VF 3 "register_operand"       " vr, vr, vr, vr")
> +           (match_operand:VF 4 "register_operand"       " vr, vr, vr,
> vr"))
> +         (match_operand:VF 2 "vector_merge_operand"     " vu,  0, vu,
> 0")))]
> +  "TARGET_VECTOR"
> +  "vf<insn>.vv\t%0,%3,%4%p1"
> +  [(set_attr "type" "<float_insn_type>")
> +   (set_attr "mode" "<MODE>")])
> +
> +(define_insn "@pred_<optab><mode>_scalar"
> +  [(set (match_operand:VF 0 "register_operand"           "=vd, vd, vr,
> vr")
> +       (if_then_else:VF
> +         (unspec:<VM>
> +           [(match_operand:<VM> 1 "vector_mask_operand" " vm, vm,Wc1,Wc1")
> +            (match_operand 5 "vector_length_operand"    " rK, rK, rK, rK")
> +            (match_operand 6 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 7 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 9 "const_int_operand"        "  i,  i,  i,  i")
> +            (reg:SI VL_REGNUM)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (commutative_float_binop:VF
>             (vec_duplicate:VF
>               (match_operand:<VEL> 4 "register_operand"  "  f,  f,  f,
> f"))
> @@ -5765,6 +5789,29 @@
>              (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
>              (reg:SI VL_REGNUM)
>              (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +         (commutative_float_binop_nofrm:VF
> +           (vec_duplicate:VF
> +             (match_operand:<VEL> 4 "register_operand"  "  f,  f,  f,
> f"))
> +           (match_operand:VF 3 "register_operand"       " vr, vr, vr,
> vr"))
> +         (match_operand:VF 2 "vector_merge_operand"     " vu,  0, vu,
> 0")))]
> +  "TARGET_VECTOR"
> +  "vf<insn>.vf\t%0,%3,%4%p1"
> +  [(set_attr "type" "<float_insn_type>")
> +   (set_attr "mode" "<MODE>")])
> +
> +(define_insn "@pred_<optab><mode>_scalar"
> +  [(set (match_operand:VF 0 "register_operand"           "=vd, vd, vr,
> vr")
> +       (if_then_else:VF
> +         (unspec:<VM>
> +           [(match_operand:<VM> 1 "vector_mask_operand" " vm, vm,Wc1,Wc1")
> +            (match_operand 5 "vector_length_operand"    " rK, rK, rK, rK")
> +            (match_operand 6 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 7 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 9 "const_int_operand"        "  i,  i,  i,  i")
> +            (reg:SI VL_REGNUM)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (non_commutative_float_binop:VF
>             (match_operand:VF 3 "register_operand"       " vr, vr, vr, vr")
>             (vec_duplicate:VF
> @@ -5784,8 +5831,10 @@
>              (match_operand 6 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 7 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 9 "const_int_operand"        "  i,  i,  i,  i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (non_commutative_float_binop:VF
>             (vec_duplicate:VF
>               (match_operand:<VEL> 4 "register_operand"  "  f,  f,  f,
> f"))
> @@ -6379,8 +6428,10 @@
>              (match_operand 5 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 6 "const_int_operand"        "  i,  i,  i,  i")
>              (match_operand 7 "const_int_operand"        "  i,  i,  i,  i")
> +            (match_operand 8 "const_int_operand"        "  i,  i,  i,  i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (unspec:VF
>             [(match_operand:VF 3 "register_operand"       " vr, vr, vr,
> vr")] VFMISC)
>           (match_operand:VF 2 "vector_merge_operand"     " vu,  0, vu,
> 0")))]
> @@ -6426,8 +6477,10 @@
>              (match_operand 6 "const_int_operand"                  "
> i,    i")
>              (match_operand 7 "const_int_operand"                  "
> i,    i")
>              (match_operand 8 "const_int_operand"                  "
> i,    i")
> +            (match_operand 9 "const_int_operand"                  "
> i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_widen_binop:VWEXTF
>             (float_extend:VWEXTF
>               (match_operand:<V_DOUBLE_TRUNC> 3 "register_operand" "
>  vr,   vr"))
> @@ -6448,8 +6501,10 @@
>              (match_operand 6 "const_int_operand"                  "
> i,    i")
>              (match_operand 7 "const_int_operand"                  "
> i,    i")
>              (match_operand 8 "const_int_operand"                  "
> i,    i")
> +            (match_operand 9 "const_int_operand"                  "
> i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_widen_binop:VWEXTF
>             (float_extend:VWEXTF
>               (match_operand:<V_DOUBLE_TRUNC> 3 "register_operand" "
>  vr,   vr"))
> @@ -6471,8 +6526,10 @@
>              (match_operand 6 "const_int_operand"                  "
> i,    i")
>              (match_operand 7 "const_int_operand"                  "
> i,    i")
>              (match_operand 8 "const_int_operand"                  "
> i,    i")
> +            (match_operand 9 "const_int_operand"                  "
> i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (match_operand:VWEXTF 3 "register_operand"             "
>  vr,   vr")
>             (float_extend:VWEXTF
> @@ -6492,8 +6549,10 @@
>              (match_operand 6 "const_int_operand"                  "
