3025 lines
308 KiB
C
3025 lines
308 KiB
C
/***
|
|
* arm_neon.h - declarations/definitions for ARM NEON specific intrinsics
|
|
*
|
|
* Copyright (c) Microsoft Corporation. All rights reserved.
|
|
*
|
|
*Purpose:
|
|
* This include file contains the declarations for ARM NEON intrinsic functions
|
|
*
|
|
****/
|
|
|
|
#pragma once
|
|
|
|
#include <stdint.h>
|
|
#include <sal.h>
|
|
|
|
#if !defined (_M_ARM)
|
|
#error This header is specific to ARM targets
|
|
#endif /* !defined (_M_ARM) */
|
|
|
|
|
|
#if defined (__cplusplus)
|
|
extern "C" {
|
|
#endif /* defined (__cplusplus) */
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
#if !defined (_ADVSIMD_ALIGN)
|
|
#if defined (__midl)
|
|
#define _ADVSIMD_ALIGN(x)
|
|
#else /* defined (__midl) */
|
|
#define _ADVSIMD_ALIGN(x) __declspec(align(x))
|
|
#endif /* defined (__midl) */
|
|
#endif /* !defined (_ADVSIMD_ALIGN) */
|
|
|
|
#ifndef DUMMYNEONSTRUCT
|
|
#define DUMMYNEONSTRUCT s
|
|
#endif /* DUMMYNEONSTRUCT */
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// ARM Advanced SIMD 64bit type
|
|
//
|
|
typedef union __declspec(intrin_type) _ADVSIMD_ALIGN(8) __n64
|
|
{
|
|
unsigned __int64 n64_u64[1];
|
|
unsigned __int32 n64_u32[2];
|
|
unsigned __int16 n64_u16[4];
|
|
unsigned __int8 n64_u8[8];
|
|
__int64 n64_i64[1];
|
|
__int32 n64_i32[2];
|
|
__int16 n64_i16[4];
|
|
__int8 n64_i8[8];
|
|
float n64_f32[2];
|
|
} __n64;
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// ARM Advanced SIMD 128bit type
|
|
//
|
|
typedef union __declspec(intrin_type) _ADVSIMD_ALIGN(8) __n128
|
|
{
|
|
unsigned __int64 n128_u64[2];
|
|
unsigned __int32 n128_u32[4];
|
|
unsigned __int16 n128_u16[8];
|
|
unsigned __int8 n128_u8[16];
|
|
__int64 n128_i64[2];
|
|
__int32 n128_i32[4];
|
|
__int16 n128_i16[8];
|
|
__int8 n128_i8[16];
|
|
float n128_f32[4];
|
|
|
|
struct
|
|
{
|
|
__n64 low64;
|
|
__n64 high64;
|
|
} DUMMYNEONSTRUCT;
|
|
|
|
} __n128;
|
|
|
|
typedef struct __n64x2
|
|
{
|
|
__n64 val[2];
|
|
} __n64x2;
|
|
|
|
typedef struct __n64x3
|
|
{
|
|
__n64 val[3];
|
|
} __n64x3;
|
|
|
|
typedef struct __n64x4
|
|
{
|
|
__n64 val[4];
|
|
} __n64x4;
|
|
|
|
typedef struct __n128x2
|
|
{
|
|
__n128 val[2];
|
|
} __n128x2;
|
|
|
|
typedef struct __n128x3
|
|
{
|
|
__n128 val[3];
|
|
} __n128x3;
|
|
|
|
typedef struct __n128x4
|
|
{
|
|
__n128 val[4];
|
|
} __n128x4;
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
typedef unsigned __int8 poly8_t;
|
|
typedef unsigned __int16 poly16_t;
|
|
|
|
typedef float float32_t;
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
__inline _Post_equal_to_(p) __n64 *__int8ToN64(_In_ int8_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__int16ToN64(_In_ int16_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__int32ToN64(_In_ int32_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__int64ToN64(_In_ int64_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__uint8ToN64(_In_ uint8_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__uint16ToN64(_In_ uint16_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__uint32ToN64(_In_ uint32_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__uint64ToN64(_In_ uint64_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__poly8ToN64(_In_ poly8_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__poly16ToN64(_In_ poly16_t *p) { return (__n64 *)p; }
|
|
__inline _Post_equal_to_(p) __n64 *__float32ToN64(_In_ float32_t *p) { return (__n64 *)p; }
|
|
|
|
__inline _Post_equal_to_(p) const __n64 *__int8ToN64_c(_In_ const int8_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__int16ToN64_c(_In_ const int16_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__int32ToN64_c(_In_ const int32_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__int64ToN64_c(_In_ const int64_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__uint8ToN64_c(_In_ const uint8_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__uint16ToN64_c(_In_ const uint16_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__uint32ToN64_c(_In_ const uint32_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__uint64ToN64_c(_In_ const uint64_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__poly8ToN64_c(_In_ const poly8_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__poly16ToN64_c(_In_ const poly16_t *p) { return (const __n64 *)p; }
|
|
__inline _Post_equal_to_(p) const __n64 *__float32ToN64_c(_In_ const float32_t *p) { return (const __n64 *)p; }
|
|
|
|
__inline int32_t __int8ToInt32(int8_t i) { return (int32_t)i; }
|
|
__inline int32_t __int16ToInt32(int16_t i) { return (int32_t)i; }
|
|
__inline int32_t __int32ToInt32(int32_t i) { return (int32_t)i; }
|
|
__inline int64_t __int64ToInt64(int64_t i) { return (int64_t)i; }
|
|
|
|
__inline int32_t __uint8ToInt32(uint8_t i) { return (int32_t)i; }
|
|
__inline int32_t __uint16ToInt32(uint16_t i) { return (int32_t)i; }
|
|
__inline int32_t __uint32ToInt32(uint32_t i) { return (int32_t)i; }
|
|
__inline int64_t __uint64ToInt64(uint64_t i) { return (int64_t)i; }
|
|
|
|
__inline int32_t __poly8ToInt32(poly8_t i) { return (int32_t)i; }
|
|
__inline int32_t __poly16ToInt32(poly16_t i) { return (int32_t)i; }
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
#define vshll_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 8, "invalid shift amount"), ((shift_amount) == 8) ? __internal_vshll_n_t2_s8((Dm)) : __internal_vshll_n_t1_s8((Dm), (shift_amount)) )
|
|
#define vshll_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 16, "invalid shift amount"), ((shift_amount) == 16) ? __internal_vshll_n_t2_s16((Dm)) : __internal_vshll_n_t1_s16((Dm), (shift_amount)) )
|
|
#define vshll_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 32, "invalid shift amount"), ((shift_amount) == 32) ? __internal_vshll_n_t2_s32((Dm)) : __internal_vshll_n_t1_s32((Dm), (shift_amount)) )
|
|
#define vshll_n_u8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 8, "invalid shift amount"), ((shift_amount) == 8) ? __internal_vshll_n_t2_u8((Dm)) : __internal_vshll_n_t1_u8((Dm), (shift_amount)) )
|
|
#define vshll_n_u16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 16, "invalid shift amount"), ((shift_amount) == 16) ? __internal_vshll_n_t2_u16((Dm)) : __internal_vshll_n_t1_u16((Dm), (shift_amount)) )
|
|
#define vshll_n_u32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) <= 32, "invalid shift amount"), ((shift_amount) == 32) ? __internal_vshll_n_t2_u32((Dm)) : __internal_vshll_n_t1_u32((Dm), (shift_amount)) )
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// { +++ auto-generated code begins (explicit types)
|
|
|
|
typedef __n64 float32x2_t;
|
|
typedef __n64x2 float32x2x2_t;
|
|
typedef __n64x3 float32x2x3_t;
|
|
typedef __n64x4 float32x2x4_t;
|
|
typedef __n64 int8x8_t;
|
|
typedef __n64x2 int8x8x2_t;
|
|
typedef __n64x3 int8x8x3_t;
|
|
typedef __n64x4 int8x8x4_t;
|
|
typedef __n64 int16x4_t;
|
|
typedef __n64x2 int16x4x2_t;
|
|
typedef __n64x3 int16x4x3_t;
|
|
typedef __n64x4 int16x4x4_t;
|
|
typedef __n64 int32x2_t;
|
|
typedef __n64x2 int32x2x2_t;
|
|
typedef __n64x3 int32x2x3_t;
|
|
typedef __n64x4 int32x2x4_t;
|
|
typedef __n64 int64x1_t;
|
|
typedef __n64x2 int64x1x2_t;
|
|
typedef __n64x3 int64x1x3_t;
|
|
typedef __n64x4 int64x1x4_t;
|
|
typedef __n64 poly8x8_t;
|
|
typedef __n64x2 poly8x8x2_t;
|
|
typedef __n64x3 poly8x8x3_t;
|
|
typedef __n64x4 poly8x8x4_t;
|
|
typedef __n64 poly16x4_t;
|
|
typedef __n64x2 poly16x4x2_t;
|
|
typedef __n64x3 poly16x4x3_t;
|
|
typedef __n64x4 poly16x4x4_t;
|
|
typedef __n64 uint8x8_t;
|
|
typedef __n64x2 uint8x8x2_t;
|
|
typedef __n64x3 uint8x8x3_t;
|
|
typedef __n64x4 uint8x8x4_t;
|
|
typedef __n64 uint16x4_t;
|
|
typedef __n64x2 uint16x4x2_t;
|
|
typedef __n64x3 uint16x4x3_t;
|
|
typedef __n64x4 uint16x4x4_t;
|
|
typedef __n64 uint32x2_t;
|
|
typedef __n64x2 uint32x2x2_t;
|
|
typedef __n64x3 uint32x2x3_t;
|
|
typedef __n64x4 uint32x2x4_t;
|
|
typedef __n64 uint64x1_t;
|
|
typedef __n64x2 uint64x1x2_t;
|
|
typedef __n64x3 uint64x1x3_t;
|
|
typedef __n64x4 uint64x1x4_t;
|
|
typedef __n128 float32x4_t;
|
|
typedef __n128x2 float32x4x2_t;
|
|
typedef __n128x3 float32x4x3_t;
|
|
typedef __n128x4 float32x4x4_t;
|
|
typedef __n128 int8x16_t;
|
|
typedef __n128x2 int8x16x2_t;
|
|
typedef __n128x3 int8x16x3_t;
|
|
typedef __n128x4 int8x16x4_t;
|
|
typedef __n128 int16x8_t;
|
|
typedef __n128x2 int16x8x2_t;
|
|
typedef __n128x3 int16x8x3_t;
|
|
typedef __n128x4 int16x8x4_t;
|
|
typedef __n128 int32x4_t;
|
|
typedef __n128x2 int32x4x2_t;
|
|
typedef __n128x3 int32x4x3_t;
|
|
typedef __n128x4 int32x4x4_t;
|
|
typedef __n128 int64x2_t;
|
|
typedef __n128x2 int64x2x2_t;
|
|
typedef __n128x3 int64x2x3_t;
|
|
typedef __n128x4 int64x2x4_t;
|
|
typedef __n128 poly8x16_t;
|
|
typedef __n128x2 poly8x16x2_t;
|
|
typedef __n128x3 poly8x16x3_t;
|
|
typedef __n128x4 poly8x16x4_t;
|
|
typedef __n128 poly16x8_t;
|
|
typedef __n128x2 poly16x8x2_t;
|
|
typedef __n128x3 poly16x8x3_t;
|
|
typedef __n128x4 poly16x8x4_t;
|
|
typedef __n128 uint8x16_t;
|
|
typedef __n128x2 uint8x16x2_t;
|
|
typedef __n128x3 uint8x16x3_t;
|
|
typedef __n128x4 uint8x16x4_t;
|
|
typedef __n128 uint16x8_t;
|
|
typedef __n128x2 uint16x8x2_t;
|
|
typedef __n128x3 uint16x8x3_t;
|
|
typedef __n128x4 uint16x8x4_t;
|
|
typedef __n128 uint32x4_t;
|
|
typedef __n128x2 uint32x4x2_t;
|
|
typedef __n128x3 uint32x4x3_t;
|
|
typedef __n128x4 uint32x4x4_t;
|
|
typedef __n128 uint64x2_t;
|
|
typedef __n128x2 uint64x2x2_t;
|
|
typedef __n128x3 uint64x2x3_t;
|
|
typedef __n128x4 uint64x2x4_t;
|
|
|
|
// } +++ auto-generated code ends (explicit types)
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// { +++ auto-generated code begins (prototypes)
|
|
|
|
__n64x2 __neon_DdDm_acc2(unsigned int _Enc, __n64, __n64);
|
|
__n64x2 __neon_Dx2Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64x2 __neon_Dx2Adr_acc(unsigned int _Enc, __n64x2, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64x3 __neon_Dx3Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64x3 __neon_Dx3Adr_acc(unsigned int _Enc, __n64x3, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64x4 __neon_Dx4Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64x4 __neon_Dx4Adr_acc(unsigned int _Enc, __n64x4, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64 __neon_DdDm(unsigned int _Enc, __n64);
|
|
__n64 __neon_DdDx2Dm(unsigned int _Enc, __n64x2, __n64);
|
|
__n64 __neon_DdDx2Dm_acc(unsigned int _Enc, __n64, __n64x2, __n64);
|
|
__n64 __neon_DdDx3Dm(unsigned int _Enc, __n64x3, __n64);
|
|
__n64 __neon_DdDx3Dm_acc(unsigned int _Enc, __n64, __n64x3, __n64);
|
|
__n64 __neon_DdDx4Dm(unsigned int _Enc, __n64x4, __n64);
|
|
__n64 __neon_DdDx4Dm_acc(unsigned int _Enc, __n64, __n64x4, __n64);
|
|
__n64 __neon_DdDm_acc(unsigned int _Enc, __n64, __n64);
|
|
__n64 __neon_DdDnDm(unsigned int _Enc, __n64, __n64);
|
|
__n64 __neon_DdDnDm_acc(unsigned int _Enc, __n64, __n64, __n64);
|
|
__n64 __neon_DdDnDmx(unsigned int _Enc, __n64, __n64);
|
|
__n64 __neon_DdDnDmx_acc(unsigned int _Enc, __n64, __n64, __n64);
|
|
__n64 __neon_DdDnFt(unsigned int, __n64, float);
|
|
__n64 __neon_DdDnFt_acc(unsigned int, __n64, __n64, float);
|
|
__n64 __neon_DdFt(unsigned int _Enc, float);
|
|
__n64 __neon_DdFt_acc(unsigned int _Enc, __n64, float);
|
|
__n64 __neon_D1Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64 __neon_D1Adr_acc(unsigned int _Enc, __n64, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n64 __neon_DdQm(unsigned int _Enc, __n128);
|
|
__n64 __neon_DdQm_high(unsigned int _Enc, __n128);
|
|
__n64 __neon_DdQm_low(unsigned int _Enc, __n128);
|
|
__n64 __neon_DdQnQm(unsigned int _Enc, __n128, __n128);
|
|
__n64 __neon_DdRt(unsigned int _Enc, int);
|
|
__n64 __neon_DdRtRt2(unsigned int _Enc, __int64);
|
|
__n64 __neon_DdRtRt2_acc(unsigned int _Enc, __n64, __int64);
|
|
__n64 __neon_DdRt_acc(unsigned int _Enc, __n64, int);
|
|
float __neon_FtDn(unsigned int _Enc, __n64);
|
|
float __neon_FtQn(unsigned int _Enc, __n128);
|
|
__n128x2 __neon_Qx2Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128x2 __neon_Qx2Adr_acc(unsigned int _Enc, __n128x2, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128x2 __neon_QdQm_acc2(unsigned int _Enc, __n128, __n128);
|
|
__n128x2 __neon_QdQm_acc3(unsigned int _Enc, __n128, __n128);
|
|
__n128x3 __neon_Qx3Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128x3 __neon_Qx3Adr_acc(unsigned int _Enc, __n128x3, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128x4 __neon_Qx4Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128x4 __neon_Qx4Adr_acc(unsigned int _Enc, __n128x4, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128 __neon_QdDm(unsigned int _Enc, __n64);
|
|
__n128 __neon_QdDnDm(unsigned int _Enc, __n64, __n64);
|
|
__n128 __neon_QdDnDm_acc(unsigned int _Enc, __n128, __n64, __n64);
|
|
__n128 __neon_QdDnDm_merge(unsigned int _Enc, __n64, __n64);
|
|
__n128 __neon_QdDnDmx(unsigned int _Enc, __n64, __n64);
|
|
__n128 __neon_QdDnDmx_acc(unsigned int _Enc, __n128, __n64, __n64);
|
|
__n128 __neon_QdFt(unsigned int _Enc, float);
|
|
__n128 __neon_QdFt_acc(unsigned int _Enc, __n128, float);
|
|
__n128 __neon_Q1Adr(unsigned int _Enc, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128 __neon_Q1Adr_acc(unsigned int _Enc, __n128, _In_reads_bytes_(_Inexpressible_(_Enc)) const __n64*);
|
|
__n128 __neon_QdQm(unsigned int _Enc, __n128);
|
|
__n128 __neon_QdQm_acc(unsigned int _Enc, __n128, __n128);
|
|
__n128 __neon_QdQnDm(unsigned int _Enc, __n128, __n64);
|
|
__n128 __neon_QdQnDmx(unsigned int _Enc, __n128, __n64);
|
|
__n128 __neon_QdQnDmx_acc(unsigned int _Enc, __n128, __n128, __n64);
|
|
__n128 __neon_QdQnFt(unsigned int, __n128, float);
|
|
__n128 __neon_QdQnFt_acc(unsigned int, __n128, __n128, float);
|
|
__n128 __neon_QdQnQm(unsigned int _Enc, __n128, __n128);
|
|
__n128 __neon_QdQnQm_acc(unsigned int _Enc, __n128, __n128, __n128);
|
|
__n128 __neon_QdRt(unsigned int _Enc, int);
|
|
__n128 __neon_QdRtRt2_acc(unsigned int _Enc, __n128, __int64);
|
|
__n128 __neon_QdRtRt2_dup(unsigned int _Enc, __int64);
|
|
__n128 __neon_QdRt_acc(unsigned int _Enc, __n128, int);
|
|
__int64 __neon_RtRt2Dm(unsigned int _Enc, __n64);
|
|
__int64 __neon_RtRt2Qm(unsigned int _Enc, __n128);
|
|
int __neon_RtDn(unsigned int _Enc, __n64);
|
|
int __neon_RtQn(unsigned int _Enc, __n128);
|
|
void __neon_AdrD1(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64);
|
|
void __neon_AdrDx2(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x2);
|
|
void __neon_AdrDx2x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x2);
|
|
void __neon_AdrDx3(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x3);
|
|
void __neon_AdrDx3x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x3);
|
|
void __neon_AdrDx4(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x4);
|
|
void __neon_AdrDx4x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n64x4);
|
|
void __neon_AdrQ1(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128);
|
|
void __neon_AdrQx2(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x2);
|
|
void __neon_AdrQx2x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x2);
|
|
void __neon_AdrQx3(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x3);
|
|
void __neon_AdrQx3x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x3);
|
|
void __neon_AdrQx4(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x4);
|
|
void __neon_AdrQx4x(unsigned int _Enc, _Out_writes_bytes_(_Inexpressible_(_Enc)) __n64*, __n128x4);
|
|
|
|
// } +++ auto-generated code ends (prototypes)
|
|
|
|
|
|
#if defined (__cplusplus)
|
|
}
|
|
#endif /* defined (__cplusplus) */
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// VLDx/VSTx alignment specifications
|
|
//
|
|
|
|
|
|
#define _NEON_ALIGN16(a) \
|
|
( \
|
|
((a) == 8) ? 0 : \
|
|
((a) == 16) ? 1 : \
|
|
-1)
|
|
|
|
#define _NEON_ALIGN32(a) \
|
|
( \
|
|
((a) == 8) ? 0 : \
|
|
((a) == 32) ? 1 : \
|
|
-1)
|
|
|
|
#define _NEON_ALIGN64(a) \
|
|
( \
|
|
((a) == 8) ? 0 : \
|
|
((a) == 64) ? 1 : \
|
|
-1)
|
|
|
|
#define _NEON_ALIGN64_128(a) \
|
|
( \
|
|
((a) == 8) ? 0 : \
|
|
((a) == 64) ? 1 : \
|
|
((a) == 128) ? 2 : \
|
|
-1)
|
|
|
|
|
|
#define _NEON_ALIGN64_128_256(a) \
|
|
( \
|
|
((a) == 8) ? 0 : \
|
|
((a) == 64) ? 1 : \
|
|
((a) == 128) ? 2 : \
|
|
((a) == 256) ? 3 : \
|
|
-1)
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// { +++ auto-generated code begins (encoding macros)
|
|
|
|
#define _NENC_0(x) ((x) & 0x1)
|
|
#define _NENC_11_8(x) (((x) << 8) & 0xf00)
|
|
#define _NENC_12(x) (((x) << 12) & 0x1000)
|
|
#define _NENC_16(x) (((x) << 16) & 0x10000)
|
|
#define _NENC_18_16(x) (((x) << 16) & 0x70000)
|
|
#define _NENC_19(x) (((x) << 19) & 0x80000)
|
|
#define _NENC_19_16(x) (((x) << 16) & 0xf0000)
|
|
#define _NENC_19_17(x) (((x) << 17) & 0xe0000)
|
|
#define _NENC_19_18(x) (((x) << 18) & 0xc0000)
|
|
#define _NENC_20_16(x) (((x) << 16) & 0x1f0000)
|
|
#define _NENC_21(x) (((x) << 21) & 0x200000)
|
|
#define _NENC_21_16(x) (((x) << 16) & 0x3f0000)
|
|
#define _NENC_21x6(x) (((x) << 6) & 0x40 | ((x) << 20) & 0x200000)
|
|
#define _NENC_21x6_5(x) (((x) << 5) & 0x60 | ((x) << 19) & 0x200000)
|
|
#define _NENC_4(x) (((x) << 4) & 0x10)
|
|
#define _NENC_5(x) (((x) << 5) & 0x20)
|
|
#define _NENC_5_4(x) (((x) << 4) & 0x30)
|
|
#define _NENC_5x3(x) (((x) << 3) & 0x8 | ((x) << 4) & 0x20)
|
|
#define _NENC_7(x) (((x) << 7) & 0x80)
|
|
#define _NENC_7_5(x) (((x) << 5) & 0xe0)
|
|
#define _NENC_7_6(x) (((x) << 6) & 0xc0)
|
|
|
|
// } +++ auto-generated code ends (encoding macros)
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// { +++ auto-generated code begins (Neon macros)
|
|
|
|
// AES
|
|
#define aesd_p8(Qm) ( __neon_QdQm( 0xf3b00340, (Qm)) )
|
|
#define aesd_s8(Qm) ( __neon_QdQm( 0xf3b00340, (Qm)) )
|
|
#define aesd_u8(Qm) ( __neon_QdQm( 0xf3b00340, (Qm)) )
|
|
#define aese_p8(Qm) ( __neon_QdQm( 0xf3b00300, (Qm)) )
|
|
#define aese_s8(Qm) ( __neon_QdQm( 0xf3b00300, (Qm)) )
|
|
#define aese_u8(Qm) ( __neon_QdQm( 0xf3b00300, (Qm)) )
|
|
#define aesimc_p8(Qm) ( __neon_QdQm( 0xf3b003c0, (Qm)) )
|
|
#define aesimc_s8(Qm) ( __neon_QdQm( 0xf3b003c0, (Qm)) )
|
|
#define aesimc_u8(Qm) ( __neon_QdQm( 0xf3b003c0, (Qm)) )
|
|
#define aesmc_p8(Qm) ( __neon_QdQm( 0xf3b00380, (Qm)) )
|
|
#define aesmc_s8(Qm) ( __neon_QdQm( 0xf3b00380, (Qm)) )
|
|
#define aesmc_u8(Qm) ( __neon_QdQm( 0xf3b00380, (Qm)) )
|
|
|
|
// SHA (2-operand)
|
|
#define sha1h_f32(Qm) ( __neon_QdQm( 0xf3b902c0, (Qm)) )
|
|
#define sha1h_s32(Qm) ( __neon_QdQm( 0xf3b902c0, (Qm)) )
|
|
#define sha1h_u32(Qm) ( __neon_QdQm( 0xf3b902c0, (Qm)) )
|
|
#define sha1su1_f32(Qm) ( __neon_QdQm( 0xf3ba0380, (Qm)) )
|
|
#define sha1su1_s32(Qm) ( __neon_QdQm( 0xf3ba0380, (Qm)) )
|
|
#define sha1su1_u32(Qm) ( __neon_QdQm( 0xf3ba0380, (Qm)) )
|
|
#define sha256su0_f32(Qm) ( __neon_QdQm( 0xf3ba03c0, (Qm)) )
|
|
#define sha256su0_s32(Qm) ( __neon_QdQm( 0xf3ba03c0, (Qm)) )
|
|
#define sha256su0_u32(Qm) ( __neon_QdQm( 0xf3ba03c0, (Qm)) )
|
|
|
|
// SHA (3-operand)
|
|
#define sha1c_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2000c40, (Qn), (Qm)) )
|
|
#define sha1c_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2000c40, (Qn), (Qm)) )
