RE: [PATCH 17/22] arm: [MVE intrinsics] rework vcreateq

[Kyrylo Tkachov <Kyrylo.Tkachov@arm.com>]

> -----Original Message-----
> From: Christophe Lyon <christophe.lyon@arm.com>
> Sent: Tuesday, April 18, 2023 2:46 PM
> To: gcc-patches@gcc.gnu.org; Kyrylo Tkachov <Kyrylo.Tkachov@arm.com>;
> Richard Earnshaw <Richard.Earnshaw@arm.com>; Richard Sandiford
> <Richard.Sandiford@arm.com>
> Cc: Christophe Lyon <Christophe.Lyon@arm.com>
> Subject: [PATCH 17/22] arm: [MVE intrinsics] rework vcreateq
> 
> Implement vcreateq using the new MVE builtins framework.

Ok.
Thanks,
Kyrill

> 
> 2022-09-08  Christophe Lyon <christophe.lyon@arm.com>
> 
> 	gcc/
> 	* config/arm/arm-mve-builtins-base.cc (FUNCTION_WITHOUT_M_N):
> New. (vcreateq): New.
> 	* config/arm/arm-mve-builtins-base.def (vcreateq): New.
> 	* config/arm/arm-mve-builtins-base.h (vcreateq): New.
> 	* config/arm/arm_mve.h (vcreateq_f16): Remove.
> 	(vcreateq_f32): Remove.
> 	(vcreateq_u8): Remove.
> 	(vcreateq_u16): Remove.
> 	(vcreateq_u32): Remove.
> 	(vcreateq_u64): Remove.
> 	(vcreateq_s8): Remove.
> 	(vcreateq_s16): Remove.
> 	(vcreateq_s32): Remove.
> 	(vcreateq_s64): Remove.
> 	(__arm_vcreateq_u8): Remove.
> 	(__arm_vcreateq_u16): Remove.
> 	(__arm_vcreateq_u32): Remove.
> 	(__arm_vcreateq_u64): Remove.
> 	(__arm_vcreateq_s8): Remove.
> 	(__arm_vcreateq_s16): Remove.
> 	(__arm_vcreateq_s32): Remove.
> 	(__arm_vcreateq_s64): Remove.
> 	(__arm_vcreateq_f16): Remove.
> 	(__arm_vcreateq_f32): Remove.
> ---
>  gcc/config/arm/arm-mve-builtins-base.cc  | 10 +++
>  gcc/config/arm/arm-mve-builtins-base.def |  2 +
>  gcc/config/arm/arm-mve-builtins-base.h   |  1 +
>  gcc/config/arm/arm_mve.h                 | 80 ------------------------
>  4 files changed, 13 insertions(+), 80 deletions(-)
> 
> diff --git a/gcc/config/arm/arm-mve-builtins-base.cc b/gcc/config/arm/arm-
> mve-builtins-base.cc
> index 499a1ef9f0e..9722c861faf 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.cc
> +++ b/gcc/config/arm/arm-mve-builtins-base.cc
> @@ -106,8 +106,18 @@ namespace arm_mve {
>      UNSPEC##_M_S, UNSPEC##_M_U, UNSPEC##_M_F,
> 	\
>      UNSPEC##_M_N_S, UNSPEC##_M_N_U, -1))
> 
> +  /* Helper for builtins without RTX codes, no _m predicated and no _n
> +     overrides.  */
> +#define FUNCTION_WITHOUT_M_N(NAME, UNSPEC) FUNCTION
> 		\
> +  (NAME, unspec_mve_function_exact_insn,				\
> +   (UNSPEC##_S, UNSPEC##_U, UNSPEC##_F,
> 	\
> +    -1, -1, -1,								\
> +    -1, -1, -1,								\
> +    -1, -1, -1))
> +
>  FUNCTION_WITH_RTX_M_N (vaddq, PLUS, VADDQ)
>  FUNCTION_WITH_RTX_M (vandq, AND, VANDQ)
> +FUNCTION_WITHOUT_M_N (vcreateq, VCREATEQ)
>  FUNCTION_WITH_RTX_M (veorq, XOR, VEORQ)
>  FUNCTION_WITH_RTX_M_N (vmulq, MULT, VMULQ)
>  FUNCTION_WITH_RTX_M_N_NO_N_F (vorrq, IOR, VORRQ)
> diff --git a/gcc/config/arm/arm-mve-builtins-base.def b/gcc/config/arm/arm-
> mve-builtins-base.def
> index c3f8c0f0eeb..1bfd15f973c 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.def
> +++ b/gcc/config/arm/arm-mve-builtins-base.def
> @@ -20,6 +20,7 @@
>  #define REQUIRES_FLOAT false
>  DEF_MVE_FUNCTION (vaddq, binary_opt_n, all_integer, mx_or_none)
>  DEF_MVE_FUNCTION (vandq, binary, all_integer, mx_or_none)
> +DEF_MVE_FUNCTION (vcreateq, create, all_integer_with_64, none)
>  DEF_MVE_FUNCTION (veorq, binary, all_integer, mx_or_none)
>  DEF_MVE_FUNCTION (vmulq, binary_opt_n, all_integer, mx_or_none)
>  DEF_MVE_FUNCTION (vorrq, binary_orrq, all_integer, mx_or_none)
> @@ -31,6 +32,7 @@ DEF_MVE_FUNCTION (vuninitializedq, inherent,
> all_integer_with_64, none)
>  #define REQUIRES_FLOAT true
>  DEF_MVE_FUNCTION (vaddq, binary_opt_n, all_float, mx_or_none)
>  DEF_MVE_FUNCTION (vandq, binary, all_float, mx_or_none)
> +DEF_MVE_FUNCTION (vcreateq, create, all_float, none)
>  DEF_MVE_FUNCTION (veorq, binary, all_float, mx_or_none)
>  DEF_MVE_FUNCTION (vmulq, binary_opt_n, all_float, mx_or_none)
>  DEF_MVE_FUNCTION (vorrq, binary_orrq, all_float, mx_or_none)
> diff --git a/gcc/config/arm/arm-mve-builtins-base.h b/gcc/config/arm/arm-
> mve-builtins-base.h
> index c450b373239..8dd6bff01bf 100644
> --- a/gcc/config/arm/arm-mve-builtins-base.h
> +++ b/gcc/config/arm/arm-mve-builtins-base.h
> @@ -25,6 +25,7 @@ namespace functions {
> 