> i,    i")
>              (match_operand 7 "const_int_operand"                  "
> i,    i")
>              (match_operand 8 "const_int_operand"                  "
> i,    i")
> +            (match_operand 9 "const_int_operand"                  "
> i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (match_operand:VWEXTF 3 "register_operand"             "
>  vr,   vr")
>             (float_extend:VWEXTF
> @@ -6521,8 +6580,10 @@
>              (match_operand 6 "const_int_operand"                    "
> i")
>              (match_operand 7 "const_int_operand"                    "
> i")
>              (match_operand 8 "const_int_operand"                    "
> i")
> +            (match_operand 9 "const_int_operand"                    "
> i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (mult:VWEXTF
>               (float_extend:VWEXTF
> @@ -6545,8 +6606,10 @@
>              (match_operand 6 "const_int_operand"                    "
> i")
>              (match_operand 7 "const_int_operand"                    "
> i")
>              (match_operand 8 "const_int_operand"                    "
> i")
> +            (match_operand 9 "const_int_operand"                    "
> i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (mult:VWEXTF
>               (float_extend:VWEXTF
> @@ -6570,8 +6633,10 @@
>              (match_operand 6 "const_int_operand"                      "
>   i")
>              (match_operand 7 "const_int_operand"                      "
>   i")
>              (match_operand 8 "const_int_operand"                      "
>   i")
> +            (match_operand 9 "const_int_operand"                      "
>   i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (neg:VWEXTF
>               (mult:VWEXTF
> @@ -6595,8 +6660,10 @@
>              (match_operand 6 "const_int_operand"                      "
>   i")
>              (match_operand 7 "const_int_operand"                      "
>   i")
>              (match_operand 8 "const_int_operand"                      "
>   i")
> +            (match_operand 9 "const_int_operand"                      "
>   i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (plus_minus:VWEXTF
>             (neg:VWEXTF
>               (mult:VWEXTF
> @@ -6910,8 +6977,10 @@
>              (match_operand 5 "const_int_operand"            "  i,  i,
> i,  i")
>              (match_operand 6 "const_int_operand"            "  i,  i,
> i,  i")
>              (match_operand 7 "const_int_operand"            "  i,  i,
> i,  i")
> +            (match_operand 8 "const_int_operand"            "  i,  i,
> i,  i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (unspec:<VCONVERT>
>              [(match_operand:VF 3 "register_operand"         " vr, vr, vr,
> vr")] VFCVTS)
>           (match_operand:<VCONVERT> 2 "vector_merge_operand" " vu,  0,
> vu,  0")))]
> @@ -6929,8 +6998,10 @@
>              (match_operand 5 "const_int_operand"            "  i,  i,
> i,  i")
>              (match_operand 6 "const_int_operand"            "  i,  i,
> i,  i")
>              (match_operand 7 "const_int_operand"            "  i,  i,
> i,  i")
> +            (match_operand 8 "const_int_operand"            "  i,  i,
> i,  i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_fix:<VCONVERT>
>              (match_operand:VF 3 "register_operand"          " vr, vr, vr,
> vr"))
>           (match_operand:<VCONVERT> 2 "vector_merge_operand" " vu,  0,
> vu,  0")))]
> @@ -6948,8 +7019,10 @@
>              (match_operand 5 "const_int_operand"           "  i,  i,  i,
> i")
>              (match_operand 6 "const_int_operand"           "  i,  i,  i,
> i")
>              (match_operand 7 "const_int_operand"           "  i,  i,  i,
> i")
> +            (match_operand 8 "const_int_operand"           "  i,  i,  i,
> i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_float:VF
>              (match_operand:<VCONVERT> 3 "register_operand" " vr, vr, vr,
> vr"))
>           (match_operand:VF 2 "vector_merge_operand"        " vu,  0, vu,
> 0")))]
> @@ -6974,8 +7047,10 @@
>              (match_operand 5 "const_int_operand"             "    i,
> i")
>              (match_operand 6 "const_int_operand"             "    i,
> i")
>              (match_operand 7 "const_int_operand"             "    i,
> i")
> +            (match_operand 8 "const_int_operand"             "    i,
> i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (unspec:VWCONVERTI
>              [(match_operand:<VNCONVERT> 3 "register_operand" "   vr,
>  vr")] VFCVTS)
>           (match_operand:VWCONVERTI 2 "vector_merge_operand"  "   vu,
> 0")))]
> @@ -6993,8 +7068,10 @@
>              (match_operand 5 "const_int_operand"            "    i,    i")
>              (match_operand 6 "const_int_operand"            "    i,    i")
>              (match_operand 7 "const_int_operand"            "    i,    i")
> +            (match_operand 8 "const_int_operand"            "    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_fix:VWCONVERTI
>              (match_operand:<VNCONVERT> 3 "register_operand" "   vr,
>  vr"))
>           (match_operand:VWCONVERTI 2 "vector_merge_operand" "   vu,