|
|
#define sha1c_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2000c40, (Qn), (Qm)) )
|
|
#define sha1m_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2200c40, (Qn), (Qm)) )
|
|
#define sha1m_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200c40, (Qn), (Qm)) )
|
|
#define sha1m_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2200c40, (Qn), (Qm)) )
|
|
#define sha1p_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2100c40, (Qn), (Qm)) )
|
|
#define sha1p_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2100c40, (Qn), (Qm)) )
|
|
#define sha1p_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2100c40, (Qn), (Qm)) )
|
|
#define sha1su0_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2300c40, (Qn), (Qm)) )
|
|
#define sha1su0_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2300c40, (Qn), (Qm)) )
|
|
#define sha1su0_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2300c40, (Qn), (Qm)) )
|
|
#define sha256h_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000c40, (Qn), (Qm)) )
|
|
#define sha256h_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3000c40, (Qn), (Qm)) )
|
|
#define sha256h_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3000c40, (Qn), (Qm)) )
|
|
#define sha256h2_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3100c40, (Qn), (Qm)) )
|
|
#define sha256h2_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3100c40, (Qn), (Qm)) )
|
|
#define sha256h2_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3100c40, (Qn), (Qm)) )
|
|
#define sha256su1_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200c40, (Qn), (Qm)) )
|
|
#define sha256su1_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3200c40, (Qn), (Qm)) )
|
|
#define sha256su1_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200c40, (Qn), (Qm)) )
|
|
|
|
// VABA, VABAL
|
|
#define vaba_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2100710, (Dd), (Dn), (Dm)) )
|
|
#define vaba_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2200710, (Dd), (Dn), (Dm)) )
|
|
#define vaba_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2000710, (Dd), (Dn), (Dm)) )
|
|
#define vaba_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100710, (Dd), (Dn), (Dm)) )
|
|
#define vaba_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200710, (Dd), (Dn), (Dm)) )
|
|
#define vaba_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3000710, (Dd), (Dn), (Dm)) )
|
|
#define vabal_s16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2900500, (Qd), (Dn), (Dm)) )
|
|
#define vabal_s32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2a00500, (Qd), (Dn), (Dm)) )
|
|
#define vabal_s8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2800500, (Qd), (Dn), (Dm)) )
|
|
#define vabal_u16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3900500, (Qd), (Dn), (Dm)) )
|
|
#define vabal_u32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3a00500, (Qd), (Dn), (Dm)) )
|
|
#define vabal_u8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3800500, (Qd), (Dn), (Dm)) )
|
|
#define vabaq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2100750, (Qd), (Qn), (Qm)) )
|
|
#define vabaq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2200750, (Qd), (Qn), (Qm)) )
|
|
#define vabaq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2000750, (Qd), (Qn), (Qm)) )
|
|
#define vabaq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100750, (Qd), (Qn), (Qm)) )
|
|
#define vabaq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200750, (Qd), (Qn), (Qm)) )
|
|
#define vabaq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3000750, (Qd), (Qn), (Qm)) )
|
|
|
|
// VABD (floating point)
|
|
#define vabd_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200d00, (Dn), (Dm)) )
|
|
#define vabdq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200d40, (Qn), (Qm)) )
|
|
|
|
// VABD[L] (integer)
|
|
#define vabd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100700, (Dn), (Dm)) )
|
|
#define vabd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200700, (Dn), (Dm)) )
|
|
#define vabd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000700, (Dn), (Dm)) )
|
|
#define vabd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100700, (Dn), (Dm)) )
|
|
#define vabd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200700, (Dn), (Dm)) )
|
|
#define vabd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000700, (Dn), (Dm)) )
|
|
#define vabdl_s16(Dn, Dm) ( __neon_QdDnDm( 0xf2900700, (Dn), (Dm)) )
|
|
#define vabdl_s32(Dn, Dm) ( __neon_QdDnDm( 0xf2a00700, (Dn), (Dm)) )
|
|
#define vabdl_s8(Dn, Dm) ( __neon_QdDnDm( 0xf2800700, (Dn), (Dm)) )
|
|
#define vabdl_u16(Dn, Dm) ( __neon_QdDnDm( 0xf3900700, (Dn), (Dm)) )
|
|
#define vabdl_u32(Dn, Dm) ( __neon_QdDnDm( 0xf3a00700, (Dn), (Dm)) )
|
|
#define vabdl_u8(Dn, Dm) ( __neon_QdDnDm( 0xf3800700, (Dn), (Dm)) )
|
|
#define vabdq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100740, (Qn), (Qm)) )
|
|
#define vabdq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200740, (Qn), (Qm)) )
|
|
#define vabdq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000740, (Qn), (Qm)) )
|
|
#define vabdq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100740, (Qn), (Qm)) )
|
|
#define vabdq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200740, (Qn), (Qm)) )
|
|
#define vabdq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000740, (Qn), (Qm)) )
|
|
|
|
// VABS, VNEG
|
|
#define vabs_f32(Dm) ( __neon_DdDm( 0xf3b90700, (Dm)) )
|
|
#define vabs_s16(Dm) ( __neon_DdDm( 0xf3b50300, (Dm)) )
|
|
#define vabs_s32(Dm) ( __neon_DdDm( 0xf3b90300, (Dm)) )
|
|
#define vabs_s8(Dm) ( __neon_DdDm( 0xf3b10300, (Dm)) )
|
|
#define vneg_f32(Dm) ( __neon_DdDm( 0xf3b90780, (Dm)) )
|
|
#define vneg_s16(Dm) ( __neon_DdDm( 0xf3b50380, (Dm)) )
|
|
#define vneg_s32(Dm) ( __neon_DdDm( 0xf3b90380, (Dm)) )
|
|
#define vneg_s8(Dm) ( __neon_DdDm( 0xf3b10380, (Dm)) )
|
|
#define vabsq_f32(Qm) ( __neon_QdQm( 0xf3b90740, (Qm)) )
|
|
#define vabsq_s16(Qm) ( __neon_QdQm( 0xf3b50340, (Qm)) )
|
|
#define vabsq_s32(Qm) ( __neon_QdQm( 0xf3b90340, (Qm)) )
|
|
#define vabsq_s8(Qm) ( __neon_QdQm( 0xf3b10340, (Qm)) )
|
|
#define vnegq_f32(Qm) ( __neon_QdQm( 0xf3b907c0, (Qm)) )
|
|
#define vnegq_s16(Qm) ( __neon_QdQm( 0xf3b503c0, (Qm)) )
|
|
#define vnegq_s32(Qm) ( __neon_QdQm( 0xf3b903c0, (Qm)) )
|
|
#define vnegq_s8(Qm) ( __neon_QdQm( 0xf3b103c0, (Qm)) )
|
|
|
|
// VACGE, VACGT, VACLE, VACLT
|
|
#define vacge_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000e10, (Dn), (Dm)) )
|
|
#define vacgt_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200e10, (Dn), (Dm)) )
|
|
#define vacle_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000e10, (Dm), (Dn)) )
|
|
#define vaclt_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200e10, (Dm), (Dn)) )
|
|
#define vacgeq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000e50, (Qn), (Qm)) )
|
|
#define vacgtq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200e50, (Qn), (Qm)) )
|
|
#define vacleq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000e50, (Qm), (Qn)) )
|
|
#define vacltq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200e50, (Qm), (Qn)) )
|
|
|
|
// VADD
|
|
#define vadd_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2000d00, (Dn), (Dm)) )
|
|
#define vadd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100800, (Dn), (Dm)) )
|
|
#define vadd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200800, (Dn), (Dm)) )
|
|
#define vadd_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300800, (Dn), (Dm)) )
|
|
#define vadd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000800, (Dn), (Dm)) )
|
|
#define vadd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2100800, (Dn), (Dm)) )
|
|
#define vadd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2200800, (Dn), (Dm)) )
|
|
#define vadd_u64(Dn, Dm) ( __neon_DdDnDm( 0xf2300800, (Dn), (Dm)) )
|
|
#define vadd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2000800, (Dn), (Dm)) )
|
|
#define vaddq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2000d40, (Qn), (Qm)) )
|
|
#define vaddq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100840, (Qn), (Qm)) )
|
|
#define vaddq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200840, (Qn), (Qm)) )
|
|
#define vaddq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300840, (Qn), (Qm)) )
|
|
#define vaddq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000840, (Qn), (Qm)) )
|
|
#define vaddq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2100840, (Qn), (Qm)) )
|
|
#define vaddq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2200840, (Qn), (Qm)) )
|
|
#define vaddq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf2300840, (Qn), (Qm)) )
|
|
#define vaddq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2000840, (Qn), (Qm)) )
|
|
|
|
// VADDHN, VRADDHN
|
|
#define vaddhn_s16(Qn, Qm) ( __neon_DdQnQm( 0xf2800400, (Qn), (Qm)) )
|
|
#define vaddhn_s32(Qn, Qm) ( __neon_DdQnQm( 0xf2900400, (Qn), (Qm)) )
|
|
#define vaddhn_s64(Qn, Qm) ( __neon_DdQnQm( 0xf2a00400, (Qn), (Qm)) )
|
|
#define vaddhn_u16(Qn, Qm) ( __neon_DdQnQm( 0xf2800400, (Qn), (Qm)) )
|
|
#define vaddhn_u32(Qn, Qm) ( __neon_DdQnQm( 0xf2900400, (Qn), (Qm)) )
|
|
#define vaddhn_u64(Qn, Qm) ( __neon_DdQnQm( 0xf2a00400, (Qn), (Qm)) )
|
|
#define vraddhn_s16(Qn, Qm) ( __neon_DdQnQm( 0xf3800400, (Qn), (Qm)) )
|
|
#define vraddhn_s32(Qn, Qm) ( __neon_DdQnQm( 0xf3900400, (Qn), (Qm)) )
|
|
#define vraddhn_s64(Qn, Qm) ( __neon_DdQnQm( 0xf3a00400, (Qn), (Qm)) )
|
|
#define vraddhn_u16(Qn, Qm) ( __neon_DdQnQm( 0xf3800400, (Qn), (Qm)) )
|
|
#define vraddhn_u32(Qn, Qm) ( __neon_DdQnQm( 0xf3900400, (Qn), (Qm)) )
|
|
#define vraddhn_u64(Qn, Qm) ( __neon_DdQnQm( 0xf3a00400, (Qn), (Qm)) )
|
|
|
|
// VADDL, VADDW
|
|
#define vaddl_s16(Dn, Dm) ( __neon_QdDnDm( 0xf2900000, (Dn), (Dm)) )
|
|
#define vaddl_s32(Dn, Dm) ( __neon_QdDnDm( 0xf2a00000, (Dn), (Dm)) )
|
|
#define vaddl_s8(Dn, Dm) ( __neon_QdDnDm( 0xf2800000, (Dn), (Dm)) )
|
|
#define vaddl_u16(Dn, Dm) ( __neon_QdDnDm( 0xf3900000, (Dn), (Dm)) )
|
|
#define vaddl_u32(Dn, Dm) ( __neon_QdDnDm( 0xf3a00000, (Dn), (Dm)) )
|
|
#define vaddl_u8(Dn, Dm) ( __neon_QdDnDm( 0xf3800000, (Dn), (Dm)) )
|
|
#define vaddw_s16(Qn, Dm) ( __neon_QdQnDm( 0xf2900100, (Qn), (Dm)) )
|
|
#define vaddw_s32(Qn, Dm) ( __neon_QdQnDm( 0xf2a00100, (Qn), (Dm)) )
|
|
#define vaddw_s8(Qn, Dm) ( __neon_QdQnDm( 0xf2800100, (Qn), (Dm)) )
|
|
#define vaddw_u16(Qn, Dm) ( __neon_QdQnDm( 0xf3900100, (Qn), (Dm)) )
|
|
#define vaddw_u32(Qn, Dm) ( __neon_QdQnDm( 0xf3a00100, (Qn), (Dm)) )
|
|
#define vaddw_u8(Qn, Dm) ( __neon_QdQnDm( 0xf3800100, (Qn), (Dm)) )
|
|
|
|
// VAND, VORR
|
|
#define vand_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_u64(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vand_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2000110, (Dn), (Dm)) )
|
|
#define vorr_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_u64(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vorr_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2200110, (Dn), (Dm)) )
|
|
#define vandq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vandq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2000150, (Qn), (Qm)) )
|
|
#define vorrq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
#define vorrq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2200150, (Qn), (Qm)) )
|
|
|
|
// VBIF, VBIT, VBSL
|
|
#define vbif_f32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_p16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_p8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_s64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_u64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbif_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3300110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_f32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_p16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_p8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_s64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_u64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbit_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_f32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_p16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_p8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_s64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_u64(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbsl_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100110, (Dd), (Dn), (Dm)) )
|
|
#define vbifq_f32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_p16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_p8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_s64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_u64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbifq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3300150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_f32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_p16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_p8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_s64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_u64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbitq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_f32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_p16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_p8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_s64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_u64(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
#define vbslq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100150, (Qd), (Qn), (Qm)) )
|
|
|
|
// VCEQ (immediate #0)
|
|
#define vceq_z_f32_ex(Dm) ( __neon_DdDm( 0xf3b90500, (Dm)) )
|
|
#define vceq_z_s16_ex(Dm) ( __neon_DdDm( 0xf3b50100, (Dm)) )
|
|
#define vceq_z_s32_ex(Dm) ( __neon_DdDm( 0xf3b90100, (Dm)) )
|
|
#define vceq_z_s8_ex(Dm) ( __neon_DdDm( 0xf3b10100, (Dm)) )
|
|
#define vceq_z_u16_ex(Dm) ( __neon_DdDm( 0xf3b50100, (Dm)) )
|
|
#define vceq_z_u32_ex(Dm) ( __neon_DdDm( 0xf3b90100, (Dm)) )
|
|
#define vceq_z_u8_ex(Dm) ( __neon_DdDm( 0xf3b10100, (Dm)) )
|
|
#define vceqq_z_f32_ex(Qm) ( __neon_QdQm( 0xf3b90540, (Qm)) )
|
|
#define vceqq_z_s16_ex(Qm) ( __neon_QdQm( 0xf3b50140, (Qm)) )
|
|
#define vceqq_z_s32_ex(Qm) ( __neon_QdQm( 0xf3b90140, (Qm)) )
|
|
#define vceqq_z_s8_ex(Qm) ( __neon_QdQm( 0xf3b10140, (Qm)) )
|
|
#define vceqq_z_u16_ex(Qm) ( __neon_QdQm( 0xf3b50140, (Qm)) )
|
|
#define vceqq_z_u32_ex(Qm) ( __neon_QdQm( 0xf3b90140, (Qm)) )
|
|
#define vceqq_z_u8_ex(Qm) ( __neon_QdQm( 0xf3b10140, (Qm)) )
|
|
|
|
// VCEQ (register)
|
|
#define vceq_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2000e00, (Dn), (Dm)) )
|
|
#define vceq_p8(Dn, Dm) ( __neon_DdDnDm( 0xf3000810, (Dn), (Dm)) )
|
|
#define vceq_s16(Dn, Dm) ( __neon_DdDnDm( 0xf3100810, (Dn), (Dm)) )
|
|
#define vceq_s32(Dn, Dm) ( __neon_DdDnDm( 0xf3200810, (Dn), (Dm)) )
|
|
#define vceq_s8(Dn, Dm) ( __neon_DdDnDm( 0xf3000810, (Dn), (Dm)) )
|
|
#define vceq_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100810, (Dn), (Dm)) )
|
|
#define vceq_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200810, (Dn), (Dm)) )
|
|
#define vceq_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000810, (Dn), (Dm)) )
|
|
#define vceqq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2000e40, (Qn), (Qm)) )
|
|
#define vceqq_p8(Qn, Qm) ( __neon_QdQnQm( 0xf3000850, (Qn), (Qm)) )
|
|
#define vceqq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf3100850, (Qn), (Qm)) )
|
|
#define vceqq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3200850, (Qn), (Qm)) )
|
|
#define vceqq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf3000850, (Qn), (Qm)) )
|
|
#define vceqq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100850, (Qn), (Qm)) )
|
|
#define vceqq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200850, (Qn), (Qm)) )
|
|
#define vceqq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000850, (Qn), (Qm)) )
|
|
|
|
// VCGE (immediate #0)
|
|
#define vcge_z_f32_ex(Dm) ( __neon_DdDm( 0xf3b90480, (Dm)) )
|
|
#define vcge_z_s16_ex(Dm) ( __neon_DdDm( 0xf3b50080, (Dm)) )
|
|
#define vcge_z_s32_ex(Dm) ( __neon_DdDm( 0xf3b90080, (Dm)) )
|
|
#define vcge_z_s8_ex(Dm) ( __neon_DdDm( 0xf3b10080, (Dm)) )
|
|
#define vcgeq_z_f32_ex(Qm) ( __neon_QdQm( 0xf3b904c0, (Qm)) )
|
|
#define vcgeq_z_s16_ex(Qm) ( __neon_QdQm( 0xf3b500c0, (Qm)) )
|
|
#define vcgeq_z_s32_ex(Qm) ( __neon_QdQm( 0xf3b900c0, (Qm)) )
|
|
#define vcgeq_z_s8_ex(Qm) ( __neon_QdQm( 0xf3b100c0, (Qm)) )
|
|
|
|
// VCGE, VCLE (register)
|
|
#define vcge_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000e00, (Dn), (Dm)) )
|
|
#define vcge_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100310, (Dn), (Dm)) )
|
|
#define vcge_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200310, (Dn), (Dm)) )
|
|
#define vcge_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000310, (Dn), (Dm)) )
|
|
#define vcge_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100310, (Dn), (Dm)) )
|
|
#define vcge_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200310, (Dn), (Dm)) )
|
|
#define vcge_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000310, (Dn), (Dm)) )
|
|
#define vcle_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000e00, (Dm), (Dn)) )
|
|
#define vcle_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100310, (Dm), (Dn)) )
|
|
#define vcle_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200310, (Dm), (Dn)) )
|
|
#define vcle_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000310, (Dm), (Dn)) )
|
|
#define vcle_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100310, (Dm), (Dn)) )
|
|
#define vcle_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200310, (Dm), (Dn)) )
|
|
#define vcle_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000310, (Dm), (Dn)) )
|
|
#define vcgeq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000e40, (Qn), (Qm)) )
|
|
#define vcgeq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100350, (Qn), (Qm)) )
|
|
#define vcgeq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200350, (Qn), (Qm)) )
|
|
#define vcgeq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000350, (Qn), (Qm)) )
|
|
#define vcgeq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100350, (Qn), (Qm)) )
|
|
#define vcgeq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200350, (Qn), (Qm)) )
|
|
#define vcgeq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000350, (Qn), (Qm)) )
|
|
#define vcleq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000e40, (Qm), (Qn)) )
|
|
#define vcleq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100350, (Qm), (Qn)) )
|
|
#define vcleq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200350, (Qm), (Qn)) )
|
|
#define vcleq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000350, (Qm), (Qn)) )
|
|
#define vcleq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100350, (Qm), (Qn)) )
|
|