>  extern const function_base *const vaddq;
>  extern const function_base *const vandq;
> +extern const function_base *const vcreateq;
>  extern const function_base *const veorq;
>  extern const function_base *const vmulq;
>  extern const function_base *const vorrq;
> diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h
> index edf8e247421..4810e2977d3 100644
> --- a/gcc/config/arm/arm_mve.h
> +++ b/gcc/config/arm/arm_mve.h
> @@ -638,20 +638,10 @@
>  #define vcvtq_n_f32_s32(__a,  __imm6) __arm_vcvtq_n_f32_s32(__a,
> __imm6)
>  #define vcvtq_n_f16_u16(__a,  __imm6) __arm_vcvtq_n_f16_u16(__a,
> __imm6)
>  #define vcvtq_n_f32_u32(__a,  __imm6) __arm_vcvtq_n_f32_u32(__a,
> __imm6)
> -#define vcreateq_f16(__a, __b) __arm_vcreateq_f16(__a, __b)
> -#define vcreateq_f32(__a, __b) __arm_vcreateq_f32(__a, __b)
>  #define vcvtq_n_s16_f16(__a,  __imm6) __arm_vcvtq_n_s16_f16(__a,
> __imm6)
>  #define vcvtq_n_s32_f32(__a,  __imm6) __arm_vcvtq_n_s32_f32(__a,
> __imm6)
>  #define vcvtq_n_u16_f16(__a,  __imm6) __arm_vcvtq_n_u16_f16(__a,
> __imm6)
>  #define vcvtq_n_u32_f32(__a,  __imm6) __arm_vcvtq_n_u32_f32(__a,
> __imm6)
> -#define vcreateq_u8(__a, __b) __arm_vcreateq_u8(__a, __b)
> -#define vcreateq_u16(__a, __b) __arm_vcreateq_u16(__a, __b)
> -#define vcreateq_u32(__a, __b) __arm_vcreateq_u32(__a, __b)
> -#define vcreateq_u64(__a, __b) __arm_vcreateq_u64(__a, __b)
> -#define vcreateq_s8(__a, __b) __arm_vcreateq_s8(__a, __b)
> -#define vcreateq_s16(__a, __b) __arm_vcreateq_s16(__a, __b)
> -#define vcreateq_s32(__a, __b) __arm_vcreateq_s32(__a, __b)
> -#define vcreateq_s64(__a, __b) __arm_vcreateq_s64(__a, __b)
>  #define vshrq_n_s8(__a,  __imm) __arm_vshrq_n_s8(__a,  __imm)
>  #define vshrq_n_s16(__a,  __imm) __arm_vshrq_n_s16(__a,  __imm)
>  #define vshrq_n_s32(__a,  __imm) __arm_vshrq_n_s32(__a,  __imm)
> @@ -3222,62 +3212,6 @@ __arm_vpnot (mve_pred16_t __a)
>    return __builtin_mve_vpnotv16bi (__a);
>  }
> 
> -__extension__ extern __inline uint8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_u8 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_uv16qi (__a, __b);
> -}
> -
> -__extension__ extern __inline uint16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_u16 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_uv8hi (__a, __b);
> -}
> -
> -__extension__ extern __inline uint32x4_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_u32 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_uv4si (__a, __b);
> -}
> -
> -__extension__ extern __inline uint64x2_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_u64 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_uv2di (__a, __b);
> -}
> -
> -__extension__ extern __inline int8x16_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_s8 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_sv16qi (__a, __b);
> -}
> -
> -__extension__ extern __inline int16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_s16 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_sv8hi (__a, __b);
> -}
> -
> -__extension__ extern __inline int32x4_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_s32 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_sv4si (__a, __b);
> -}
> -
> -__extension__ extern __inline int64x2_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_s64 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_sv2di (__a, __b);
> -}
> -
>  __extension__ extern __inline int8x16_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vshrq_n_s8 (int8x16_t __a, const int __imm)
> @@ -15580,20 +15514,6 @@ __arm_vcvtq_n_f32_u32 (uint32x4_t __a, const
> int __imm6)
>    return __builtin_mve_vcvtq_n_to_f_uv4sf (__a, __imm6);
>  }
> 
> -__extension__ extern __inline float16x8_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_f16 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_fv8hf (__a, __b);
> -}
> -
> -__extension__ extern __inline float32x4_t
> -__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
> -__arm_vcreateq_f32 (uint64_t __a, uint64_t __b)
> -{
> -  return __builtin_mve_vcreateq_fv4sf (__a, __b);
> -}
> -
>  __extension__ extern __inline int16x8_t
>  __attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
>  __arm_vcvtq_n_s16_f16 (float16x8_t __a, const int __imm6)
> --
> 2.34.1