> 0")))]
> @@ -7012,8 +7089,10 @@
>              (match_operand 5 "const_int_operand"            "    i,    i")
>              (match_operand 6 "const_int_operand"            "    i,    i")
>              (match_operand 7 "const_int_operand"            "    i,    i")
> +            (match_operand 8 "const_int_operand"            "    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_float:VF
>              (match_operand:<VNCONVERT> 3 "register_operand" "   vr,
>  vr"))
>           (match_operand:VF 2 "vector_merge_operand"         "   vu,
> 0")))]
> @@ -7031,8 +7110,10 @@
>              (match_operand 5 "const_int_operand"                 "    i,
>   i")
>              (match_operand 6 "const_int_operand"                 "    i,
>   i")
>              (match_operand 7 "const_int_operand"                 "    i,
>   i")
> +            (match_operand 8 "const_int_operand"                 "    i,
>   i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (float_extend:VWEXTF
>              (match_operand:<V_DOUBLE_TRUNC> 3 "register_operand" "   vr,
>  vr"))
>           (match_operand:VWEXTF 2 "vector_merge_operand"          "   vu,
>   0")))]
> @@ -7057,8 +7138,10 @@
>              (match_operand 5 "const_int_operand"              "  i,  i,
> i,  i,    i,    i")
>              (match_operand 6 "const_int_operand"              "  i,  i,
> i,  i,    i,    i")
>              (match_operand 7 "const_int_operand"              "  i,  i,
> i,  i,    i,    i")
> +            (match_operand 8 "const_int_operand"              "  i,  i,
> i,  i,    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (unspec:<VNCONVERT>
>              [(match_operand:VF 3 "register_operand"           "  0,  0,
> 0,  0,   vr,   vr")] VFCVTS)
>           (match_operand:<VNCONVERT> 2 "vector_merge_operand"  " vu,  0,
> vu,  0,   vu,    0")))]
> @@ -7076,8 +7159,10 @@
>              (match_operand 5 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (match_operand 6 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (match_operand 7 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
> +            (match_operand 8 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_fix:<VNCONVERT>
>              (match_operand:VF 3 "register_operand"           "  0,  0,
> 0,  0,   vr,   vr"))
>           (match_operand:<VNCONVERT> 2 "vector_merge_operand" " vu,  0,
> vu,  0,   vu,    0")))]
> @@ -7095,8 +7180,10 @@
>              (match_operand 5 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (match_operand 6 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (match_operand 7 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
> +            (match_operand 8 "const_int_operand"             "  i,  i,
> i,  i,    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (any_float:<VNCONVERT>
>              (match_operand:VWCONVERTI 3 "register_operand"   "  0,  0,
> 0,  0,   vr,   vr"))
>           (match_operand:<VNCONVERT> 2 "vector_merge_operand" " vu,  0,
> vu,  0,   vu,    0")))]
> @@ -7114,8 +7201,10 @@
>              (match_operand 5 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (match_operand 6 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (match_operand 7 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
> +            (match_operand 8 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (float_truncate:<V_DOUBLE_TRUNC>
>              (match_operand:VWEXTF 3 "register_operand"            "  0,
> 0,  0,  0,   vr,   vr"))
>           (match_operand:<V_DOUBLE_TRUNC> 2 "vector_merge_operand" " vu,
> 0, vu,  0,   vu,    0")))]
> @@ -7133,8 +7222,10 @@
>              (match_operand 5 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (match_operand 6 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (match_operand 7 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
> +            (match_operand 8 "const_int_operand"                  "  i,
> i,  i,  i,    i,    i")
>              (reg:SI VL_REGNUM)
> -            (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +            (reg:SI VTYPE_REGNUM)
> +            (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>           (unspec:<V_DOUBLE_TRUNC>
>             [(float_truncate:<V_DOUBLE_TRUNC>
>                (match_operand:VWEXTF 3 "register_operand"          "  0,
> 0,  0,  0,   vr,   vr"))] UNSPEC_ROD)
> @@ -7289,8 +7380,10 @@
>               (match_operand 5 "vector_length_operand"         "   rK,
>  rK")
>               (match_operand 6 "const_int_operand"             "    i,
> i")
>               (match_operand 7 "const_int_operand"             "    i,
> i")
> +             (match_operand 8 "const_int_operand"             "    i,
> i")
>               (reg:SI VL_REGNUM)
> -             (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +             (reg:SI VTYPE_REGNUM)
> +             (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>            (any_freduc:VF
>              (vec_duplicate:VF
>                (vec_select:<VEL>
> @@ -7311,8 +7404,10 @@
>               (match_operand 5 "vector_length_operand"              "