#define vcleq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200350, (Qm), (Qn)) )
|
|
#define vcleq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000350, (Qm), (Qn)) )
|
|
|
|
// VCGT (immediate #0)
|
|
#define vcgt_z_f32_ex(Dm) ( __neon_DdDm( 0xf3b90400, (Dm)) )
|
|
#define vcgt_z_s16_ex(Dm) ( __neon_DdDm( 0xf3b50000, (Dm)) )
|
|
#define vcgt_z_s32_ex(Dm) ( __neon_DdDm( 0xf3b90000, (Dm)) )
|
|
#define vcgt_z_s8_ex(Dm) ( __neon_DdDm( 0xf3b10000, (Dm)) )
|
|
#define vcgtq_z_f32_ex(Qm) ( __neon_QdQm( 0xf3b90440, (Qm)) )
|
|
#define vcgtq_z_s16_ex(Qm) ( __neon_QdQm( 0xf3b50040, (Qm)) )
|
|
#define vcgtq_z_s32_ex(Qm) ( __neon_QdQm( 0xf3b90040, (Qm)) )
|
|
#define vcgtq_z_s8_ex(Qm) ( __neon_QdQm( 0xf3b10040, (Qm)) )
|
|
|
|
// VCGT, VCLT (register)
|
|
#define vcgt_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200e00, (Dn), (Dm)) )
|
|
#define vcgt_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100300, (Dn), (Dm)) )
|
|
#define vcgt_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200300, (Dn), (Dm)) )
|
|
#define vcgt_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000300, (Dn), (Dm)) )
|
|
#define vcgt_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100300, (Dn), (Dm)) )
|
|
#define vcgt_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200300, (Dn), (Dm)) )
|
|
#define vcgt_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000300, (Dn), (Dm)) )
|
|
#define vclt_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200e00, (Dm), (Dn)) )
|
|
#define vclt_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100300, (Dm), (Dn)) )
|
|
#define vclt_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200300, (Dm), (Dn)) )
|
|
#define vclt_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000300, (Dm), (Dn)) )
|
|
#define vclt_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100300, (Dm), (Dn)) )
|
|
#define vclt_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200300, (Dm), (Dn)) )
|
|
#define vclt_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000300, (Dm), (Dn)) )
|
|
#define vcgtq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200e40, (Qn), (Qm)) )
|
|
#define vcgtq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100340, (Qn), (Qm)) )
|
|
#define vcgtq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200340, (Qn), (Qm)) )
|
|
#define vcgtq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000340, (Qn), (Qm)) )
|
|
#define vcgtq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100340, (Qn), (Qm)) )
|
|
#define vcgtq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200340, (Qn), (Qm)) )
|
|
#define vcgtq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000340, (Qn), (Qm)) )
|
|
#define vcltq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200e40, (Qm), (Qn)) )
|
|
#define vcltq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100340, (Qm), (Qn)) )
|
|
#define vcltq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200340, (Qm), (Qn)) )
|
|
#define vcltq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000340, (Qm), (Qn)) )
|
|
#define vcltq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100340, (Qm), (Qn)) )
|
|
#define vcltq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200340, (Qm), (Qn)) )
|
|
#define vcltq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000340, (Qm), (Qn)) )
|
|
|
|
// VCLE (immediate #0)
|
|
#define vcle_z_f32_ex(Dm) ( __neon_DdDm( 0xf3b90580, (Dm)) )
|
|
#define vcle_z_s16_ex(Dm) ( __neon_DdDm( 0xf3b50180, (Dm)) )
|
|
#define vcle_z_s32_ex(Dm) ( __neon_DdDm( 0xf3b90180, (Dm)) )
|
|
#define vcle_z_s8_ex(Dm) ( __neon_DdDm( 0xf3b10180, (Dm)) )
|
|
#define vcleq_z_f32_ex(Qm) ( __neon_QdQm( 0xf3b905c0, (Qm)) )
|
|
#define vcleq_z_s16_ex(Qm) ( __neon_QdQm( 0xf3b501c0, (Qm)) )
|
|
#define vcleq_z_s32_ex(Qm) ( __neon_QdQm( 0xf3b901c0, (Qm)) )
|
|
#define vcleq_z_s8_ex(Qm) ( __neon_QdQm( 0xf3b101c0, (Qm)) )
|
|
|
|
// VCLS, VCLZ
|
|
#define vcls_s16(Dm) ( __neon_DdDm( 0xf3b40400, (Dm)) )
|
|
#define vcls_s32(Dm) ( __neon_DdDm( 0xf3b80400, (Dm)) )
|
|
#define vcls_s8(Dm) ( __neon_DdDm( 0xf3b00400, (Dm)) )
|
|
#define vclz_s16(Dm) ( __neon_DdDm( 0xf3b40480, (Dm)) )
|
|
#define vclz_s32(Dm) ( __neon_DdDm( 0xf3b80480, (Dm)) )
|
|
#define vclz_s8(Dm) ( __neon_DdDm( 0xf3b00480, (Dm)) )
|
|
#define vclz_u16(Dm) ( __neon_DdDm( 0xf3b40480, (Dm)) )
|
|
#define vclz_u32(Dm) ( __neon_DdDm( 0xf3b80480, (Dm)) )
|
|
#define vclz_u8(Dm) ( __neon_DdDm( 0xf3b00480, (Dm)) )
|
|
#define vclsq_s16(Qm) ( __neon_QdQm( 0xf3b40440, (Qm)) )
|
|
#define vclsq_s32(Qm) ( __neon_QdQm( 0xf3b80440, (Qm)) )
|
|
#define vclsq_s8(Qm) ( __neon_QdQm( 0xf3b00440, (Qm)) )
|
|
#define vclzq_s16(Qm) ( __neon_QdQm( 0xf3b404c0, (Qm)) )
|
|
#define vclzq_s32(Qm) ( __neon_QdQm( 0xf3b804c0, (Qm)) )
|
|
#define vclzq_s8(Qm) ( __neon_QdQm( 0xf3b004c0, (Qm)) )
|
|
#define vclzq_u16(Qm) ( __neon_QdQm( 0xf3b404c0, (Qm)) )
|
|
#define vclzq_u32(Qm) ( __neon_QdQm( 0xf3b804c0, (Qm)) )
|
|
#define vclzq_u8(Qm) ( __neon_QdQm( 0xf3b004c0, (Qm)) )
|
|
|
|
// VCLT (immediate #0)
|
|
#define vclt_z_f32_ex(Dm) ( __neon_DdDm( 0xf3b90600, (Dm)) )
|
|
#define vclt_z_s16_ex(Dm) ( __neon_DdDm( 0xf3b50200, (Dm)) )
|
|
#define vclt_z_s32_ex(Dm) ( __neon_DdDm( 0xf3b90200, (Dm)) )
|
|
#define vclt_z_s8_ex(Dm) ( __neon_DdDm( 0xf3b10200, (Dm)) )
|
|
#define vcltq_z_f32_ex(Qm) ( __neon_QdQm( 0xf3b90640, (Qm)) )
|
|
#define vcltq_z_s16_ex(Qm) ( __neon_QdQm( 0xf3b50240, (Qm)) )
|
|
#define vcltq_z_s32_ex(Qm) ( __neon_QdQm( 0xf3b90240, (Qm)) )
|
|
#define vcltq_z_s8_ex(Qm) ( __neon_QdQm( 0xf3b10240, (Qm)) )
|
|
|
|
// VCNT
|
|
#define vcnt_p8(Dm) ( __neon_DdDm( 0xf3b00500, (Dm)) )
|
|
#define vcnt_s8(Dm) ( __neon_DdDm( 0xf3b00500, (Dm)) )
|
|
#define vcnt_u8(Dm) ( __neon_DdDm( 0xf3b00500, (Dm)) )
|
|
#define vcntq_p8(Qm) ( __neon_QdQm( 0xf3b00540, (Qm)) )
|
|
#define vcntq_s8(Qm) ( __neon_QdQm( 0xf3b00540, (Qm)) )
|
|
#define vcntq_u8(Qm) ( __neon_QdQm( 0xf3b00540, (Qm)) )
|
|
|
|
// VCOMBINE (combine 2x64bit into a 128bit register)
|
|
#define vcombine_f32(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_p16(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_p8(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_s16(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_s32(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_s64(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_s8(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_u16(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_u32(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_u64(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
#define vcombine_u8(Dn, Dm) ( __neon_QdDnDm_merge( 0x00000000, (Dn), (Dm)) )
|
|
|
|
// VCREATE (ARM core register pair to Neon 64bit register)
|
|
#define vcreate_f32(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_p16(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_p8(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_s16(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_s32(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_s64(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_s8(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_u16(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_u32(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_u64(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vcreate_u8(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
|
|
// VCVT (between floating-point and fixed-point)
|
|
#define vcvt_n_f32_s32(Dm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_DdDm( 0xf2800e10 | _NENC_21_16(64 - (fbits)), (Dm)) )
|
|
#define vcvt_n_f32_u32(Dm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_DdDm( 0xf3800e10 | _NENC_21_16(64 - (fbits)), (Dm)) )
|
|
#define vcvt_n_s32_f32(Dm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_DdDm( 0xf2800f10 | _NENC_21_16(64 - (fbits)), (Dm)) )
|
|
#define vcvt_n_u32_f32(Dm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_DdDm( 0xf3800f10 | _NENC_21_16(64 - (fbits)), (Dm)) )
|
|
#define vcvtq_n_f32_s32(Qm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_QdQm( 0xf2800e50 | _NENC_21_16(64 - (fbits)), (Qm)) )
|
|
#define vcvtq_n_f32_u32(Qm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_QdQm( 0xf3800e50 | _NENC_21_16(64 - (fbits)), (Qm)) )
|
|
#define vcvtq_n_s32_f32(Qm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_QdQm( 0xf2800f50 | _NENC_21_16(64 - (fbits)), (Qm)) )
|
|
#define vcvtq_n_u32_f32(Qm, fbits) ( __static_assert((fbits) >= 1 && (fbits) <= 32, "invalid fbits value"), __neon_QdQm( 0xf3800f50 | _NENC_21_16(64 - (fbits)), (Qm)) )
|
|
|
|
// VCVT (between floating-point and integer with directed rounding)
|
|
#define vcvta_s32_f32(Dm) ( __neon_DdDm( 0xf3bb0000, (Dm)) )
|
|
#define vcvta_u32_f32(Dm) ( __neon_DdDm( 0xf3bb0080, (Dm)) )
|
|
#define vcvtm_s32_f32(Dm) ( __neon_DdDm( 0xf3bb0300, (Dm)) )
|
|
#define vcvtm_u32_f32(Dm) ( __neon_DdDm( 0xf3bb0380, (Dm)) )
|
|
#define vcvtn_s32_f32(Dm) ( __neon_DdDm( 0xf3bb0100, (Dm)) )
|
|
#define vcvtn_u32_f32(Dm) ( __neon_DdDm( 0xf3bb0180, (Dm)) )
|
|
#define vcvtp_s32_f32(Dm) ( __neon_DdDm( 0xf3bb0200, (Dm)) )
|
|
#define vcvtp_u32_f32(Dm) ( __neon_DdDm( 0xf3bb0280, (Dm)) )
|
|
#define vcvtaq_s32_f32(Qm) ( __neon_QdQm( 0xf3bb0040, (Qm)) )
|
|
#define vcvtaq_u32_f32(Qm) ( __neon_QdQm( 0xf3bb00c0, (Qm)) )
|
|
#define vcvtmq_s32_f32(Qm) ( __neon_QdQm( 0xf3bb0340, (Qm)) )
|
|
#define vcvtmq_u32_f32(Qm) ( __neon_QdQm( 0xf3bb03c0, (Qm)) )
|
|
#define vcvtnq_s32_f32(Qm) ( __neon_QdQm( 0xf3bb0140, (Qm)) )
|
|
#define vcvtnq_u32_f32(Qm) ( __neon_QdQm( 0xf3bb01c0, (Qm)) )
|
|
#define vcvtpq_s32_f32(Qm) ( __neon_QdQm( 0xf3bb0240, (Qm)) )
|
|
#define vcvtpq_u32_f32(Qm) ( __neon_QdQm( 0xf3bb02c0, (Qm)) )
|
|
|
|
// VCVT (between floating-point and integer)
|
|
#define vcvt_f32_s32(Dm) ( __neon_DdDm( 0xf3bb0600, (Dm)) )
|
|
#define vcvt_f32_u32(Dm) ( __neon_DdDm( 0xf3bb0680, (Dm)) )
|
|
#define vcvt_s32_f32(Dm) ( __neon_DdDm( 0xf3bb0700, (Dm)) )
|
|
#define vcvt_u32_f32(Dm) ( __neon_DdDm( 0xf3bb0780, (Dm)) )
|
|
#define vcvtq_f32_s32(Qm) ( __neon_QdQm( 0xf3bb0640, (Qm)) )
|
|
#define vcvtq_f32_u32(Qm) ( __neon_QdQm( 0xf3bb06c0, (Qm)) )
|
|
#define vcvtq_s32_f32(Qm) ( __neon_QdQm( 0xf3bb0740, (Qm)) )
|
|
#define vcvtq_u32_f32(Qm) ( __neon_QdQm( 0xf3bb07c0, (Qm)) )
|
|
|
|
// VDUP (scalar)
|
|
#define vdup_lane_f32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDm( 0xf3b40c00 | _NENC_19(lane), (Dm)) )
|
|
#define vdup_lane_p16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDm( 0xf3b20c00 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdup_lane_p8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdDm( 0xf3b10c00 | _NENC_19_17(lane), (Dm)) )
|
|
#define vdup_lane_s16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDm( 0xf3b20c00 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdup_lane_s32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDm( 0xf3b40c00 | _NENC_19(lane), (Dm)) )
|
|
#define vdup_lane_s8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdDm( 0xf3b10c00 | _NENC_19_17(lane), (Dm)) )
|
|
#define vdup_lane_u16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDm( 0xf3b20c00 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdup_lane_u32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDm( 0xf3b40c00 | _NENC_19(lane), (Dm)) )
|
|
#define vdup_lane_u8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdDm( 0xf3b10c00 | _NENC_19_17(lane), (Dm)) )
|
|
#define vdupq_lane_f32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDm( 0xf3b40c40 | _NENC_19(lane), (Dm)) )
|
|
#define vdupq_lane_p16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDm( 0xf3b20c40 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdupq_lane_p8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdDm( 0xf3b10c40 | _NENC_19_17(lane), (Dm)) )
|
|
#define vdupq_lane_s16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDm( 0xf3b20c40 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdupq_lane_s32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDm( 0xf3b40c40 | _NENC_19(lane), (Dm)) )
|
|
#define vdupq_lane_s8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdDm( 0xf3b10c40 | _NENC_19_17(lane), (Dm)) )
|
|
#define vdupq_lane_u16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDm( 0xf3b20c40 | _NENC_19_18(lane), (Dm)) )
|
|
#define vdupq_lane_u32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDm( 0xf3b40c40 | _NENC_19(lane), (Dm)) )
|
|
#define vdupq_lane_u8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdDm( 0xf3b10c40 | _NENC_19_17(lane), (Dm)) )
|
|
|
|
// VDUP, VMOV (ARM core register to Neon register)
|
|
#define vdup_n_f32(Ft) ( __neon_DdFt( 0xee800b10, (Ft)) )
|
|
#define vmov_n_f32(Ft) ( __neon_DdFt( 0xee800b10, (Ft)) )
|
|
#define vdup_n_p16(Rt) ( __neon_DdRt( 0xee800b30, __poly16ToInt32(Rt)) )
|
|
#define vdup_n_p8(Rt) ( __neon_DdRt( 0xeec00b10, __poly8ToInt32(Rt)) )
|
|
#define vdup_n_s16(Rt) ( __neon_DdRt( 0xee800b30, __int16ToInt32(Rt)) )
|
|
#define vdup_n_s32(Rt) ( __neon_DdRt( 0xee800b10, __int32ToInt32(Rt)) )
|
|
#define vdup_n_s8(Rt) ( __neon_DdRt( 0xeec00b10, __int8ToInt32(Rt)) )
|
|
#define vdup_n_u16(Rt) ( __neon_DdRt( 0xee800b30, __uint16ToInt32(Rt)) )
|
|
#define vdup_n_u32(Rt) ( __neon_DdRt( 0xee800b10, __uint32ToInt32(Rt)) )
|
|
#define vdup_n_u8(Rt) ( __neon_DdRt( 0xeec00b10, __uint8ToInt32(Rt)) )
|
|
#define vmov_n_p16(Rt) ( __neon_DdRt( 0xee800b30, __poly16ToInt32(Rt)) )
|
|
#define vmov_n_p8(Rt) ( __neon_DdRt( 0xeec00b10, __poly8ToInt32(Rt)) )
|
|
#define vmov_n_s16(Rt) ( __neon_DdRt( 0xee800b30, __int16ToInt32(Rt)) )
|
|
#define vmov_n_s32(Rt) ( __neon_DdRt( 0xee800b10, __int32ToInt32(Rt)) )
|
|
#define vmov_n_s8(Rt) ( __neon_DdRt( 0xeec00b10, __int8ToInt32(Rt)) )
|
|
#define vmov_n_u16(Rt) ( __neon_DdRt( 0xee800b30, __uint16ToInt32(Rt)) )
|
|
#define vmov_n_u32(Rt) ( __neon_DdRt( 0xee800b10, __uint32ToInt32(Rt)) )
|
|
#define vmov_n_u8(Rt) ( __neon_DdRt( 0xeec00b10, __uint8ToInt32(Rt)) )
|
|
#define vdupq_n_f32(Ft) ( __neon_QdFt( 0xeea00b10, (Ft)) )
|
|
#define vmovq_n_f32(Ft) ( __neon_QdFt( 0xeea00b10, (Ft)) )
|
|
#define vdupq_n_p16(Rt) ( __neon_QdRt( 0xeea00b30, __poly16ToInt32(Rt)) )
|
|
#define vdupq_n_p8(Rt) ( __neon_QdRt( 0xeee00b10, __poly8ToInt32(Rt)) )
|
|
#define vdupq_n_s16(Rt) ( __neon_QdRt( 0xeea00b30, __int16ToInt32(Rt)) )
|
|
#define vdupq_n_s32(Rt) ( __neon_QdRt( 0xeea00b10, __int32ToInt32(Rt)) )
|
|
#define vdupq_n_s8(Rt) ( __neon_QdRt( 0xeee00b10, __int8ToInt32(Rt)) )
|
|
#define vdupq_n_u16(Rt) ( __neon_QdRt( 0xeea00b30, __uint16ToInt32(Rt)) )
|
|
#define vdupq_n_u32(Rt) ( __neon_QdRt( 0xeea00b10, __uint32ToInt32(Rt)) )
|
|
#define vdupq_n_u8(Rt) ( __neon_QdRt( 0xeee00b10, __uint8ToInt32(Rt)) )
|
|
#define vmovq_n_p16(Rt) ( __neon_QdRt( 0xeea00b30, __poly16ToInt32(Rt)) )
|
|
#define vmovq_n_p8(Rt) ( __neon_QdRt( 0xeee00b10, __poly8ToInt32(Rt)) )
|
|
#define vmovq_n_s16(Rt) ( __neon_QdRt( 0xeea00b30, __int16ToInt32(Rt)) )
|
|
#define vmovq_n_s32(Rt) ( __neon_QdRt( 0xeea00b10, __int32ToInt32(Rt)) )
|
|
#define vmovq_n_s8(Rt) ( __neon_QdRt( 0xeee00b10, __int8ToInt32(Rt)) )
|
|
#define vmovq_n_u16(Rt) ( __neon_QdRt( 0xeea00b30, __uint16ToInt32(Rt)) )
|
|
#define vmovq_n_u32(Rt) ( __neon_QdRt( 0xeea00b10, __uint32ToInt32(Rt)) )
|
|
#define vmovq_n_u8(Rt) ( __neon_QdRt( 0xeee00b10, __uint8ToInt32(Rt)) )
|
|
|
|
// VDUP.64, VMOV.64 (ARM core register pair to Neon registers)
|
|
#define vdup_n_s64(R64t) ( __neon_DdRtRt2( 0xec400b10, __int64ToInt64(R64t)) )
|
|
#define vdup_n_u64(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vmov_n_s64(R64t) ( __neon_DdRtRt2( 0xec400b10, __int64ToInt64(R64t)) )
|
|
#define vmov_n_u64(R64t) ( __neon_DdRtRt2( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vdupq_n_s64(R64t) ( __neon_QdRtRt2_dup( 0xec400b10, __int64ToInt64(R64t)) )
|
|
#define vdupq_n_u64(R64t) ( __neon_QdRtRt2_dup( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
#define vmovq_n_s64(R64t) ( __neon_QdRtRt2_dup( 0xec400b10, __int64ToInt64(R64t)) )
|
|
#define vmovq_n_u64(R64t) ( __neon_QdRtRt2_dup( 0xec400b10, __uint64ToInt64(R64t)) )
|
|
|
|
// VEOR, VBIC, VORN
|
|
#define vbic_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_u64(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define vbic_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2100110, (Dn), (Dm)) )
|
|
#define veor_s16(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_s32(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_s64(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_s8(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define veor_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000110, (Dn), (Dm)) )
|
|
#define vorn_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_u64(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vorn_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2300110, (Dn), (Dm)) )
|
|
#define vbicq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define vbicq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2100150, (Qn), (Qm)) )
|
|
#define veorq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define veorq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000150, (Qn), (Qm)) )
|
|
#define vornq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
#define vornq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2300150, (Qn), (Qm)) )
|
|
|
|
// VEXT
|
|
#define vext_f32(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 2, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 4), (Dn), (Dm)) )
|
|
#define vext_p16(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 2), (Dn), (Dm)) )
|
|
#define vext_p8(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8(pos), (Dn), (Dm)) )
|
|
#define vext_s16(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 2), (Dn), (Dm)) )
|
|
#define vext_s32(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 2, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 4), (Dn), (Dm)) )
|
|
#define vext_s64(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 1, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 8), (Dn), (Dm)) )
|
|
#define vext_s8(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8(pos), (Dn), (Dm)) )
|
|
#define vext_u16(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 2), (Dn), (Dm)) )
|
|
#define vext_u32(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 2, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 4), (Dn), (Dm)) )
|
|
#define vext_u64(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 1, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8((pos) * 8), (Dn), (Dm)) )
|
|
#define vext_u8(Dn, Dm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_DdDnDm( 0xf2b00000 | _NENC_11_8(pos), (Dn), (Dm)) )
|
|
#define vextq_f32(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 4), (Qn), (Qm)) )
|
|
#define vextq_p16(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 2), (Qn), (Qm)) )
|
|
#define vextq_p8(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 16, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8(pos), (Qn), (Qm)) )
|
|
#define vextq_s16(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 2), (Qn), (Qm)) )
|
|