>  rK,   rK")
>               (match_operand 6 "const_int_operand"                  "
> i,    i")
>               (match_operand 7 "const_int_operand"                  "
> i,    i")
> +             (match_operand 8 "const_int_operand"                  "
> i,    i")
>               (reg:SI VL_REGNUM)
> -             (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +             (reg:SI VTYPE_REGNUM)
> +             (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>            (any_freduc:VF_ZVE64
>              (vec_duplicate:VF_ZVE64
>                (vec_select:<VEL>
> @@ -7333,8 +7428,10 @@
>               (match_operand 5 "vector_length_operand"              " rK,
> rK, rK, rK")
>               (match_operand 6 "const_int_operand"                  "  i,
> i,  i,  i")
>               (match_operand 7 "const_int_operand"                  "  i,
> i,  i,  i")
> +             (match_operand 8 "const_int_operand"                  "  i,
> i,  i,  i")
>               (reg:SI VL_REGNUM)
> -             (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +             (reg:SI VTYPE_REGNUM)
> +             (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>            (any_freduc:VF_ZVE32
>              (vec_duplicate:VF_ZVE32
>                (vec_select:<VEL>
> @@ -7356,8 +7453,10 @@
>                 (match_operand 5 "vector_length_operand"         "   rK,
>  rK")
>                 (match_operand 6 "const_int_operand"             "    i,
>   i")
>                 (match_operand 7 "const_int_operand"             "    i,
>   i")
> +               (match_operand 8 "const_int_operand"             "    i,
>   i")
>                 (reg:SI VL_REGNUM)
> -               (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +               (reg:SI VTYPE_REGNUM)
> +               (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>              (plus:VF
>                (vec_duplicate:VF
>                  (vec_select:<VEL>
> @@ -7379,8 +7478,10 @@
>                 (match_operand 5 "vector_length_operand"              "
>  rK,   rK")
>                 (match_operand 6 "const_int_operand"                  "
> i,    i")
>                 (match_operand 7 "const_int_operand"                  "
> i,    i")
> +               (match_operand 8 "const_int_operand"                  "
> i,    i")
>                 (reg:SI VL_REGNUM)
> -               (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +               (reg:SI VTYPE_REGNUM)
> +               (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>              (plus:VF_ZVE64
>                (vec_duplicate:VF_ZVE64
>                  (vec_select:<VEL>
> @@ -7402,8 +7503,10 @@
>                 (match_operand 5 "vector_length_operand"              "
> rK, rK, rK, rK")
>                 (match_operand 6 "const_int_operand"                  "
> i,  i,  i,  i")
>                 (match_operand 7 "const_int_operand"                  "
> i,  i,  i,  i")
> +               (match_operand 8 "const_int_operand"                  "
> i,  i,  i,  i")
>                 (reg:SI VL_REGNUM)
> -               (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +               (reg:SI VTYPE_REGNUM)
> +               (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>              (plus:VF_ZVE32
>                (vec_duplicate:VF_ZVE32
>                  (vec_select:<VEL>
> @@ -7425,8 +7528,10 @@
>                 (match_operand 5 "vector_length_operand"         "   rK,
>  rK")
>                 (match_operand 6 "const_int_operand"             "    i,
>   i")
>                 (match_operand 7 "const_int_operand"             "    i,
>   i")
> +               (match_operand 8 "const_int_operand"             "    i,
>   i")
>                 (reg:SI VL_REGNUM)
> -               (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +               (reg:SI VTYPE_REGNUM)
> +               (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>              (match_operand:VWF 3 "register_operand"             "   vr,
>  vr")
>              (match_operand:<VWLMUL1> 4 "register_operand"       "   vr,
>  vr")
>              (match_operand:<VWLMUL1> 2 "vector_merge_operand"   "   vu,
>   0")] UNSPEC_WREDUC_SUM)] ORDER))]
> @@ -7444,8 +7549,10 @@
>                 (match_operand 5 "vector_length_operand"         "   rK,
>  rK")
>                 (match_operand 6 "const_int_operand"             "    i,
>   i")
>                 (match_operand 7 "const_int_operand"             "    i,
>   i")
> +               (match_operand 8 "const_int_operand"             "    i,
>   i")
>                 (reg:SI VL_REGNUM)
> -               (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> +               (reg:SI VTYPE_REGNUM)
> +               (reg:SI FRM_REGNUM)] UNSPEC_VPREDICATE)
>              (match_operand:VWF_ZVE64 3 "register_operand"             "
>  vr,   vr")
>              (match_operand:<VWLMUL1_ZVE64> 4 "register_operand"       "
>  vr,   vr")
>              (match_operand:<VWLMUL1_ZVE64> 2 "vector_merge_operand"   "
>  vu,    0")] UNSPEC_WREDUC_SUM)] ORDER))]
> --
> 2.36.1
>
>
  

On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> 
> This patch is adding rounding mode operand and FRM_REGNUM dependency
> into floating-point instructions.
> 
> The floating-point instructions we added FRM and rounding mode operand:
> 1. vfadd/vfsub
> 2. vfwadd/vfwsub
> 3. vfmul
> 4. vfdiv
> 5. vfwmul
> 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> 7. vfsqrt7/vfrec7
> 8. floating-point conversions.
> 9. floating-point reductions.
> 
> The floating-point instructions we did NOT add FRM and rounding mode operand:
> 1. vfsqrt/vfneg
Assuming vfsqrt is actually an estimator the best place to handle 
rounding modes is at the last step(s) after N-R or Goldschmidt 
refinement steps.  I haven't paid too much attention to FP yet, but this 
is an area I've got fairly extensive experience.

Sadly RISC-V's estimator is fairly poor and the single instance FMACs 
are going to result in an implementation that may not actually be any 
better than what glibc can do.

Jeff
  

On 5/15/23 07:44, Kito Cheng wrote:
> LGTM
Agreed.
jeff
  

I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
not add frm into vsqrt.v. Maybe kito knows the reason ?

https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226 




juzhe.zhong@rivai.ai
 
From: Jeff Law
Date: 2023-05-15 21:52
To: juzhe.zhong; gcc-patches
CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
 
 
On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> 
> This patch is adding rounding mode operand and FRM_REGNUM dependency
> into floating-point instructions.
> 
> The floating-point instructions we added FRM and rounding mode operand:
> 1. vfadd/vfsub
> 2. vfwadd/vfwsub
> 3. vfmul
> 4. vfdiv
> 5. vfwmul
> 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> 7. vfsqrt7/vfrec7
> 8. floating-point conversions.
> 9. floating-point reductions.
> 
> The floating-point instructions we did NOT add FRM and rounding mode operand:
> 1. vfsqrt/vfneg
Assuming vfsqrt is actually an estimator the best place to handle 
rounding modes is at the last step(s) after N-R or Goldschmidt 
refinement steps.  I haven't paid too much attention to FP yet, but this 
is an area I've got fairly extensive experience.
 
Sadly RISC-V's estimator is fairly poor and the single instance FMACs 
are going to result in an implementation that may not actually be any 
better than what glibc can do.
 
Jeff
  

Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
checked spike that match that.

On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> not add frm into vsqrt.v. Maybe kito knows the reason ?
>
> https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
>
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Jeff Law
> Date: 2023-05-15 21:52
> To: juzhe.zhong; gcc-patches
> CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
>
>
> On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> >
> > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > into floating-point instructions.
> >
> > The floating-point instructions we added FRM and rounding mode operand:
> > 1. vfadd/vfsub
> > 2. vfwadd/vfwsub
> > 3. vfmul
> > 4. vfdiv
> > 5. vfwmul
> > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > 7. vfsqrt7/vfrec7
> > 8. floating-point conversions.
> > 9. floating-point reductions.
> >
> > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > 1. vfsqrt/vfneg
> Assuming vfsqrt is actually an estimator the best place to handle
> rounding modes is at the last step(s) after N-R or Goldschmidt
> refinement steps.  I haven't paid too much attention to FP yet, but this
> is an area I've got fairly extensive experience.
>
> Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> are going to result in an implementation that may not actually be any
> better than what glibc can do.
>
> Jeff
>
  

So, you mean I also need to add frm into vsqrt? 
If yes, I am now send another patch to add it.



juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:07
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
checked spike that match that.
 
On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> not add frm into vsqrt.v. Maybe kito knows the reason ?
>
> https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
>
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Jeff Law
> Date: 2023-05-15 21:52
> To: juzhe.zhong; gcc-patches
> CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
>
>
> On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> >
> > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > into floating-point instructions.
> >
> > The floating-point instructions we added FRM and rounding mode operand:
> > 1. vfadd/vfsub
> > 2. vfwadd/vfwsub
> > 3. vfmul
> > 4. vfdiv
> > 5. vfwmul
> > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > 7. vfsqrt7/vfrec7
> > 8. floating-point conversions.
> > 9. floating-point reductions.
> >
> > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > 1. vfsqrt/vfneg
> Assuming vfsqrt is actually an estimator the best place to handle
> rounding modes is at the last step(s) after N-R or Goldschmidt
> refinement steps.  I haven't paid too much attention to FP yet, but this
> is an area I've got fairly extensive experience.
>
> Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> are going to result in an implementation that may not actually be any
> better than what glibc can do.
>
> Jeff
>
  

Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
Is that rigth? If yes, I am gonna send a patch to fix it immediately.



juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:07
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
checked spike that match that.
 