#define vextq_s32(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 4), (Qn), (Qm)) )
|
|
#define vextq_s64(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 2, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 8), (Qn), (Qm)) )
|
|
#define vextq_s8(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 16, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8(pos), (Qn), (Qm)) )
|
|
#define vextq_u16(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 8, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 2), (Qn), (Qm)) )
|
|
#define vextq_u32(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 4, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 4), (Qn), (Qm)) )
|
|
#define vextq_u64(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 2, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8((pos) * 8), (Qn), (Qm)) )
|
|
#define vextq_u8(Qn, Qm, pos) ( __static_assert((pos) >= 0 && (pos) < 16, "invalid position value"), __neon_QdQnQm( 0xf2b00040 | _NENC_11_8(pos), (Qn), (Qm)) )
|
|
|
|
// VGET (access the 64bit high/low part of a 128bit register)
|
|
#define vget_high_f32(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_p16(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_p8(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_s16(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_s32(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_s64(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_s8(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_u16(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_u32(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_u64(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_high_u8(Qm) ( __neon_DdQm_high( 0x00000000, (Qm)) )
|
|
#define vget_low_f32(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_p16(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_p8(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_s16(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_s32(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_s64(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_s8(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_u16(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_u32(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_u64(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
#define vget_low_u8(Qm) ( __neon_DdQm_low( 0x00000000, (Qm)) )
|
|
|
|
// VHADD, VRHADD, VHSUB
|
|
#define vhadd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100000, (Dn), (Dm)) )
|
|
#define vhadd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200000, (Dn), (Dm)) )
|
|
#define vhadd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000000, (Dn), (Dm)) )
|
|
#define vhadd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100000, (Dn), (Dm)) )
|
|
#define vhadd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200000, (Dn), (Dm)) )
|
|
#define vhadd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000000, (Dn), (Dm)) )
|
|
#define vhsub_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100200, (Dn), (Dm)) )
|
|
#define vhsub_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200200, (Dn), (Dm)) )
|
|
#define vhsub_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000200, (Dn), (Dm)) )
|
|
#define vhsub_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100200, (Dn), (Dm)) )
|
|
#define vhsub_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200200, (Dn), (Dm)) )
|
|
#define vhsub_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000200, (Dn), (Dm)) )
|
|
#define vrhadd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100100, (Dn), (Dm)) )
|
|
#define vrhadd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200100, (Dn), (Dm)) )
|
|
#define vrhadd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000100, (Dn), (Dm)) )
|
|
#define vrhadd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100100, (Dn), (Dm)) )
|
|
#define vrhadd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200100, (Dn), (Dm)) )
|
|
#define vrhadd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000100, (Dn), (Dm)) )
|
|
#define vhaddq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100040, (Qn), (Qm)) )
|
|
#define vhaddq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200040, (Qn), (Qm)) )
|
|
#define vhaddq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000040, (Qn), (Qm)) )
|
|
#define vhaddq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100040, (Qn), (Qm)) )
|
|
#define vhaddq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200040, (Qn), (Qm)) )
|
|
#define vhaddq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000040, (Qn), (Qm)) )
|
|
#define vhsubq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100240, (Qn), (Qm)) )
|
|
#define vhsubq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200240, (Qn), (Qm)) )
|
|
#define vhsubq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000240, (Qn), (Qm)) )
|
|
#define vhsubq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100240, (Qn), (Qm)) )
|
|
#define vhsubq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200240, (Qn), (Qm)) )
|
|
#define vhsubq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000240, (Qn), (Qm)) )
|
|
#define vrhaddq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100140, (Qn), (Qm)) )
|
|
#define vrhaddq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200140, (Qn), (Qm)) )
|
|
#define vrhaddq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000140, (Qn), (Qm)) )
|
|
#define vrhaddq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100140, (Qn), (Qm)) )
|
|
#define vrhaddq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200140, (Qn), (Qm)) )
|
|
#define vrhaddq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000140, (Qn), (Qm)) )
|
|
|
|
// VLD1 (multiple single elements)
|
|
#define vld1_f32(pcD) ( __neon_D1Adr( 0xf420078f, __float32ToN64_c(pcD)) )
|
|
#define vld1_p16(pcD) ( __neon_D1Adr( 0xf420074f, __poly16ToN64_c(pcD)) )
|
|
#define vld1_p8(pcD) ( __neon_D1Adr( 0xf420070f, __poly8ToN64_c(pcD)) )
|
|
#define vld1_s16(pcD) ( __neon_D1Adr( 0xf420074f, __int16ToN64_c(pcD)) )
|
|
#define vld1_s32(pcD) ( __neon_D1Adr( 0xf420078f, __int32ToN64_c(pcD)) )
|
|
#define vld1_s64(pcD) ( __neon_D1Adr( 0xf42007cf, __int64ToN64_c(pcD)) )
|
|
#define vld1_s8(pcD) ( __neon_D1Adr( 0xf420070f, __int8ToN64_c(pcD)) )
|
|
#define vld1_u16(pcD) ( __neon_D1Adr( 0xf420074f, __uint16ToN64_c(pcD)) )
|
|
#define vld1_u32(pcD) ( __neon_D1Adr( 0xf420078f, __uint32ToN64_c(pcD)) )
|
|
#define vld1_u64(pcD) ( __neon_D1Adr( 0xf42007cf, __uint64ToN64_c(pcD)) )
|
|
#define vld1_u8(pcD) ( __neon_D1Adr( 0xf420070f, __uint8ToN64_c(pcD)) )
|
|
#define vld1_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420078f | _NENC_5_4(_NEON_ALIGN64(align)), __float32ToN64_c(pcD)) )
|
|
#define vld1_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420074f | _NENC_5_4(_NEON_ALIGN64(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld1_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420070f | _NENC_5_4(_NEON_ALIGN64(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld1_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420074f | _NENC_5_4(_NEON_ALIGN64(align)), __int16ToN64_c(pcD)) )
|
|
#define vld1_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420078f | _NENC_5_4(_NEON_ALIGN64(align)), __int32ToN64_c(pcD)) )
|
|
#define vld1_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf42007cf | _NENC_5_4(_NEON_ALIGN64(align)), __int64ToN64_c(pcD)) )
|
|
#define vld1_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420070f | _NENC_5_4(_NEON_ALIGN64(align)), __int8ToN64_c(pcD)) )
|
|
#define vld1_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420074f | _NENC_5_4(_NEON_ALIGN64(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld1_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420078f | _NENC_5_4(_NEON_ALIGN64(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld1_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf42007cf | _NENC_5_4(_NEON_ALIGN64(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld1_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_D1Adr( 0xf420070f | _NENC_5_4(_NEON_ALIGN64(align)), __uint8ToN64_c(pcD)) )
|
|
#define vld1q_f32(pcD) ( __neon_Q1Adr( 0xf4200a8f, __float32ToN64_c(pcD)) )
|
|
#define vld1q_p16(pcD) ( __neon_Q1Adr( 0xf4200a4f, __poly16ToN64_c(pcD)) )
|
|
#define vld1q_p8(pcD) ( __neon_Q1Adr( 0xf4200a0f, __poly8ToN64_c(pcD)) )
|
|
#define vld1q_s16(pcD) ( __neon_Q1Adr( 0xf4200a4f, __int16ToN64_c(pcD)) )
|
|
#define vld1q_s32(pcD) ( __neon_Q1Adr( 0xf4200a8f, __int32ToN64_c(pcD)) )
|
|
#define vld1q_s64(pcD) ( __neon_Q1Adr( 0xf4200acf, __int64ToN64_c(pcD)) )
|
|
#define vld1q_s8(pcD) ( __neon_Q1Adr( 0xf4200a0f, __int8ToN64_c(pcD)) )
|
|
#define vld1q_u16(pcD) ( __neon_Q1Adr( 0xf4200a4f, __uint16ToN64_c(pcD)) )
|
|
#define vld1q_u32(pcD) ( __neon_Q1Adr( 0xf4200a8f, __uint32ToN64_c(pcD)) )
|
|
#define vld1q_u64(pcD) ( __neon_Q1Adr( 0xf4200acf, __uint64ToN64_c(pcD)) )
|
|
#define vld1q_u8(pcD) ( __neon_Q1Adr( 0xf4200a0f, __uint8ToN64_c(pcD)) )
|
|
#define vld1q_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64_c(pcD)) )
|
|
#define vld1q_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld1q_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld1q_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64_c(pcD)) )
|
|
#define vld1q_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64_c(pcD)) )
|
|
#define vld1q_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __int64ToN64_c(pcD)) )
|
|
#define vld1q_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64_c(pcD)) )
|
|
#define vld1q_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld1q_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld1q_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld1q_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4200a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD1 (single element to all lanes)
|
|
#define vld1_dup_f32(pcD) ( __neon_D1Adr( 0xf4a00c8f, __float32ToN64_c(pcD)) )
|
|
#define vld1_dup_p16(pcD) ( __neon_D1Adr( 0xf4a00c4f, __poly16ToN64_c(pcD)) )
|
|
#define vld1_dup_p8(pcD) ( __neon_D1Adr( 0xf4a00c0f, __poly8ToN64_c(pcD)) )
|
|
#define vld1_dup_s16(pcD) ( __neon_D1Adr( 0xf4a00c4f, __int16ToN64_c(pcD)) )
|
|
#define vld1_dup_s32(pcD) ( __neon_D1Adr( 0xf4a00c8f, __int32ToN64_c(pcD)) )
|
|
#define vld1_dup_s8(pcD) ( __neon_D1Adr( 0xf4a00c0f, __int8ToN64_c(pcD)) )
|
|
#define vld1_dup_u16(pcD) ( __neon_D1Adr( 0xf4a00c4f, __uint16ToN64_c(pcD)) )
|
|
#define vld1_dup_u32(pcD) ( __neon_D1Adr( 0xf4a00c8f, __uint32ToN64_c(pcD)) )
|
|
#define vld1_dup_u8(pcD) ( __neon_D1Adr( 0xf4a00c0f, __uint8ToN64_c(pcD)) )
|
|
#define vld1q_dup_f32(pcD) ( __neon_Q1Adr( 0xf4a00caf, __float32ToN64_c(pcD)) )
|
|
#define vld1q_dup_p16(pcD) ( __neon_Q1Adr( 0xf4a00c6f, __poly16ToN64_c(pcD)) )
|
|
#define vld1q_dup_p8(pcD) ( __neon_Q1Adr( 0xf4a00c2f, __poly8ToN64_c(pcD)) )
|
|
#define vld1q_dup_s16(pcD) ( __neon_Q1Adr( 0xf4a00c6f, __int16ToN64_c(pcD)) )
|
|
#define vld1q_dup_s32(pcD) ( __neon_Q1Adr( 0xf4a00caf, __int32ToN64_c(pcD)) )
|
|
#define vld1q_dup_s8(pcD) ( __neon_Q1Adr( 0xf4a00c2f, __int8ToN64_c(pcD)) )
|
|
#define vld1q_dup_u16(pcD) ( __neon_Q1Adr( 0xf4a00c6f, __uint16ToN64_c(pcD)) )
|
|
#define vld1q_dup_u32(pcD) ( __neon_Q1Adr( 0xf4a00caf, __uint32ToN64_c(pcD)) )
|
|
#define vld1q_dup_u8(pcD) ( __neon_Q1Adr( 0xf4a00c2f, __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD1 (single element to all lanes, aligned)
|
|
#define vld1_dup_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c8f | _NENC_4(_NEON_ALIGN32(align)), __float32ToN64_c(pcD)) )
|
|
#define vld1_dup_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c4f | _NENC_4(_NEON_ALIGN16(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld1_dup_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c4f | _NENC_4(_NEON_ALIGN16(align)), __int16ToN64_c(pcD)) )
|
|
#define vld1_dup_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c8f | _NENC_4(_NEON_ALIGN32(align)), __int32ToN64_c(pcD)) )
|
|
#define vld1_dup_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c4f | _NENC_4(_NEON_ALIGN16(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld1_dup_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr( 0xf4a00c8f | _NENC_4(_NEON_ALIGN32(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld1q_dup_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00caf | _NENC_4(_NEON_ALIGN32(align)), __float32ToN64_c(pcD)) )
|
|
#define vld1q_dup_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00c6f | _NENC_4(_NEON_ALIGN16(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld1q_dup_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00c6f | _NENC_4(_NEON_ALIGN16(align)), __int16ToN64_c(pcD)) )
|
|
#define vld1q_dup_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00caf | _NENC_4(_NEON_ALIGN32(align)), __int32ToN64_c(pcD)) )
|
|
#define vld1q_dup_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00c6f | _NENC_4(_NEON_ALIGN16(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld1q_dup_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr( 0xf4a00caf | _NENC_4(_NEON_ALIGN32(align)), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD1 (single element to one lane)
|
|
#define vld1_lane_f32(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane), (Dd), __float32ToN64_c(pcD)) )
|
|
#define vld1_lane_p16(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane), (Dd), __poly16ToN64_c(pcD)) )
|
|
#define vld1_lane_p8(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0000f | _NENC_7_5(lane), (Dd), __poly8ToN64_c(pcD)) )
|
|
#define vld1_lane_s16(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane), (Dd), __int16ToN64_c(pcD)) )
|
|
#define vld1_lane_s32(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane), (Dd), __int32ToN64_c(pcD)) )
|
|
#define vld1_lane_s8(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0000f | _NENC_7_5(lane), (Dd), __int8ToN64_c(pcD)) )
|
|
#define vld1_lane_u16(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane), (Dd), __uint16ToN64_c(pcD)) )
|
|
#define vld1_lane_u32(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane), (Dd), __uint32ToN64_c(pcD)) )
|
|
#define vld1_lane_u8(pcD, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_D1Adr_acc( 0xf4a0000f | _NENC_7_5(lane), (Dd), __uint8ToN64_c(pcD)) )
|
|
#define vld1q_lane_f32(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Qd), __float32ToN64_c(pcD)) )
|
|
#define vld1q_lane_p16(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Qd), __poly16ToN64_c(pcD)) )
|
|
#define vld1q_lane_p8(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), (Qd), __poly8ToN64_c(pcD)) )
|
|
#define vld1q_lane_s16(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Qd), __int16ToN64_c(pcD)) )
|
|
#define vld1q_lane_s32(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Qd), __int32ToN64_c(pcD)) )
|
|
#define vld1q_lane_s8(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), (Qd), __int8ToN64_c(pcD)) )
|
|
#define vld1q_lane_u16(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Qd), __uint16ToN64_c(pcD)) )
|
|
#define vld1q_lane_u32(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Qd), __uint32ToN64_c(pcD)) )
|
|
#define vld1q_lane_u8(pcD, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_Q1Adr_acc( 0xf4a0000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), (Qd), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD1 (single element to one lane, aligned)
|
|
#define vld1_lane_f32_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Dd), __float32ToN64_c(pcD)) )
|
|
#define vld1_lane_p16_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), (Dd), __poly16ToN64_c(pcD)) )
|
|
#define vld1_lane_s16_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), (Dd), __int16ToN64_c(pcD)) )
|
|
#define vld1_lane_s32_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Dd), __int32ToN64_c(pcD)) )
|
|
#define vld1_lane_u16_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), (Dd), __uint16ToN64_c(pcD)) )
|
|
#define vld1_lane_u32_ex(pcD, Dd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_D1Adr_acc( 0xf4a0080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Dd), __uint32ToN64_c(pcD)) )
|
|
#define vld1q_lane_f32_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Qd), __float32ToN64_c(pcD)) )
|
|
#define vld1q_lane_p16_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), (Qd), __poly16ToN64_c(pcD)) )
|
|
#define vld1q_lane_s16_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), (Qd), __int16ToN64_c(pcD)) )
|
|
#define vld1q_lane_s32_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Qd), __int32ToN64_c(pcD)) )
|
|
#define vld1q_lane_u16_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), (Qd), __uint16ToN64_c(pcD)) )
|
|
#define vld1q_lane_u32_ex(pcD, Qd, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Q1Adr_acc( 0xf4a0080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), (Qd), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD2 (multiple 2-element structures)
|
|
#define vld2_f32(pcD) ( __neon_Dx2Adr( 0xf420088f, __float32ToN64_c(pcD)) )
|
|
#define vld2_p16(pcD) ( __neon_Dx2Adr( 0xf420084f, __poly16ToN64_c(pcD)) )
|
|
#define vld2_p8(pcD) ( __neon_Dx2Adr( 0xf420080f, __poly8ToN64_c(pcD)) )
|
|
#define vld2_s16(pcD) ( __neon_Dx2Adr( 0xf420084f, __int16ToN64_c(pcD)) )
|
|
#define vld2_s32(pcD) ( __neon_Dx2Adr( 0xf420088f, __int32ToN64_c(pcD)) )
|
|
#define vld2_s8(pcD) ( __neon_Dx2Adr( 0xf420080f, __int8ToN64_c(pcD)) )
|
|
#define vld2_u16(pcD) ( __neon_Dx2Adr( 0xf420084f, __uint16ToN64_c(pcD)) )
|
|
#define vld2_u32(pcD) ( __neon_Dx2Adr( 0xf420088f, __uint32ToN64_c(pcD)) )
|
|
#define vld2_u8(pcD) ( __neon_Dx2Adr( 0xf420080f, __uint8ToN64_c(pcD)) )
|
|
#define vld2_s64(pcD) ( __neon_Dx2Adr( 0xf4200acf, __int64ToN64_c(pcD)) )
|
|
#define vld2_u64(pcD) ( __neon_Dx2Adr( 0xf4200acf, __uint64ToN64_c(pcD)) )
|
|
#define vld2_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __int64ToN64_c(pcD)) )
|
|
#define vld2_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld2_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64_c(pcD)) )
|
|
#define vld2_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld2_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld2_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64_c(pcD)) )
|
|
#define vld2_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64_c(pcD)) )
|
|
#define vld2_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64_c(pcD)) )
|
|
#define vld2_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld2_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld2_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf420080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64_c(pcD)) )
|
|
#define vld2q_f32(pcD) ( __neon_Qx2Adr( 0xf420098f, __float32ToN64_c(pcD)) )
|
|
#define vld2q_p16(pcD) ( __neon_Qx2Adr( 0xf420094f, __poly16ToN64_c(pcD)) )
|
|
#define vld2q_p8(pcD) ( __neon_Qx2Adr( 0xf420090f, __poly8ToN64_c(pcD)) )
|
|
#define vld2q_s16(pcD) ( __neon_Qx2Adr( 0xf420094f, __int16ToN64_c(pcD)) )
|
|
#define vld2q_s32(pcD) ( __neon_Qx2Adr( 0xf420098f, __int32ToN64_c(pcD)) )