On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> not add frm into vsqrt.v. Maybe kito knows the reason ?
>
> https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
>
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Jeff Law
> Date: 2023-05-15 21:52
> To: juzhe.zhong; gcc-patches
> CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
>
>
> On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> >
> > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > into floating-point instructions.
> >
> > The floating-point instructions we added FRM and rounding mode operand:
> > 1. vfadd/vfsub
> > 2. vfwadd/vfwsub
> > 3. vfmul
> > 4. vfdiv
> > 5. vfwmul
> > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > 7. vfsqrt7/vfrec7
> > 8. floating-point conversions.
> > 9. floating-point reductions.
> >
> > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > 1. vfsqrt/vfneg
> Assuming vfsqrt is actually an estimator the best place to handle
> rounding modes is at the last step(s) after N-R or Goldschmidt
> refinement steps.  I haven't paid too much attention to FP yet, but this
> is an area I've got fairly extensive experience.
>
> Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> are going to result in an implementation that may not actually be any
> better than what glibc can do.
>
> Jeff
>
  

> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.

Yes, and I also double checked spike implementation :P

and it seems like you're not committed yet, so let's send V2 :)

On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:07
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> checked spike that match that.
>
> On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > not add frm into vsqrt.v. Maybe kito knows the reason ?
> >
> > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> >
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Jeff Law
> > Date: 2023-05-15 21:52
> > To: juzhe.zhong; gcc-patches
> > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> >
> >
> > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > >
> > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > into floating-point instructions.
> > >
> > > The floating-point instructions we added FRM and rounding mode operand:
> > > 1. vfadd/vfsub
> > > 2. vfwadd/vfwsub
> > > 3. vfmul
> > > 4. vfdiv
> > > 5. vfwmul
> > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > 7. vfsqrt7/vfrec7
> > > 8. floating-point conversions.
> > > 9. floating-point reductions.
> > >
> > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > 1. vfsqrt/vfneg
> > Assuming vfsqrt is actually an estimator the best place to handle
> > rounding modes is at the last step(s) after N-R or Goldschmidt
> > refinement steps.  I haven't paid too much attention to FP yet, but this
> > is an area I've got fairly extensive experience.
> >
> > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > are going to result in an implementation that may not actually be any
> > better than what glibc can do.
> >
> > Jeff
> >
>
  

What about vfsnjn ? Do they have FRM ? I want to double check it since I don't trust document.



juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:22
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
 
Yes, and I also double checked spike implementation :P
 
and it seems like you're not committed yet, so let's send V2 :)
 
On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:07
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> checked spike that match that.
>
> On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > not add frm into vsqrt.v. Maybe kito knows the reason ?
> >
> > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> >
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Jeff Law
> > Date: 2023-05-15 21:52
> > To: juzhe.zhong; gcc-patches
> > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> >
> >
> > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > >
> > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > into floating-point instructions.
> > >
> > > The floating-point instructions we added FRM and rounding mode operand:
> > > 1. vfadd/vfsub
> > > 2. vfwadd/vfwsub
> > > 3. vfmul
> > > 4. vfdiv
> > > 5. vfwmul
> > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > 7. vfsqrt7/vfrec7
> > > 8. floating-point conversions.
> > > 9. floating-point reductions.
> > >
> > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > 1. vfsqrt/vfneg
> > Assuming vfsqrt is actually an estimator the best place to handle
> > rounding modes is at the last step(s) after N-R or Goldschmidt
> > refinement steps.  I haven't paid too much attention to FP yet, but this
> > is an area I've got fairly extensive experience.
> >
> > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > are going to result in an implementation that may not actually be any
> > better than what glibc can do.
> >
> > Jeff
> >
>
  

The reason I ask vfsgnjn since according to RVV ISA:
vfneg.v vd,vs = vfsgnjn.vv vd,vs,vs.

It's really confusing here that document has FRM in vfneg but no FRM in vfsgnjn ?
It's really odd here.


juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:22
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
 
Yes, and I also double checked spike implementation :P
 
and it seems like you're not committed yet, so let's send V2 :)
 
On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:07
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> checked spike that match that.
>
> On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > not add frm into vsqrt.v. Maybe kito knows the reason ?
> >
> > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> >
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Jeff Law
> > Date: 2023-05-15 21:52
> > To: juzhe.zhong; gcc-patches
> > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> >
> >
> > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > >
> > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > into floating-point instructions.
> > >
> > > The floating-point instructions we added FRM and rounding mode operand:
> > > 1. vfadd/vfsub
> > > 2. vfwadd/vfwsub
> > > 3. vfmul
> > > 4. vfdiv
> > > 5. vfwmul
> > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > 7. vfsqrt7/vfrec7
> > > 8. floating-point conversions.
> > > 9. floating-point reductions.
> > >
> > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > 1. vfsqrt/vfneg
> > Assuming vfsqrt is actually an estimator the best place to handle
> > rounding modes is at the last step(s) after N-R or Goldschmidt
> > refinement steps.  I haven't paid too much attention to FP yet, but this
> > is an area I've got fairly extensive experience.
> >
> > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > are going to result in an implementation that may not actually be any
> > better than what glibc can do.
> >
> > Jeff
> >
>
  

And what about vfabs ? I guess it also need FRM ?
vfneg/vfabs/vfsgnj/vfsgnj/vfsgnjx
vfneg.v vd,vs = vfsgnjn.vv vd,vs,vs
vfabs.v vd,vs = vfsgnjx.vv vd,vs,vs

That's all questions I have, plz double check for me.
Thanks.


juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:22
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
 
Yes, and I also double checked spike implementation :P
 
and it seems like you're not committed yet, so let's send V2 :)
 
On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:07
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> checked spike that match that.
>
> On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > not add frm into vsqrt.v. Maybe kito knows the reason ?
> >
> > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> >
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Jeff Law
> > Date: 2023-05-15 21:52
> > To: juzhe.zhong; gcc-patches
> > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> >
> >
> > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > >
> > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > into floating-point instructions.
> > >
> > > The floating-point instructions we added FRM and rounding mode operand:
> > > 1. vfadd/vfsub
> > > 2. vfwadd/vfwsub
> > > 3. vfmul
> > > 4. vfdiv
> > > 5. vfwmul
> > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > 7. vfsqrt7/vfrec7
> > > 8. floating-point conversions.
> > > 9. floating-point reductions.
> > >
> > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > 1. vfsqrt/vfneg
> > Assuming vfsqrt is actually an estimator the best place to handle
> > rounding modes is at the last step(s) after N-R or Goldschmidt
> > refinement steps.  I haven't paid too much attention to FP yet, but this
> > is an area I've got fairly extensive experience.
> >
> > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > are going to result in an implementation that may not actually be any
> > better than what glibc can do.
> >
> > Jeff
> >
>
  

all sign injection operations (vfsgnjn/ vfsgnj/vfsgnjx and its
friends) didn't involve rounding in the operation, so vfneg.v and
vfabs.v don't need FRM.

On Mon, May 15, 2023 at 10:38 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> And what about vfabs ? I guess it also need FRM ?
> vfneg/vfabs/vfsgnj/vfsgnj/vfsgnjx
> vfneg.v vd,vs = vfsgnjn.vv vd,vs,vs
> vfabs.v vd,vs = vfsgnjx.vv vd,vs,vs
>
> That's all questions I have, plz double check for me.
> Thanks.
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:22
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> > Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
> Yes, and I also double checked spike implementation :P
>
> and it seems like you're not committed yet, so let's send V2 :)
>
> On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> > Is that rigth? If yes, I am gonna send a patch to fix it immediately.
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Kito Cheng
> > Date: 2023-05-15 22:07
> > To: 钟居哲
> > CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> > checked spike that match that.
> >
> > On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> > >
> > > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > > not add frm into vsqrt.v. Maybe kito knows the reason ?
> > >
> > > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> > >
> > >
> > >
> > >
> > > juzhe.zhong@rivai.ai
> > >
> > > From: Jeff Law
> > > Date: 2023-05-15 21:52
> > > To: juzhe.zhong; gcc-patches
> > > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > >
> > >
> > > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > > >
> > > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > > into floating-point instructions.
> > > >
> > > > The floating-point instructions we added FRM and rounding mode operand:
> > > > 1. vfadd/vfsub
> > > > 2. vfwadd/vfwsub
> > > > 3. vfmul
> > > > 4. vfdiv
> > > > 5. vfwmul
> > > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > > 7. vfsqrt7/vfrec7
> > > > 8. floating-point conversions.
> > > > 9. floating-point reductions.
> > > >
> > > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > > 1. vfsqrt/vfneg
> > > Assuming vfsqrt is actually an estimator the best place to handle
> > > rounding modes is at the last step(s) after N-R or Goldschmidt
> > > refinement steps.  I haven't paid too much attention to FP yet, but this
> > > is an area I've got fairly extensive experience.
> > >
> > > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > > are going to result in an implementation that may not actually be any
> > > better than what glibc can do.
> > >
> > > Jeff
> > >
> >
>
  

Thanks.
https://gcc.gnu.org/pipermail/gcc-patches/2023-May/618614.html 
here is the V2 patch.
I have description about instructions are adding FRM or not.
Would you mind check it again now?



juzhe.zhong@rivai.ai
 
From: Kito Cheng
Date: 2023-05-15 22:41
To: 钟居哲
CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
all sign injection operations (vfsgnjn/ vfsgnj/vfsgnjx and its
friends) didn't involve rounding in the operation, so vfneg.v and
vfabs.v don't need FRM.
 