|
|
#define vld2q_s8(pcD) ( __neon_Qx2Adr( 0xf420090f, __int8ToN64_c(pcD)) )
|
|
#define vld2q_u16(pcD) ( __neon_Qx2Adr( 0xf420094f, __uint16ToN64_c(pcD)) )
|
|
#define vld2q_u32(pcD) ( __neon_Qx2Adr( 0xf420098f, __uint32ToN64_c(pcD)) )
|
|
#define vld2q_u8(pcD) ( __neon_Qx2Adr( 0xf420090f, __uint8ToN64_c(pcD)) )
|
|
#define vld2q_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64_c(pcD)) )
|
|
#define vld2q_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld2q_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld2q_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64_c(pcD)) )
|
|
#define vld2q_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64_c(pcD)) )
|
|
#define vld2q_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64_c(pcD)) )
|
|
#define vld2q_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld2q_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld2q_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx2Adr( 0xf420090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD2 (single 2-element structure to all lanes)
|
|
#define vld2_dup_f32(pcD) ( __neon_Dx2Adr( 0xf4a00d8f, __float32ToN64_c(pcD)) )
|
|
#define vld2_dup_p16(pcD) ( __neon_Dx2Adr( 0xf4a00d4f, __poly16ToN64_c(pcD)) )
|
|
#define vld2_dup_p8(pcD) ( __neon_Dx2Adr( 0xf4a00d0f, __poly8ToN64_c(pcD)) )
|
|
#define vld2_dup_s16(pcD) ( __neon_Dx2Adr( 0xf4a00d4f, __int16ToN64_c(pcD)) )
|
|
#define vld2_dup_s32(pcD) ( __neon_Dx2Adr( 0xf4a00d8f, __int32ToN64_c(pcD)) )
|
|
#define vld2_dup_s8(pcD) ( __neon_Dx2Adr( 0xf4a00d0f, __int8ToN64_c(pcD)) )
|
|
#define vld2_dup_u16(pcD) ( __neon_Dx2Adr( 0xf4a00d4f, __uint16ToN64_c(pcD)) )
|
|
#define vld2_dup_u32(pcD) ( __neon_Dx2Adr( 0xf4a00d8f, __uint32ToN64_c(pcD)) )
|
|
#define vld2_dup_u8(pcD) ( __neon_Dx2Adr( 0xf4a00d0f, __uint8ToN64_c(pcD)) )
|
|
#define vld2_dup_s64(pcD) ( __neon_Dx2Adr( 0xf4200acf, __int64ToN64_c(pcD)) )
|
|
#define vld2_dup_u64(pcD) ( __neon_Dx2Adr( 0xf4200acf, __uint64ToN64_c(pcD)) )
|
|
|
|
// VLD2 (single 2-element structure to all lanes, aligned)
|
|
#define vld2_dup_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __int64ToN64_c(pcD)) )
|
|
#define vld2_dup_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4200acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld2_dup_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d8f | _NENC_4(_NEON_ALIGN64(align)), __float32ToN64_c(pcD)) )
|
|
#define vld2_dup_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d4f | _NENC_4(_NEON_ALIGN32(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld2_dup_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d0f | _NENC_4(_NEON_ALIGN16(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld2_dup_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d4f | _NENC_4(_NEON_ALIGN32(align)), __int16ToN64_c(pcD)) )
|
|
#define vld2_dup_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d8f | _NENC_4(_NEON_ALIGN64(align)), __int32ToN64_c(pcD)) )
|
|
#define vld2_dup_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d0f | _NENC_4(_NEON_ALIGN16(align)), __int8ToN64_c(pcD)) )
|
|
#define vld2_dup_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d4f | _NENC_4(_NEON_ALIGN32(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld2_dup_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d8f | _NENC_4(_NEON_ALIGN64(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld2_dup_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr( 0xf4a00d0f | _NENC_4(_NEON_ALIGN16(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD2 (single 2-element structure to one lane)
|
|
#define vld2_lane_f32(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane), (D2), __float32ToN64_c(pcD)) )
|
|
#define vld2_lane_p16(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane), (D2), __poly16ToN64_c(pcD)) )
|
|
#define vld2_lane_p8(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane), (D2), __poly8ToN64_c(pcD)) )
|
|
#define vld2_lane_s16(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane), (D2), __int16ToN64_c(pcD)) )
|
|
#define vld2_lane_s32(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane), (D2), __int32ToN64_c(pcD)) )
|
|
#define vld2_lane_s8(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane), (D2), __int8ToN64_c(pcD)) )
|
|
#define vld2_lane_u16(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane), (D2), __uint16ToN64_c(pcD)) )
|
|
#define vld2_lane_u32(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane), (D2), __uint32ToN64_c(pcD)) )
|
|
#define vld2_lane_u8(pcD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane), (D2), __uint8ToN64_c(pcD)) )
|
|
#define vld2q_lane_f32(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q2), __float32ToN64_c(pcD)) )
|
|
#define vld2q_lane_p16(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q2), __poly16ToN64_c(pcD)) )
|
|
#define vld2q_lane_s16(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q2), __int16ToN64_c(pcD)) )
|
|
#define vld2q_lane_s32(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q2), __int32ToN64_c(pcD)) )
|
|
#define vld2q_lane_u16(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q2), __uint16ToN64_c(pcD)) )
|
|
#define vld2q_lane_u32(pcD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q2), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD2 (single 2-element structure to one lane, aligned)
|
|
#define vld2_lane_f32_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), (D2), __float32ToN64_c(pcD)) )
|
|
#define vld2_lane_p16_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), (D2), __poly16ToN64_c(pcD)) )
|
|
#define vld2_lane_p8_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), (D2), __poly8ToN64_c(pcD)) )
|
|
#define vld2_lane_s16_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), (D2), __int16ToN64_c(pcD)) )
|
|
#define vld2_lane_s32_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), (D2), __int32ToN64_c(pcD)) )
|
|
#define vld2_lane_s8_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), (D2), __int8ToN64_c(pcD)) )
|
|
#define vld2_lane_u16_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), (D2), __uint16ToN64_c(pcD)) )
|
|
#define vld2_lane_u32_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), (D2), __uint32ToN64_c(pcD)) )
|
|
#define vld2_lane_u8_ex(pcD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_Dx2Adr_acc( 0xf4a0010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), (D2), __uint8ToN64_c(pcD)) )
|
|
#define vld2q_lane_f32_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q2), __float32ToN64_c(pcD)) )
|
|
#define vld2q_lane_p16_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), (Q2), __poly16ToN64_c(pcD)) )
|
|
#define vld2q_lane_s16_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), (Q2), __int16ToN64_c(pcD)) )
|
|
#define vld2q_lane_s32_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q2), __int32ToN64_c(pcD)) )
|
|
#define vld2q_lane_u16_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), (Q2), __uint16ToN64_c(pcD)) )
|
|
#define vld2q_lane_u32_ex(pcD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx2Adr_acc( 0xf4a0094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q2), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD3 (multiple 3-element structures)
|
|
#define vld3_f32(pcD) ( __neon_Dx3Adr( 0xf420048f, __float32ToN64_c(pcD)) )
|
|
#define vld3_p16(pcD) ( __neon_Dx3Adr( 0xf420044f, __poly16ToN64_c(pcD)) )
|
|
#define vld3_p8(pcD) ( __neon_Dx3Adr( 0xf420040f, __poly8ToN64_c(pcD)) )
|
|
#define vld3_s16(pcD) ( __neon_Dx3Adr( 0xf420044f, __int16ToN64_c(pcD)) )
|
|
#define vld3_s32(pcD) ( __neon_Dx3Adr( 0xf420048f, __int32ToN64_c(pcD)) )
|
|
#define vld3_s8(pcD) ( __neon_Dx3Adr( 0xf420040f, __int8ToN64_c(pcD)) )
|
|
#define vld3_u16(pcD) ( __neon_Dx3Adr( 0xf420044f, __uint16ToN64_c(pcD)) )
|
|
#define vld3_u32(pcD) ( __neon_Dx3Adr( 0xf420048f, __uint32ToN64_c(pcD)) )
|
|
#define vld3_u8(pcD) ( __neon_Dx3Adr( 0xf420040f, __uint8ToN64_c(pcD)) )
|
|
#define vld3_s64(pcD) ( __neon_Dx3Adr( 0xf42006cf, __int64ToN64_c(pcD)) )
|
|
#define vld3_u64(pcD) ( __neon_Dx3Adr( 0xf42006cf, __uint64ToN64_c(pcD)) )
|
|
#define vld3_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf42006cf | _NENC_5_4(_NEON_ALIGN64(align)), __int64ToN64_c(pcD)) )
|
|
#define vld3_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf42006cf | _NENC_5_4(_NEON_ALIGN64(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld3_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420048f | _NENC_5_4(_NEON_ALIGN64(align)), __float32ToN64_c(pcD)) )
|
|
#define vld3_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420044f | _NENC_5_4(_NEON_ALIGN64(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld3_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420040f | _NENC_5_4(_NEON_ALIGN64(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld3_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420044f | _NENC_5_4(_NEON_ALIGN64(align)), __int16ToN64_c(pcD)) )
|
|
#define vld3_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420048f | _NENC_5_4(_NEON_ALIGN64(align)), __int32ToN64_c(pcD)) )
|
|
#define vld3_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420040f | _NENC_5_4(_NEON_ALIGN64(align)), __int8ToN64_c(pcD)) )
|
|
#define vld3_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420044f | _NENC_5_4(_NEON_ALIGN64(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld3_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420048f | _NENC_5_4(_NEON_ALIGN64(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld3_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx3Adr( 0xf420040f | _NENC_5_4(_NEON_ALIGN64(align)), __uint8ToN64_c(pcD)) )
|
|
#define vld3q_f32(pcD) ( __neon_Qx3Adr( 0xf420058f, __float32ToN64_c(pcD)) )
|
|
#define vld3q_p16(pcD) ( __neon_Qx3Adr( 0xf420054f, __poly16ToN64_c(pcD)) )
|
|
#define vld3q_p8(pcD) ( __neon_Qx3Adr( 0xf420050f, __poly8ToN64_c(pcD)) )
|
|
#define vld3q_s16(pcD) ( __neon_Qx3Adr( 0xf420054f, __int16ToN64_c(pcD)) )
|
|
#define vld3q_s32(pcD) ( __neon_Qx3Adr( 0xf420058f, __int32ToN64_c(pcD)) )
|
|
#define vld3q_s8(pcD) ( __neon_Qx3Adr( 0xf420050f, __int8ToN64_c(pcD)) )
|
|
#define vld3q_u16(pcD) ( __neon_Qx3Adr( 0xf420054f, __uint16ToN64_c(pcD)) )
|
|
#define vld3q_u32(pcD) ( __neon_Qx3Adr( 0xf420058f, __uint32ToN64_c(pcD)) )
|
|
#define vld3q_u8(pcD) ( __neon_Qx3Adr( 0xf420050f, __uint8ToN64_c(pcD)) )
|
|
#define vld3q_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420058f | _NENC_5_4(_NEON_ALIGN64(align)), __float32ToN64_c(pcD)) )
|
|
#define vld3q_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420054f | _NENC_5_4(_NEON_ALIGN64(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld3q_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420050f | _NENC_5_4(_NEON_ALIGN64(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld3q_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420054f | _NENC_5_4(_NEON_ALIGN64(align)), __int16ToN64_c(pcD)) )
|
|
#define vld3q_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420058f | _NENC_5_4(_NEON_ALIGN64(align)), __int32ToN64_c(pcD)) )
|
|
#define vld3q_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420050f | _NENC_5_4(_NEON_ALIGN64(align)), __int8ToN64_c(pcD)) )
|
|
#define vld3q_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420054f | _NENC_5_4(_NEON_ALIGN64(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld3q_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420058f | _NENC_5_4(_NEON_ALIGN64(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld3q_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx3Adr( 0xf420050f | _NENC_5_4(_NEON_ALIGN64(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD3 (single 3-element structure to all lanes)
|
|
#define vld3_dup_f32(pcD) ( __neon_Dx3Adr( 0xf4a00e8f, __float32ToN64_c(pcD)) )
|
|
#define vld3_dup_p16(pcD) ( __neon_Dx3Adr( 0xf4a00e4f, __poly16ToN64_c(pcD)) )
|
|
#define vld3_dup_p8(pcD) ( __neon_Dx3Adr( 0xf4a00e0f, __poly8ToN64_c(pcD)) )
|
|
#define vld3_dup_s16(pcD) ( __neon_Dx3Adr( 0xf4a00e4f, __int16ToN64_c(pcD)) )
|
|
#define vld3_dup_s32(pcD) ( __neon_Dx3Adr( 0xf4a00e8f, __int32ToN64_c(pcD)) )
|
|
#define vld3_dup_s8(pcD) ( __neon_Dx3Adr( 0xf4a00e0f, __int8ToN64_c(pcD)) )
|
|
#define vld3_dup_u16(pcD) ( __neon_Dx3Adr( 0xf4a00e4f, __uint16ToN64_c(pcD)) )
|
|
#define vld3_dup_u32(pcD) ( __neon_Dx3Adr( 0xf4a00e8f, __uint32ToN64_c(pcD)) )
|
|
#define vld3_dup_u8(pcD) ( __neon_Dx3Adr( 0xf4a00e0f, __uint8ToN64_c(pcD)) )
|
|
#define vld3_dup_s64(pcD) ( __neon_Dx3Adr( 0xf42006cf, __int64ToN64_c(pcD)) )
|
|
#define vld3_dup_u64(pcD) ( __neon_Dx3Adr( 0xf42006cf, __uint64ToN64_c(pcD)) )
|
|
|
|
// VLD3 (single 3-element structure to one lane)
|
|
#define vld3_lane_f32(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a00a0f | _NENC_7(lane), (D3), __float32ToN64_c(pcD)) )
|
|
#define vld3_lane_p16(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0060f | _NENC_7_6(lane), (D3), __poly16ToN64_c(pcD)) )
|
|
#define vld3_lane_p8(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0020f | _NENC_7_5(lane), (D3), __poly8ToN64_c(pcD)) )
|
|
#define vld3_lane_s16(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0060f | _NENC_7_6(lane), (D3), __int16ToN64_c(pcD)) )
|
|
#define vld3_lane_s32(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a00a0f | _NENC_7(lane), (D3), __int32ToN64_c(pcD)) )
|
|
#define vld3_lane_s8(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0020f | _NENC_7_5(lane), (D3), __int8ToN64_c(pcD)) )
|
|
#define vld3_lane_u16(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0060f | _NENC_7_6(lane), (D3), __uint16ToN64_c(pcD)) )
|
|
#define vld3_lane_u32(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a00a0f | _NENC_7(lane), (D3), __uint32ToN64_c(pcD)) )
|
|
#define vld3_lane_u8(pcD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx3Adr_acc( 0xf4a0020f | _NENC_7_5(lane), (D3), __uint8ToN64_c(pcD)) )
|
|
#define vld3q_lane_f32(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a00a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q3), __float32ToN64_c(pcD)) )
|
|
#define vld3q_lane_p16(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a0062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q3), __poly16ToN64_c(pcD)) )
|
|
#define vld3q_lane_s16(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a0062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q3), __int16ToN64_c(pcD)) )
|
|
#define vld3q_lane_s32(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a00a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q3), __int32ToN64_c(pcD)) )
|
|
#define vld3q_lane_u16(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a0062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q3), __uint16ToN64_c(pcD)) )
|
|
#define vld3q_lane_u32(pcD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx3Adr_acc( 0xf4a00a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q3), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD4 (multiple 4-element structures)
|
|
#define vld4_f32(pcD) ( __neon_Dx4Adr( 0xf420008f, __float32ToN64_c(pcD)) )
|
|
#define vld4_p16(pcD) ( __neon_Dx4Adr( 0xf420004f, __poly16ToN64_c(pcD)) )
|
|
#define vld4_p8(pcD) ( __neon_Dx4Adr( 0xf420000f, __poly8ToN64_c(pcD)) )
|
|
#define vld4_s16(pcD) ( __neon_Dx4Adr( 0xf420004f, __int16ToN64_c(pcD)) )
|
|
#define vld4_s32(pcD) ( __neon_Dx4Adr( 0xf420008f, __int32ToN64_c(pcD)) )
|
|
#define vld4_s8(pcD) ( __neon_Dx4Adr( 0xf420000f, __int8ToN64_c(pcD)) )
|
|
#define vld4_u16(pcD) ( __neon_Dx4Adr( 0xf420004f, __uint16ToN64_c(pcD)) )
|
|
#define vld4_u32(pcD) ( __neon_Dx4Adr( 0xf420008f, __uint32ToN64_c(pcD)) )
|
|
#define vld4_u8(pcD) ( __neon_Dx4Adr( 0xf420000f, __uint8ToN64_c(pcD)) )
|
|
#define vld4_s64(pcD) ( __neon_Dx4Adr( 0xf42002cf, __int64ToN64_c(pcD)) )
|
|
#define vld4_u64(pcD) ( __neon_Dx4Adr( 0xf42002cf, __uint64ToN64_c(pcD)) )
|
|
#define vld4_s64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf42002cf | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int64ToN64_c(pcD)) )
|
|
#define vld4_u64_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf42002cf | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint64ToN64_c(pcD)) )
|
|
#define vld4_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __float32ToN64_c(pcD)) )
|
|
#define vld4_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld4_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld4_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int16ToN64_c(pcD)) )
|
|
#define vld4_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int32ToN64_c(pcD)) )
|
|
#define vld4_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int8ToN64_c(pcD)) )
|
|
#define vld4_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld4_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld4_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf420000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint8ToN64_c(pcD)) )
|
|
#define vld4q_f32(pcD) ( __neon_Qx4Adr( 0xf420018f, __float32ToN64_c(pcD)) )
|
|
#define vld4q_p16(pcD) ( __neon_Qx4Adr( 0xf420014f, __poly16ToN64_c(pcD)) )
|
|
#define vld4q_p8(pcD) ( __neon_Qx4Adr( 0xf420010f, __poly8ToN64_c(pcD)) )
|
|
#define vld4q_s16(pcD) ( __neon_Qx4Adr( 0xf420014f, __int16ToN64_c(pcD)) )
|
|
#define vld4q_s32(pcD) ( __neon_Qx4Adr( 0xf420018f, __int32ToN64_c(pcD)) )
|
|
#define vld4q_s8(pcD) ( __neon_Qx4Adr( 0xf420010f, __int8ToN64_c(pcD)) )
|
|
#define vld4q_u16(pcD) ( __neon_Qx4Adr( 0xf420014f, __uint16ToN64_c(pcD)) )
|
|
#define vld4q_u32(pcD) ( __neon_Qx4Adr( 0xf420018f, __uint32ToN64_c(pcD)) )
|
|
#define vld4q_u8(pcD) ( __neon_Qx4Adr( 0xf420010f, __uint8ToN64_c(pcD)) )
|
|
#define vld4q_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __float32ToN64_c(pcD)) )
|
|
#define vld4q_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld4q_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld4q_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int16ToN64_c(pcD)) )
|
|
#define vld4q_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int32ToN64_c(pcD)) )
|
|
#define vld4q_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int8ToN64_c(pcD)) )
|
|
#define vld4q_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld4q_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint32ToN64_c(pcD)) )
|
|
#define vld4q_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_Qx4Adr( 0xf420010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD4 (single 4-element structure to all lanes)
|
|
#define vld4_dup_f32(pcD) ( __neon_Dx4Adr( 0xf4a00f8f, __float32ToN64_c(pcD)) )
|
|
#define vld4_dup_p16(pcD) ( __neon_Dx4Adr( 0xf4a00f4f, __poly16ToN64_c(pcD)) )
|
|
#define vld4_dup_p8(pcD) ( __neon_Dx4Adr( 0xf4a00f0f, __poly8ToN64_c(pcD)) )
|
|
#define vld4_dup_s16(pcD) ( __neon_Dx4Adr( 0xf4a00f4f, __int16ToN64_c(pcD)) )
|
|
#define vld4_dup_s32(pcD) ( __neon_Dx4Adr( 0xf4a00f8f, __int32ToN64_c(pcD)) )
|
|
#define vld4_dup_s8(pcD) ( __neon_Dx4Adr( 0xf4a00f0f, __int8ToN64_c(pcD)) )
|
|
#define vld4_dup_u16(pcD) ( __neon_Dx4Adr( 0xf4a00f4f, __uint16ToN64_c(pcD)) )
|
|
#define vld4_dup_u32(pcD) ( __neon_Dx4Adr( 0xf4a00f8f, __uint32ToN64_c(pcD)) )
|
|
#define vld4_dup_u8(pcD) ( __neon_Dx4Adr( 0xf4a00f0f, __uint8ToN64_c(pcD)) )
|
|
#define vld4_dup_s64(pcD) ( __neon_Dx4Adr( 0xf42002cf, __int64ToN64_c(pcD)) )
|
|
#define vld4_dup_u64(pcD) ( __neon_Dx4Adr( 0xf42002cf, __uint64ToN64_c(pcD)) )
|
|
|
|
// VLD4 (single 4-element structure to all lanes, aligned)
|
|
#define vld4_dup_f32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_7_6(_NEON_ALIGN64_128(align) > 1 ? 3 : 2) | _NENC_4(_NEON_ALIGN64_128(align) > 0 ? 1 : 0), __float32ToN64_c(pcD)) )
|
|
#define vld4_dup_p16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f4f | _NENC_4(_NEON_ALIGN64(align)), __poly16ToN64_c(pcD)) )
|
|
#define vld4_dup_p8_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_4(_NEON_ALIGN32(align)), __poly8ToN64_c(pcD)) )
|
|
#define vld4_dup_s16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f4f | _NENC_4(_NEON_ALIGN64(align)), __int16ToN64_c(pcD)) )
|
|
#define vld4_dup_s32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_7_6(_NEON_ALIGN64_128(align) > 1 ? 3 : 2) | _NENC_4(_NEON_ALIGN64_128(align) > 0 ? 1 : 0), __int32ToN64_c(pcD)) )
|
|
#define vld4_dup_s8_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_4(_NEON_ALIGN32(align)), __int8ToN64_c(pcD)) )
|
|
#define vld4_dup_u16_ex(pcD, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f4f | _NENC_4(_NEON_ALIGN64(align)), __uint16ToN64_c(pcD)) )
|
|
#define vld4_dup_u32_ex(pcD, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_7_6(_NEON_ALIGN64_128(align) > 1 ? 3 : 2) | _NENC_4(_NEON_ALIGN64_128(align) > 0 ? 1 : 0), __uint32ToN64_c(pcD)) )
|
|
#define vld4_dup_u8_ex(pcD, align) ( __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr( 0xf4a00f0f | _NENC_4(_NEON_ALIGN32(align)), __uint8ToN64_c(pcD)) )
|
|
|
|
// VLD4 (single 4-element structure to one lane)
|
|
#define vld4_lane_f32(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane), (D4), __float32ToN64_c(pcD)) )
|
|
#define vld4_lane_p16(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane), (D4), __poly16ToN64_c(pcD)) )
|
|
#define vld4_lane_p8(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane), (D4), __poly8ToN64_c(pcD)) )
|
|
#define vld4_lane_s16(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane), (D4), __int16ToN64_c(pcD)) )
|
|
#define vld4_lane_s32(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane), (D4), __int32ToN64_c(pcD)) )
|
|
#define vld4_lane_s8(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane), (D4), __int8ToN64_c(pcD)) )
|
|
#define vld4_lane_u16(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane), (D4), __uint16ToN64_c(pcD)) )
|
|
#define vld4_lane_u32(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane), (D4), __uint32ToN64_c(pcD)) )
|
|
#define vld4_lane_u8(pcD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane), (D4), __uint8ToN64_c(pcD)) )
|
|
#define vld4q_lane_f32(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q4), __float32ToN64_c(pcD)) )
|
|
#define vld4q_lane_p16(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q4), __poly16ToN64_c(pcD)) )
|
|
#define vld4q_lane_s16(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q4), __int16ToN64_c(pcD)) )
|
|
#define vld4q_lane_s32(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q4), __int32ToN64_c(pcD)) )
|
|
#define vld4q_lane_u16(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), (Q4), __uint16ToN64_c(pcD)) )
|
|
#define vld4q_lane_u32(pcD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), (Q4), __uint32ToN64_c(pcD)) )
|
|
|
|
// VLD4 (single 4-element structure to one lane, aligned)
|
|
#define vld4_lane_f32_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), (D4), __float32ToN64_c(pcD)) )
|
|
#define vld4_lane_p16_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), (D4), __poly16ToN64_c(pcD)) )
|
|
#define vld4_lane_p8_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), (D4), __poly8ToN64_c(pcD)) )
|
|
#define vld4_lane_s16_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), (D4), __int16ToN64_c(pcD)) )
|
|
#define vld4_lane_s32_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), (D4), __int32ToN64_c(pcD)) )
|
|
#define vld4_lane_s8_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), (D4), __int8ToN64_c(pcD)) )
|
|
#define vld4_lane_u16_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), (D4), __uint16ToN64_c(pcD)) )
|
|
#define vld4_lane_u32_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a00b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), (D4), __uint32ToN64_c(pcD)) )
|
|
#define vld4_lane_u8_ex(pcD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_Dx4Adr_acc( 0xf4a0030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), (D4), __uint8ToN64_c(pcD)) )
|
|
#define vld4q_lane_f32_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), (Q4), __float32ToN64_c(pcD)) )
|
|
#define vld4q_lane_p16_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q4), __poly16ToN64_c(pcD)) )
|
|
#define vld4q_lane_s16_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q4), __int16ToN64_c(pcD)) )
|
|
#define vld4q_lane_s32_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), (Q4), __int32ToN64_c(pcD)) )
|
|
#define vld4q_lane_u16_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a0072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), (Q4), __uint16ToN64_c(pcD)) )
|
|
#define vld4q_lane_u32_ex(pcD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_Qx4Adr_acc( 0xf4a00b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), (Q4), __uint32ToN64_c(pcD)) )
|
|
|
|
// VMAX, VMIN (floating point)
|
|
#define vmax_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2000f00, (Dn), (Dm)) )
|
|
#define vmaxnm_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000f10, (Dn), (Dm)) )
|
|
#define vmin_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2200f00, (Dn), (Dm)) )
|
|
#define vminnm_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200f10, (Dn), (Dm)) )
|
|
#define vmaxq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2000f40, (Qn), (Qm)) )
|
|
#define vmaxnmq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000f50, (Qn), (Qm)) )
|
|
#define vminq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2200f40, (Qn), (Qm)) )
|
|
#define vminnmq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3200f50, (Qn), (Qm)) )
|
|
|
|
// VMAX, VMIN (integer)
|
|
#define vmax_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100600, (Dn), (Dm)) )
|
|
#define vmax_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200600, (Dn), (Dm)) )
|
|
#define vmax_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000600, (Dn), (Dm)) )
|
|
#define vmax_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100600, (Dn), (Dm)) )
|
|
#define vmax_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200600, (Dn), (Dm)) )
|
|
#define vmax_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000600, (Dn), (Dm)) )
|
|
#define vmin_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100610, (Dn), (Dm)) )
|
|
#define vmin_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200610, (Dn), (Dm)) )
|
|
#define vmin_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000610, (Dn), (Dm)) )
|
|
#define vmin_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100610, (Dn), (Dm)) )
|
|
#define vmin_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200610, (Dn), (Dm)) )
|
|
#define vmin_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000610, (Dn), (Dm)) )
|
|
#define vmaxq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100640, (Qn), (Qm)) )
|
|
#define vmaxq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200640, (Qn), (Qm)) )
|
|
#define vmaxq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000640, (Qn), (Qm)) )
|
|
#define vmaxq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100640, (Qn), (Qm)) )
|
|
#define vmaxq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200640, (Qn), (Qm)) )
|
|
#define vmaxq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000640, (Qn), (Qm)) )
|
|
#define vminq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100650, (Qn), (Qm)) )
|
|
#define vminq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200650, (Qn), (Qm)) )
|
|
#define vminq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000650, (Qn), (Qm)) )
|
|
#define vminq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100650, (Qn), (Qm)) )
|
|
#define vminq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200650, (Qn), (Qm)) )
|
|
#define vminq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000650, (Qn), (Qm)) )
|
|
|
|
// VMLA, VMLS (by scalar)
|
|
#define vmla_lane_f32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00140 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmla_lane_s16(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2900040 | _NENC_5x3(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmla_lane_s32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00040 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmla_lane_u16(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2900040 | _NENC_5x3(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmla_lane_u32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00040 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmls_lane_f32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00540 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmls_lane_s16(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2900440 | _NENC_5x3(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmls_lane_s32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00440 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmls_lane_u16(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2900440 | _NENC_5x3(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmls_lane_u32(Dd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx_acc( 0xf2a00440 | _NENC_5(lane), (Dd), (Dn), (Dm)) )
|
|
#define vmlaq_lane_f32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00140 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlaq_lane_s16(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3900040 | _NENC_5x3(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlaq_lane_s32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00040 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlaq_lane_u16(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3900040 | _NENC_5x3(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlaq_lane_u32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00040 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlsq_lane_f32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00540 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlsq_lane_s16(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3900440 | _NENC_5x3(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlsq_lane_s32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00440 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlsq_lane_u16(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3900440 | _NENC_5x3(lane), (Qd), (Qn), (Dm)) )
|
|
#define vmlsq_lane_u32(Qd, Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx_acc( 0xf3a00440 | _NENC_5(lane), (Qd), (Qn), (Dm)) )
|
|
|
|
// VMLA, VMLS (float, by scalar)
|
|
#define vmla_n_f32(Dd, Dn, Ft) ( __neon_DdDnFt_acc( 0xf2a00140, (Dd), (Dn), (Ft)) )
|
|
#define vmls_n_f32(Dd, Dn, Ft) ( __neon_DdDnFt_acc( 0xf2a00540, (Dd), (Dn), (Ft)) )
|
|
#define vmlaq_n_f32(Qd, Qn, Ft) ( __neon_QdQnFt_acc( 0xf3a00140, (Qd), (Qn), (Ft)) )
|
|
#define vmlsq_n_f32(Qd, Qn, Ft) ( __neon_QdQnFt_acc( 0xf3a00540, (Qd), (Qn), (Ft)) )
|
|
|
|
// VMLA, VMLS (floating point)
|
|
#define vmla_f32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2000d10, (Dd), (Dn), (Dm)) )
|
|
#define vmls_f32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2200d10, (Dd), (Dn), (Dm)) )
|
|
#define vmlaq_f32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2000d50, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_f32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2200d50, (Qd), (Qn), (Qm)) )
|
|
|
|
// VMLA, VMLS (integer)
|
|
#define vmla_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2100900, (Dd), (Dn), (Dm)) )
|
|
#define vmla_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2200900, (Dd), (Dn), (Dm)) )
|
|
#define vmla_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2000900, (Dd), (Dn), (Dm)) )
|
|
#define vmla_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2100900, (Dd), (Dn), (Dm)) )
|
|
#define vmla_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2200900, (Dd), (Dn), (Dm)) )
|
|
#define vmla_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf2000900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_s16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_s32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3000900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_u16(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3100900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_u32(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3200900, (Dd), (Dn), (Dm)) )
|
|
#define vmls_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3000900, (Dd), (Dn), (Dm)) )
|
|
#define vmlaq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2100940, (Qd), (Qn), (Qm)) )
|
|
#define vmlaq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2200940, (Qd), (Qn), (Qm)) )
|
|
#define vmlaq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2000940, (Qd), (Qn), (Qm)) )
|
|
#define vmlaq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2100940, (Qd), (Qn), (Qm)) )
|
|
#define vmlaq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2200940, (Qd), (Qn), (Qm)) )
|
|
#define vmlaq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf2000940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_s16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_s32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_s8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3000940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_u16(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3100940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_u32(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3200940, (Qd), (Qn), (Qm)) )
|
|
#define vmlsq_u8(Qd, Qn, Qm) ( __neon_QdQnQm_acc( 0xf3000940, (Qd), (Qn), (Qm)) )
|
|
|
|
// VMLAL, VMLSL
|
|
#define vmlal_s16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2900800, (Qd), (Dn), (Dm)) )
|
|
#define vmlal_s32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2a00800, (Qd), (Dn), (Dm)) )
|
|
#define vmlal_s8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2800800, (Qd), (Dn), (Dm)) )
|
|
#define vmlal_u16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3900800, (Qd), (Dn), (Dm)) )
|
|
#define vmlal_u32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3a00800, (Qd), (Dn), (Dm)) )
|
|
#define vmlal_u8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3800800, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_s16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2900a00, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_s32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2a00a00, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_s8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2800a00, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_u16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3900a00, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_u32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3a00a00, (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_u8(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf3800a00, (Qd), (Dn), (Dm)) )
|
|
|
|
// VMLAL, VMLSL (by scalar)
|
|
#define vmlal_lane_s16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2900240 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlal_lane_s32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2a00240 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlal_lane_u16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf3900240 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlal_lane_u32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf3a00240 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_lane_s16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2900640 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_lane_s32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2a00640 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_lane_u16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf3900640 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vmlsl_lane_u32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf3a00640 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
|
|
// VMOV (ARM core register to scalar)
|
|
#define vset_lane_f32(Ft, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdFt_acc( 0xee000b10 | _NENC_21(lane), (Dd), (Ft)) )
|
|
#define vset_lane_p16(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdRt_acc( 0xee000b30 | _NENC_21x6(lane), (Dd), __poly16ToInt32(Rt)) )
|
|
#define vset_lane_p8(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdRt_acc( 0xee400b10 | _NENC_21x6_5(lane), (Dd), __poly8ToInt32(Rt)) )
|
|
#define vset_lane_s16(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdRt_acc( 0xee000b30 | _NENC_21x6(lane), (Dd), __int16ToInt32(Rt)) )
|
|
#define vset_lane_s32(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdRt_acc( 0xee000b10 | _NENC_21(lane), (Dd), __int32ToInt32(Rt)) )
|
|
#define vset_lane_s8(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdRt_acc( 0xee400b10 | _NENC_21x6_5(lane), (Dd), __int8ToInt32(Rt)) )
|
|
#define vset_lane_u16(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdRt_acc( 0xee000b30 | _NENC_21x6(lane), (Dd), __uint16ToInt32(Rt)) )
|
|
#define vset_lane_u32(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdRt_acc( 0xee000b10 | _NENC_21(lane), (Dd), __uint32ToInt32(Rt)) )
|
|
#define vset_lane_u8(Rt, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_DdRt_acc( 0xee400b10 | _NENC_21x6_5(lane), (Dd), __uint8ToInt32(Rt)) )
|
|
|
|
// VMOV (scalar to ARM core register)
|
|
#define vget_lane_f32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_FtDn( 0xee100b10 | _NENC_21(lane), (Dm)) )
|
|
#define vget_lane_p16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), (poly16_t)__neon_RtDn( 0xee900b30 | _NENC_21x6(lane), (Dm)) )
|
|
#define vget_lane_p8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (poly8_t)__neon_RtDn( 0xeed00b10 | _NENC_21x6_5(lane), (Dm)) )
|
|
#define vget_lane_s16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), (int16_t)__neon_RtDn( 0xee100b30 | _NENC_21x6(lane), (Dm)) )
|
|
#define vget_lane_s8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (int8_t)__neon_RtDn( 0xee500b10 | _NENC_21x6_5(lane), (Dm)) )
|
|
#define vget_lane_s32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), (int32_t)__neon_RtDn( 0xee100b10 | _NENC_21(lane), (Dm)) )
|
|
#define vget_lane_u16(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), (uint16_t)__neon_RtDn( 0xee900b30 | _NENC_21x6(lane), (Dm)) )
|
|