On Mon, May 15, 2023 at 10:38 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
>
> And what about vfabs ? I guess it also need FRM ?
> vfneg/vfabs/vfsgnj/vfsgnj/vfsgnjx
> vfneg.v vd,vs = vfsgnjn.vv vd,vs,vs
> vfabs.v vd,vs = vfsgnjx.vv vd,vs,vs
>
> That's all questions I have, plz double check for me.
> Thanks.
>
>
> juzhe.zhong@rivai.ai
>
> From: Kito Cheng
> Date: 2023-05-15 22:22
> To: 钟居哲
> CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> > Is that rigth? If yes, I am gonna send a patch to fix it immediately.
>
> Yes, and I also double checked spike implementation :P
>
> and it seems like you're not committed yet, so let's send V2 :)
>
> On Mon, May 15, 2023 at 10:12 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> >
> > Oh, do you mean vfsqrt7/vfrec7 doesn't have frm, but vfsqrt/vfneg should have frm.
> > Is that rigth? If yes, I am gonna send a patch to fix it immediately.
> >
> >
> >
> > juzhe.zhong@rivai.ai
> >
> > From: Kito Cheng
> > Date: 2023-05-15 22:07
> > To: 钟居哲
> > CC: Jeff Law; gcc-patches; kito.cheng; palmer; palmer; rdapp.gcc
> > Subject: Re: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > Oh, Craig says vfrsqrt7.v not have frm but vsqrt.v have frm, and
> > checked spike that match that.
> >
> > On Mon, May 15, 2023 at 9:55 PM 钟居哲 <juzhe.zhong@rivai.ai> wrote:
> > >
> > > I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM maintainer) said we should
> > > not add frm into vsqrt.v. Maybe kito knows the reason ?
> > >
> > > https://github.com/riscv-non-isa/rvv-intrinsic-doc/pull/226
> > >
> > >
> > >
> > >
> > > juzhe.zhong@rivai.ai
> > >
> > > From: Jeff Law
> > > Date: 2023-05-15 21:52
> > > To: juzhe.zhong; gcc-patches
> > > CC: kito.cheng; kito.cheng; palmer; palmer; rdapp.gcc
> > > Subject: Re: [PATCH] RISC-V: Add rounding mode operand for floating point instructions
> > >
> > >
> > > On 5/15/23 05:49, juzhe.zhong@rivai.ai wrote:
> > > > From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> > > >
> > > > This patch is adding rounding mode operand and FRM_REGNUM dependency
> > > > into floating-point instructions.
> > > >
> > > > The floating-point instructions we added FRM and rounding mode operand:
> > > > 1. vfadd/vfsub
> > > > 2. vfwadd/vfwsub
> > > > 3. vfmul
> > > > 4. vfdiv
> > > > 5. vfwmul
> > > > 6. vfwmacc/vfwnmacc/vfwmsac/vfwnmsac
> > > > 7. vfsqrt7/vfrec7
> > > > 8. floating-point conversions.
> > > > 9. floating-point reductions.
> > > >
> > > > The floating-point instructions we did NOT add FRM and rounding mode operand:
> > > > 1. vfsqrt/vfneg
> > > Assuming vfsqrt is actually an estimator the best place to handle
> > > rounding modes is at the last step(s) after N-R or Goldschmidt
> > > refinement steps.  I haven't paid too much attention to FP yet, but this
> > > is an area I've got fairly extensive experience.
> > >
> > > Sadly RISC-V's estimator is fairly poor and the single instance FMACs
> > > are going to result in an implementation that may not actually be any
> > > better than what glibc can do.
> > >
> > > Jeff
> > >
> >
>
  

On 5/15/23 07:54, 钟居哲 wrote:
> I don't know why we should not add frm vfsqrt.v since I saw topper (LLVM 
> maintainer) said we should
> not add frm into vsqrt.v. Maybe kito knows the reason ?
I'm pretty sure this is referring to the estimator.   The documentation 
is very clear that the sqrt estimator is independent of the rounding mode.

While it's not as explicit in the RV manual, a real sqrt instruction 
must round to be usable in an IEEE compliant way.  If it didn't honor 
rounding modes we would largely be unable to use the [v]fsqrt instructions.

Jeff