#define vget_lane_u8(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (uint8_t)__neon_RtDn( 0xeed00b10 | _NENC_21x6_5(lane), (Dm)) )
|
|
#define vget_lane_u32(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), (uint32_t)__neon_RtDn( 0xee100b10 | _NENC_21(lane), (Dm)) )
|
|
|
|
// VMOV.64 (ARM core register pair to scalar)
|
|
#define vset_lane_s64(R64t, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 1, "invalid lane index"), __neon_DdRtRt2_acc( 0xec400b10, (Dd), __int64ToInt64(R64t)) )
|
|
#define vset_lane_u64(R64t, Dd, lane) ( __static_assert((lane) >= 0 && (lane) < 1, "invalid lane index"), __neon_DdRtRt2_acc( 0xec400b10, (Dd), __uint64ToInt64(R64t)) )
|
|
#define vsetq_lane_s64(R64t, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdRtRt2_acc( 0xec400b10 | _NENC_0(lane), (Qd), __int64ToInt64(R64t)) )
|
|
#define vsetq_lane_u64(R64t, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdRtRt2_acc( 0xec400b10 | _NENC_0(lane), (Qd), __uint64ToInt64(R64t)) )
|
|
|
|
// VMOV.64 (scalar to ARM core register pair)
|
|
#define vget_lane_s64(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 1, "invalid lane index"), (int64_t)__neon_RtRt2Dm( 0xec500b10, (Dm)) )
|
|
#define vget_lane_u64(Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 1, "invalid lane index"), (uint64_t)__neon_RtRt2Dm( 0xec500b10, (Dm)) )
|
|
#define vgetq_lane_s64(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), (int64_t)__neon_RtRt2Qm( 0xec500b10 | _NENC_0(lane), (Qm)) )
|
|
#define vgetq_lane_u64(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), (uint64_t)__neon_RtRt2Qm( 0xec500b10 | _NENC_0(lane), (Qm)) )
|
|
|
|
// VMOV.Q (ARM core register to scalar)
|
|
#define vsetq_lane_f32(Ft, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdFt_acc( 0xee000b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qd), (Ft)) )
|
|
#define vsetq_lane_p16(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdRt_acc( 0xee000b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qd), __poly16ToInt32(Rt)) )
|
|
#define vsetq_lane_p8(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_QdRt_acc( 0xee400b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qd), __poly8ToInt32(Rt)) )
|
|
#define vsetq_lane_s16(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdRt_acc( 0xee000b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qd), __int16ToInt32(Rt)) )
|
|
#define vsetq_lane_s32(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdRt_acc( 0xee000b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qd), __int32ToInt32(Rt)) )
|
|
#define vsetq_lane_s8(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_QdRt_acc( 0xee400b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qd), __int8ToInt32(Rt)) )
|
|
#define vsetq_lane_u16(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_QdRt_acc( 0xee000b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qd), __uint16ToInt32(Rt)) )
|
|
#define vsetq_lane_u32(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdRt_acc( 0xee000b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qd), __uint32ToInt32(Rt)) )
|
|
#define vsetq_lane_u8(Rt, Qd, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_QdRt_acc( 0xee400b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qd), __uint8ToInt32(Rt)) )
|
|
|
|
// VMOV.Q (scalar to ARM core register)
|
|
#define vgetq_lane_f32(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_FtQn( 0xee100b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qm)) )
|
|
#define vgetq_lane_p16(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (poly16_t)__neon_RtQn( 0xee900b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qm)) )
|
|
#define vgetq_lane_p8(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), (poly8_t)__neon_RtQn( 0xeed00b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qm)) )
|
|
#define vgetq_lane_s16(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (int16_t)__neon_RtQn( 0xee100b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qm)) )
|
|
#define vgetq_lane_s8(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), (int8_t)__neon_RtQn( 0xee500b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qm)) )
|
|
#define vgetq_lane_s32(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), (int32_t)__neon_RtQn( 0xee100b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qm)) )
|
|
#define vgetq_lane_u16(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), (uint16_t)__neon_RtQn( 0xee900b30 | _NENC_16((lane) >= 4 ? 1 : 0) | _NENC_21x6((lane) % 4), (Qm)) )
|
|
#define vgetq_lane_u8(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), (uint8_t)__neon_RtQn( 0xeed00b10 | _NENC_16((lane) >= 8 ? 1 : 0) | _NENC_21x6_5((lane) % 8), (Qm)) )
|
|
#define vgetq_lane_u32(Qm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), (uint32_t)__neon_RtQn( 0xee100b10 | _NENC_16((lane) >= 2 ? 1 : 0) | _NENC_21((lane) % 2), (Qm)) )
|
|
|
|
// VMOVL
|
|
#define vmovl_s16(Dm) ( __neon_QdDm( 0xf2900a10, (Dm)) )
|
|
#define vmovl_s32(Dm) ( __neon_QdDm( 0xf2a00a10, (Dm)) )
|
|
#define vmovl_s8(Dm) ( __neon_QdDm( 0xf2880a10, (Dm)) )
|
|
#define vmovl_u16(Dm) ( __neon_QdDm( 0xf3900a10, (Dm)) )
|
|
#define vmovl_u32(Dm) ( __neon_QdDm( 0xf3a00a10, (Dm)) )
|
|
#define vmovl_u8(Dm) ( __neon_QdDm( 0xf3880a10, (Dm)) )
|
|
|
|
// VMOVN
|
|
#define vmovn_s16(Qm) ( __neon_DdQm( 0xf3b20200, (Qm)) )
|
|
#define vmovn_s32(Qm) ( __neon_DdQm( 0xf3b60200, (Qm)) )
|
|
#define vmovn_s64(Qm) ( __neon_DdQm( 0xf3ba0200, (Qm)) )
|
|
#define vmovn_u16(Qm) ( __neon_DdQm( 0xf3b20200, (Qm)) )
|
|
#define vmovn_u32(Qm) ( __neon_DdQm( 0xf3b60200, (Qm)) )
|
|
#define vmovn_u64(Qm) ( __neon_DdQm( 0xf3ba0200, (Qm)) )
|
|
|
|
// VMUL
|
|
#define vmul_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000d10, (Dn), (Dm)) )
|
|
#define vmul_p8(Dn, Dm) ( __neon_DdDnDm( 0xf3000910, (Dn), (Dm)) )
|
|
#define vmul_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100910, (Dn), (Dm)) )
|
|
#define vmul_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200910, (Dn), (Dm)) )
|
|
#define vmul_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000910, (Dn), (Dm)) )
|
|
#define vmul_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2100910, (Dn), (Dm)) )
|
|
#define vmul_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2200910, (Dn), (Dm)) )
|
|
#define vmul_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2000910, (Dn), (Dm)) )
|
|
#define vmulq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf3000d50, (Qn), (Qm)) )
|
|
#define vmulq_p8(Qn, Qm) ( __neon_QdQnQm( 0xf3000950, (Qn), (Qm)) )
|
|
#define vmulq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100950, (Qn), (Qm)) )
|
|
#define vmulq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200950, (Qn), (Qm)) )
|
|
#define vmulq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000950, (Qn), (Qm)) )
|
|
#define vmulq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2100950, (Qn), (Qm)) )
|
|
#define vmulq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2200950, (Qn), (Qm)) )
|
|
#define vmulq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2000950, (Qn), (Qm)) )
|
|
|
|
// VMUL (by scalar - float)
|
|
#define vmul_n_f32(Dn, Ft) ( __neon_DdDnFt( 0xf2a00940, (Dn), (Ft)) )
|
|
#define vmulq_n_f32(Qn, Ft) ( __neon_QdQnFt( 0xf3a00940, (Qn), (Ft)) )
|
|
|
|
// VMUL (by scalar)
|
|
#define vmul_lane_f32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx( 0xf2a00940 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vmul_lane_s16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx( 0xf2900840 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vmul_lane_s32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx( 0xf2a00840 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vmul_lane_u16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx( 0xf2900840 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vmul_lane_u32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx( 0xf2a00840 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vmulq_lane_f32(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx( 0xf3a00940 | _NENC_5(lane), (Qn), (Dm)) )
|
|
#define vmulq_lane_s16(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx( 0xf3900840 | _NENC_5x3(lane), (Qn), (Dm)) )
|
|
#define vmulq_lane_s32(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx( 0xf3a00840 | _NENC_5(lane), (Qn), (Dm)) )
|
|
#define vmulq_lane_u16(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx( 0xf3900840 | _NENC_5x3(lane), (Qn), (Dm)) )
|
|
#define vmulq_lane_u32(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx( 0xf3a00840 | _NENC_5(lane), (Qn), (Dm)) )
|
|
|
|
// VMULL
|
|
#define vmull_p64(Dn, Dm) ( __neon_QdDnDm( 0xf2a00e00, (Dn), (Dm)) )
|
|
#define vmull_p8(Dn, Dm) ( __neon_QdDnDm( 0xf2800e00, (Dn), (Dm)) )
|
|
#define vmull_s16(Dn, Dm) ( __neon_QdDnDm( 0xf2900c00, (Dn), (Dm)) )
|
|
#define vmull_s32(Dn, Dm) ( __neon_QdDnDm( 0xf2a00c00, (Dn), (Dm)) )
|
|
#define vmull_s8(Dn, Dm) ( __neon_QdDnDm( 0xf2800c00, (Dn), (Dm)) )
|
|
#define vmull_u16(Dn, Dm) ( __neon_QdDnDm( 0xf3900c00, (Dn), (Dm)) )
|
|
#define vmull_u32(Dn, Dm) ( __neon_QdDnDm( 0xf3a00c00, (Dn), (Dm)) )
|
|
#define vmull_u8(Dn, Dm) ( __neon_QdDnDm( 0xf3800c00, (Dn), (Dm)) )
|
|
|
|
// VMULL (by scalar)
|
|
#define vmull_lane_s16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx( 0xf2900a40 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vmull_lane_s32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx( 0xf2a00a40 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vmull_lane_u16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx( 0xf3900a40 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vmull_lane_u32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx( 0xf3a00a40 | _NENC_5(lane), (Dn), (Dm)) )
|
|
|
|
// VMVN
|
|
#define vmvn_p16(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_p8(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_s16(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_s32(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_s8(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_u16(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_u32(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvn_u8(Dm) ( __neon_DdDm( 0xf3b00580, (Dm)) )
|
|
#define vmvnq_p16(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_p8(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_s16(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_s32(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_s8(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_u16(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_u32(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
#define vmvnq_u8(Qm) ( __neon_QdQm( 0xf3b005c0, (Qm)) )
|
|
|
|
// VPADAL
|
|
#define vpadal_s16(Dd, Dm) ( __neon_DdDm_acc( 0xf3b40600, (Dd), (Dm)) )
|
|
#define vpadal_s32(Dd, Dm) ( __neon_DdDm_acc( 0xf3b80600, (Dd), (Dm)) )
|
|
#define vpadal_s8(Dd, Dm) ( __neon_DdDm_acc( 0xf3b00600, (Dd), (Dm)) )
|
|
#define vpadal_u16(Dd, Dm) ( __neon_DdDm_acc( 0xf3b40680, (Dd), (Dm)) )
|
|
#define vpadal_u32(Dd, Dm) ( __neon_DdDm_acc( 0xf3b80680, (Dd), (Dm)) )
|
|
#define vpadal_u8(Dd, Dm) ( __neon_DdDm_acc( 0xf3b00680, (Dd), (Dm)) )
|
|
#define vpadalq_s16(Qd, Qm) ( __neon_QdQm_acc( 0xf3b40640, (Qd), (Qm)) )
|
|
#define vpadalq_s32(Qd, Qm) ( __neon_QdQm_acc( 0xf3b80640, (Qd), (Qm)) )
|
|
#define vpadalq_s8(Qd, Qm) ( __neon_QdQm_acc( 0xf3b00640, (Qd), (Qm)) )
|
|
#define vpadalq_u16(Qd, Qm) ( __neon_QdQm_acc( 0xf3b406c0, (Qd), (Qm)) )
|
|
#define vpadalq_u32(Qd, Qm) ( __neon_QdQm_acc( 0xf3b806c0, (Qd), (Qm)) )
|
|
#define vpadalq_u8(Qd, Qm) ( __neon_QdQm_acc( 0xf3b006c0, (Qd), (Qm)) )
|
|
|
|
// VPADD (floating point)
|
|
#define vpadd_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000d00, (Dn), (Dm)) )
|
|
|
|
// VPADD (integer)
|
|
#define vpadd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100b10, (Dn), (Dm)) )
|
|
#define vpadd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200b10, (Dn), (Dm)) )
|
|
#define vpadd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000b10, (Dn), (Dm)) )
|
|
#define vpadd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2100b10, (Dn), (Dm)) )
|
|
#define vpadd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2200b10, (Dn), (Dm)) )
|
|
#define vpadd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2000b10, (Dn), (Dm)) )
|
|
|
|
// VPADDL
|
|
#define vpaddl_s16(Dm) ( __neon_DdDm( 0xf3b40200, (Dm)) )
|
|
#define vpaddl_s32(Dm) ( __neon_DdDm( 0xf3b80200, (Dm)) )
|
|
#define vpaddl_s8(Dm) ( __neon_DdDm( 0xf3b00200, (Dm)) )
|
|
#define vpaddl_u16(Dm) ( __neon_DdDm( 0xf3b40280, (Dm)) )
|
|
#define vpaddl_u32(Dm) ( __neon_DdDm( 0xf3b80280, (Dm)) )
|
|
#define vpaddl_u8(Dm) ( __neon_DdDm( 0xf3b00280, (Dm)) )
|
|
#define vpaddlq_s16(Qm) ( __neon_QdQm( 0xf3b40240, (Qm)) )
|
|
#define vpaddlq_s32(Qm) ( __neon_QdQm( 0xf3b80240, (Qm)) )
|
|
#define vpaddlq_s8(Qm) ( __neon_QdQm( 0xf3b00240, (Qm)) )
|
|
#define vpaddlq_u16(Qm) ( __neon_QdQm( 0xf3b402c0, (Qm)) )
|
|
#define vpaddlq_u32(Qm) ( __neon_QdQm( 0xf3b802c0, (Qm)) )
|
|
#define vpaddlq_u8(Qm) ( __neon_QdQm( 0xf3b002c0, (Qm)) )
|
|
|
|
// VPMAX, VPMIN (floating point)
|
|
#define vpmax_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3000f00, (Dn), (Dm)) )
|
|
#define vpmin_f32(Dn, Dm) ( __neon_DdDnDm( 0xf3200f00, (Dn), (Dm)) )
|
|
|
|
// VPMAX, VPMIN (integer)
|
|
#define vpmax_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100a00, (Dn), (Dm)) )
|
|
#define vpmax_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200a00, (Dn), (Dm)) )
|
|
#define vpmax_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000a00, (Dn), (Dm)) )
|
|
#define vpmax_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100a00, (Dn), (Dm)) )
|
|
#define vpmax_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200a00, (Dn), (Dm)) )
|
|
#define vpmax_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000a00, (Dn), (Dm)) )
|
|
#define vpmin_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100a10, (Dn), (Dm)) )
|
|
#define vpmin_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200a10, (Dn), (Dm)) )
|
|
#define vpmin_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000a10, (Dn), (Dm)) )
|
|
#define vpmin_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100a10, (Dn), (Dm)) )
|
|
#define vpmin_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200a10, (Dn), (Dm)) )
|
|
#define vpmin_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000a10, (Dn), (Dm)) )
|
|
|
|
// VQABS, VQNEG
|
|
#define vqabs_s16(Dm) ( __neon_DdDm( 0xf3b40700, (Dm)) )
|
|
#define vqabs_s32(Dm) ( __neon_DdDm( 0xf3b80700, (Dm)) )
|
|
#define vqabs_s8(Dm) ( __neon_DdDm( 0xf3b00700, (Dm)) )
|
|
#define vqneg_s16(Dm) ( __neon_DdDm( 0xf3b40780, (Dm)) )
|
|
#define vqneg_s32(Dm) ( __neon_DdDm( 0xf3b80780, (Dm)) )
|
|
#define vqneg_s8(Dm) ( __neon_DdDm( 0xf3b00780, (Dm)) )
|
|
#define vqabsq_s16(Qm) ( __neon_QdQm( 0xf3b40740, (Qm)) )
|
|
#define vqabsq_s32(Qm) ( __neon_QdQm( 0xf3b80740, (Qm)) )
|
|
#define vqabsq_s8(Qm) ( __neon_QdQm( 0xf3b00740, (Qm)) )
|
|
#define vqnegq_s16(Qm) ( __neon_QdQm( 0xf3b407c0, (Qm)) )
|
|
#define vqnegq_s32(Qm) ( __neon_QdQm( 0xf3b807c0, (Qm)) )
|
|
#define vqnegq_s8(Qm) ( __neon_QdQm( 0xf3b007c0, (Qm)) )
|
|
|
|
// VQADD
|
|
#define vqadd_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100010, (Dn), (Dm)) )
|
|
#define vqadd_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200010, (Dn), (Dm)) )
|
|
#define vqadd_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300010, (Dn), (Dm)) )
|
|
#define vqadd_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000010, (Dn), (Dm)) )
|
|
#define vqadd_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100010, (Dn), (Dm)) )
|
|
#define vqadd_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200010, (Dn), (Dm)) )
|
|
#define vqadd_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300010, (Dn), (Dm)) )
|
|
#define vqadd_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000010, (Dn), (Dm)) )
|
|
#define vqaddq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100050, (Qn), (Qm)) )
|
|
#define vqaddq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200050, (Qn), (Qm)) )
|
|
#define vqaddq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300050, (Qn), (Qm)) )
|
|
#define vqaddq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000050, (Qn), (Qm)) )
|
|
#define vqaddq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100050, (Qn), (Qm)) )
|
|
#define vqaddq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200050, (Qn), (Qm)) )
|
|
#define vqaddq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300050, (Qn), (Qm)) )
|
|
#define vqaddq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000050, (Qn), (Qm)) )
|
|
|
|
// VQDMLAL, VQDMLSL
|
|
#define vqdmlal_s16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2900900, (Qd), (Dn), (Dm)) )
|
|
#define vqdmlal_s32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2a00900, (Qd), (Dn), (Dm)) )
|
|
#define vqdmlsl_s16(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2900b00, (Qd), (Dn), (Dm)) )
|
|
#define vqdmlsl_s32(Qd, Dn, Dm) ( __neon_QdDnDm_acc( 0xf2a00b00, (Qd), (Dn), (Dm)) )
|
|
|
|
// VQDMLAL, VQDMLSL (by scalar)
|
|
#define vqdmlal_lane_s16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2900340 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vqdmlal_lane_s32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2a00340 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
#define vqdmlsl_lane_s16(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2900740 | _NENC_5x3(lane), (Qd), (Dn), (Dm)) )
|
|
#define vqdmlsl_lane_s32(Qd, Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx_acc( 0xf2a00740 | _NENC_5(lane), (Qd), (Dn), (Dm)) )
|
|
|
|
// VQDMULH (by scalar)
|
|
#define vqdmulh_lane_s16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx( 0xf2900c40 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vqdmulh_lane_s32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx( 0xf2a00c40 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vqrdmulh_lane_s16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_DdDnDmx( 0xf2900d40 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vqrdmulh_lane_s32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_DdDnDmx( 0xf2a00d40 | _NENC_5(lane), (Dn), (Dm)) )
|
|
#define vqdmulhq_lane_s16(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx( 0xf3900c40 | _NENC_5x3(lane), (Qn), (Dm)) )
|
|
#define vqdmulhq_lane_s32(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx( 0xf3a00c40 | _NENC_5(lane), (Qn), (Dm)) )
|
|
#define vqrdmulhq_lane_s16(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdQnDmx( 0xf3900d40 | _NENC_5x3(lane), (Qn), (Dm)) )
|
|
#define vqrdmulhq_lane_s32(Qn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdQnDmx( 0xf3a00d40 | _NENC_5(lane), (Qn), (Dm)) )
|
|
|
|
// VQDMULH, VQRDMULH
|
|
#define vqdmulh_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100b00, (Dn), (Dm)) )
|
|
#define vqdmulh_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200b00, (Dn), (Dm)) )
|
|
#define vqrdmulh_s16(Dn, Dm) ( __neon_DdDnDm( 0xf3100b00, (Dn), (Dm)) )
|
|
#define vqrdmulh_s32(Dn, Dm) ( __neon_DdDnDm( 0xf3200b00, (Dn), (Dm)) )
|
|
#define vqdmulhq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100b40, (Qn), (Qm)) )
|
|
#define vqdmulhq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200b40, (Qn), (Qm)) )
|
|
#define vqrdmulhq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf3100b40, (Qn), (Qm)) )
|
|
#define vqrdmulhq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3200b40, (Qn), (Qm)) )
|
|
|
|
// VQDMULL
|
|
#define vqdmull_s16(Dn, Dm) ( __neon_QdDnDm( 0xf2900d00, (Dn), (Dm)) )
|
|
#define vqdmull_s32(Dn, Dm) ( __neon_QdDnDm( 0xf2a00d00, (Dn), (Dm)) )
|
|
|
|
// VQDMULL (by scalar)
|
|
#define vqdmull_lane_s16(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_QdDnDmx( 0xf2900b40 | _NENC_5x3(lane), (Dn), (Dm)) )
|
|
#define vqdmull_lane_s32(Dn, Dm, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_QdDnDmx( 0xf2a00b40 | _NENC_5(lane), (Dn), (Dm)) )
|
|
|
|
// VQMOVN, VQMOVUN
|
|
#define vqmovn_s16(Qm) ( __neon_DdQm( 0xf3b20280, (Qm)) )
|
|
#define vqmovn_s32(Qm) ( __neon_DdQm( 0xf3b60280, (Qm)) )
|
|
#define vqmovn_s64(Qm) ( __neon_DdQm( 0xf3ba0280, (Qm)) )
|
|
#define vqmovn_u16(Qm) ( __neon_DdQm( 0xf3b202c0, (Qm)) )
|
|
#define vqmovn_u32(Qm) ( __neon_DdQm( 0xf3b602c0, (Qm)) )
|
|
#define vqmovn_u64(Qm) ( __neon_DdQm( 0xf3ba02c0, (Qm)) )
|
|
#define vqmovun_s16(Qm) ( __neon_DdQm( 0xf3b20240, (Qm)) )
|
|
#define vqmovun_s32(Qm) ( __neon_DdQm( 0xf3b60240, (Qm)) )
|
|
#define vqmovun_s64(Qm) ( __neon_DdQm( 0xf3ba0240, (Qm)) )
|
|
|
|
// VQSHL, VQSHLU (immediate)
|
|
#define vqshl_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm( 0xf2900710 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm( 0xf2a00710 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_s64(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm( 0xf2800790 | _NENC_21_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm( 0xf2880710 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_u16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm( 0xf3900710 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_u32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm( 0xf3a00710 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_u64(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm( 0xf3800790 | _NENC_21_16(shift_amount), (Dm)) )
|
|
#define vqshl_n_u8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm( 0xf3880710 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define vqshlu_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm( 0xf3900610 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define vqshlu_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm( 0xf3a00610 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define vqshlu_n_s64(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm( 0xf3800690 | _NENC_21_16(shift_amount), (Dm)) )
|
|
#define vqshlu_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm( 0xf3880610 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define vqshlq_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm( 0xf2900750 | _NENC_19_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm( 0xf2a00750 | _NENC_20_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm( 0xf28007d0 | _NENC_21_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_s8(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm( 0xf2880750 | _NENC_18_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm( 0xf3900750 | _NENC_19_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm( 0xf3a00750 | _NENC_20_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm( 0xf38007d0 | _NENC_21_16(shift_amount), (Qm)) )
|
|
#define vqshlq_n_u8(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm( 0xf3880750 | _NENC_18_16(shift_amount), (Qm)) )
|
|
#define vqshluq_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm( 0xf3900650 | _NENC_19_16(shift_amount), (Qm)) )
|
|
#define vqshluq_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm( 0xf3a00650 | _NENC_20_16(shift_amount), (Qm)) )
|
|
#define vqshluq_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm( 0xf38006d0 | _NENC_21_16(shift_amount), (Qm)) )
|
|
#define vqshluq_n_s8(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm( 0xf3880650 | _NENC_18_16(shift_amount), (Qm)) )
|
|
|
|
// VQSHRN, VQSHRUN, VQRSHRN, VQRSHRUN (immediate)
|
|
#define vqrshrn_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880950 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqrshrn_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900950 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqrshrn_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00950 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vqrshrn_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf3880950 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqrshrn_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf3900950 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqrshrn_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf3a00950 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vqrshrun_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf3880850 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqrshrun_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf3900850 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqrshrun_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf3a00850 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880910 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900910 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00910 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf3880910 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf3900910 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqshrn_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf3a00910 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vqshrun_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf3880810 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vqshrun_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf3900810 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vqshrun_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf3a00810 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
|
|
// VQSUB
|
|
#define vqsub_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100210, (Dn), (Dm)) )
|
|
#define vqsub_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200210, (Dn), (Dm)) )
|
|
#define vqsub_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300210, (Dn), (Dm)) )
|
|
#define vqsub_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000210, (Dn), (Dm)) )
|
|
#define vqsub_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100210, (Dn), (Dm)) )
|
|
#define vqsub_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200210, (Dn), (Dm)) )
|
|
#define vqsub_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300210, (Dn), (Dm)) )
|
|
#define vqsub_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000210, (Dn), (Dm)) )
|
|
#define vqsubq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100250, (Qn), (Qm)) )
|
|
#define vqsubq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200250, (Qn), (Qm)) )
|
|
#define vqsubq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300250, (Qn), (Qm)) )
|
|
#define vqsubq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000250, (Qn), (Qm)) )
|
|
#define vqsubq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100250, (Qn), (Qm)) )
|
|
#define vqsubq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200250, (Qn), (Qm)) )
|
|
#define vqsubq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300250, (Qn), (Qm)) )
|
|
#define vqsubq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000250, (Qn), (Qm)) )
|
|
|
|
// VRECPE, VRSQRTE
|
|
#define vrecpe_f32(Dm) ( __neon_DdDm( 0xf3bb0500, (Dm)) )
|
|
#define vrecpe_u32(Dm) ( __neon_DdDm( 0xf3bb0400, (Dm)) )
|
|
#define vrsqrte_f32(Dm) ( __neon_DdDm( 0xf3bb0580, (Dm)) )
|
|
#define vrsqrte_u32(Dm) ( __neon_DdDm( 0xf3bb0480, (Dm)) )
|
|
#define vrecpeq_f32(Qm) ( __neon_QdQm( 0xf3bb0540, (Qm)) )
|
|
#define vrecpeq_u32(Qm) ( __neon_QdQm( 0xf3bb0440, (Qm)) )
|
|
#define vrsqrteq_f32(Qm) ( __neon_QdQm( 0xf3bb05c0, (Qm)) )
|
|
#define vrsqrteq_u32(Qm) ( __neon_QdQm( 0xf3bb04c0, (Qm)) )
|
|
|
|
// VRECPS
|
|
#define vrecps_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2000f10, (Dn), (Dm)) )
|
|
#define vrecpsq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2000f50, (Qn), (Qm)) )
|
|
|
|
// VREV
|
|
#define vrev16_p8(Dm) ( __neon_DdDm( 0xf3b00100, (Dm)) )
|
|
#define vrev16_s8(Dm) ( __neon_DdDm( 0xf3b00100, (Dm)) )
|
|
#define vrev16_u8(Dm) ( __neon_DdDm( 0xf3b00100, (Dm)) )
|
|
#define vrev32_p16(Dm) ( __neon_DdDm( 0xf3b40080, (Dm)) )
|
|
#define vrev32_p8(Dm) ( __neon_DdDm( 0xf3b00080, (Dm)) )
|
|
#define vrev32_s16(Dm) ( __neon_DdDm( 0xf3b40080, (Dm)) )
|
|
#define vrev32_s8(Dm) ( __neon_DdDm( 0xf3b00080, (Dm)) )
|
|
#define vrev32_u16(Dm) ( __neon_DdDm( 0xf3b40080, (Dm)) )
|
|
#define vrev32_u8(Dm) ( __neon_DdDm( 0xf3b00080, (Dm)) )
|
|
#define vrev64_f32(Dm) ( __neon_DdDm( 0xf3b80000, (Dm)) )
|
|
#define vrev64_p16(Dm) ( __neon_DdDm( 0xf3b40000, (Dm)) )
|
|
#define vrev64_p8(Dm) ( __neon_DdDm( 0xf3b00000, (Dm)) )
|
|
#define vrev64_s16(Dm) ( __neon_DdDm( 0xf3b40000, (Dm)) )
|
|
#define vrev64_s32(Dm) ( __neon_DdDm( 0xf3b80000, (Dm)) )
|
|
#define vrev64_s8(Dm) ( __neon_DdDm( 0xf3b00000, (Dm)) )
|
|
#define vrev64_u16(Dm) ( __neon_DdDm( 0xf3b40000, (Dm)) )
|
|
#define vrev64_u32(Dm) ( __neon_DdDm( 0xf3b80000, (Dm)) )
|
|
#define vrev64_u8(Dm) ( __neon_DdDm( 0xf3b00000, (Dm)) )
|
|
#define vrev16q_p8(Qm) ( __neon_QdQm( 0xf3b00140, (Qm)) )
|
|
#define vrev16q_s8(Qm) ( __neon_QdQm( 0xf3b00140, (Qm)) )
|
|
#define vrev16q_u8(Qm) ( __neon_QdQm( 0xf3b00140, (Qm)) )
|
|
#define vrev32q_p16(Qm) ( __neon_QdQm( 0xf3b400c0, (Qm)) )
|
|
#define vrev32q_p8(Qm) ( __neon_QdQm( 0xf3b000c0, (Qm)) )
|
|
#define vrev32q_s16(Qm) ( __neon_QdQm( 0xf3b400c0, (Qm)) )
|
|
#define vrev32q_s8(Qm) ( __neon_QdQm( 0xf3b000c0, (Qm)) )
|
|
#define vrev32q_u16(Qm) ( __neon_QdQm( 0xf3b400c0, (Qm)) )
|
|
#define vrev32q_u8(Qm) ( __neon_QdQm( 0xf3b000c0, (Qm)) )
|
|
#define vrev64q_f32(Qm) ( __neon_QdQm( 0xf3b80040, (Qm)) )
|
|
#define vrev64q_p16(Qm) ( __neon_QdQm( 0xf3b40040, (Qm)) )
|
|
#define vrev64q_p8(Qm) ( __neon_QdQm( 0xf3b00040, (Qm)) )
|
|
#define vrev64q_s16(Qm) ( __neon_QdQm( 0xf3b40040, (Qm)) )
|
|
#define vrev64q_s32(Qm) ( __neon_QdQm( 0xf3b80040, (Qm)) )
|
|
#define vrev64q_s8(Qm) ( __neon_QdQm( 0xf3b00040, (Qm)) )
|
|
#define vrev64q_u16(Qm) ( __neon_QdQm( 0xf3b40040, (Qm)) )
|
|
#define vrev64q_u32(Qm) ( __neon_QdQm( 0xf3b80040, (Qm)) )
|
|
#define vrev64q_u8(Qm) ( __neon_QdQm( 0xf3b00040, (Qm)) )
|
|
|
|
// VRINT
|
|
#define vrnd_f32(Dm) ( __neon_DdDm( 0xf3ba0580, (Dm)) )
|
|
#define vrnda_f32(Dm) ( __neon_DdDm( 0xf3ba0500, (Dm)) )
|
|
#define vrndm_f32(Dm) ( __neon_DdDm( 0xf3ba0680, (Dm)) )
|
|
#define vrndn_f32(Dm) ( __neon_DdDm( 0xf3ba0400, (Dm)) )
|
|
#define vrndp_f32(Dm) ( __neon_DdDm( 0xf3ba0780, (Dm)) )
|
|
#define vrndx_f32(Dm) ( __neon_DdDm( 0xf3ba0480, (Dm)) )
|
|
#define vrndq_f32(Qm) ( __neon_QdQm( 0xf3ba05c0, (Qm)) )
|
|
#define vrndaq_f32(Qm) ( __neon_QdQm( 0xf3ba0540, (Qm)) )
|
|
#define vrndmq_f32(Qm) ( __neon_QdQm( 0xf3ba06c0, (Qm)) )
|
|
#define vrndnq_f32(Qm) ( __neon_QdQm( 0xf3ba0440, (Qm)) )
|
|
#define vrndpq_f32(Qm) ( __neon_QdQm( 0xf3ba07c0, (Qm)) )
|
|
#define vrndxq_f32(Qm) ( __neon_QdQm( 0xf3ba04c0, (Qm)) )
|
|
|
|
// VRSQRTS
|
|
#define vrsqrts_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2200f10, (Dn), (Dm)) )
|
|
#define vrsqrtsq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2200f50, (Qn), (Qm)) )
|
|
|
|
// VSHL (immediate)
|
|
#define vshl_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm( 0xf2900510 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define vshl_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm( 0xf2a00510 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define vshl_n_s64(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm( 0xf2800590 | _NENC_21_16(shift_amount), (Dm)) )
|
|
#define vshl_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm( 0xf2880510 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define vshl_n_u16(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm( 0xf2900510 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define vshl_n_u32(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm( 0xf2a00510 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define vshl_n_u64(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm( 0xf2800590 | _NENC_21_16(shift_amount), (Dm)) )
|
|
#define vshl_n_u8(Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm( 0xf2880510 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define vshlq_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm( 0xf2900550 | _NENC_19_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm( 0xf2a00550 | _NENC_20_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm( 0xf28005d0 | _NENC_21_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_s8(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm( 0xf2880550 | _NENC_18_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm( 0xf2900550 | _NENC_19_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm( 0xf2a00550 | _NENC_20_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm( 0xf28005d0 | _NENC_21_16(shift_amount), (Qm)) )
|
|
#define vshlq_n_u8(Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm( 0xf2880550 | _NENC_18_16(shift_amount), (Qm)) )
|
|
|
|
// VSHL, VQSHL, VRSHL, VQRSHL (register)
|
|
#define vqrshl_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100510, (Dm), (Dn)) )
|
|
#define vqrshl_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200510, (Dm), (Dn)) )
|
|
#define vqrshl_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300510, (Dm), (Dn)) )
|
|
#define vqrshl_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000510, (Dm), (Dn)) )
|
|
#define vqrshl_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100510, (Dm), (Dn)) )
|
|
#define vqrshl_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200510, (Dm), (Dn)) )
|
|
#define vqrshl_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300510, (Dm), (Dn)) )
|
|
#define vqrshl_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000510, (Dm), (Dn)) )
|
|
#define vqshl_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100410, (Dm), (Dn)) )
|
|
#define vqshl_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200410, (Dm), (Dn)) )
|
|
#define vqshl_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300410, (Dm), (Dn)) )
|
|
#define vqshl_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000410, (Dm), (Dn)) )
|
|
#define vqshl_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100410, (Dm), (Dn)) )
|
|
#define vqshl_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200410, (Dm), (Dn)) )
|
|
#define vqshl_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300410, (Dm), (Dn)) )
|
|
#define vqshl_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000410, (Dm), (Dn)) )
|
|
#define vrshl_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100500, (Dm), (Dn)) )
|
|
#define vrshl_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200500, (Dm), (Dn)) )
|
|
#define vrshl_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300500, (Dm), (Dn)) )
|
|
#define vrshl_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000500, (Dm), (Dn)) )
|
|
#define vrshl_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100500, (Dm), (Dn)) )
|
|
#define vrshl_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200500, (Dm), (Dn)) )
|
|
#define vrshl_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300500, (Dm), (Dn)) )
|
|
#define vrshl_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000500, (Dm), (Dn)) )
|
|
#define vshl_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100400, (Dm), (Dn)) )
|
|
#define vshl_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200400, (Dm), (Dn)) )
|
|
#define vshl_s64(Dn, Dm) ( __neon_DdDnDm( 0xf2300400, (Dm), (Dn)) )
|
|
#define vshl_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000400, (Dm), (Dn)) )
|
|
#define vshl_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100400, (Dm), (Dn)) )
|
|
#define vshl_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200400, (Dm), (Dn)) )
|
|
#define vshl_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300400, (Dm), (Dn)) )
|
|
#define vshl_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000400, (Dm), (Dn)) )
|
|
#define vqrshlq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100550, (Qm), (Qn)) )
|
|
#define vqrshlq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200550, (Qm), (Qn)) )
|
|
#define vqrshlq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300550, (Qm), (Qn)) )
|
|
#define vqrshlq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000550, (Qm), (Qn)) )
|
|
#define vqrshlq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100550, (Qm), (Qn)) )
|
|
#define vqrshlq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200550, (Qm), (Qn)) )
|
|
#define vqrshlq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300550, (Qm), (Qn)) )
|
|
#define vqrshlq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000550, (Qm), (Qn)) )
|
|
#define vqshlq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100450, (Qm), (Qn)) )
|
|
#define vqshlq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200450, (Qm), (Qn)) )
|
|
#define vqshlq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300450, (Qm), (Qn)) )
|
|
#define vqshlq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000450, (Qm), (Qn)) )
|
|
#define vqshlq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100450, (Qm), (Qn)) )
|
|
#define vqshlq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200450, (Qm), (Qn)) )
|
|
#define vqshlq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300450, (Qm), (Qn)) )
|
|
#define vqshlq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000450, (Qm), (Qn)) )
|
|
#define vrshlq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100540, (Qm), (Qn)) )
|
|
#define vrshlq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200540, (Qm), (Qn)) )
|
|
#define vrshlq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300540, (Qm), (Qn)) )
|
|
#define vrshlq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000540, (Qm), (Qn)) )
|
|
#define vrshlq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100540, (Qm), (Qn)) )
|
|
#define vrshlq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200540, (Qm), (Qn)) )
|
|
#define vrshlq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300540, (Qm), (Qn)) )
|
|
#define vrshlq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000540, (Qm), (Qn)) )
|
|
#define vshlq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100440, (Qm), (Qn)) )
|
|
#define vshlq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200440, (Qm), (Qn)) )
|
|
#define vshlq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf2300440, (Qm), (Qn)) )
|
|
#define vshlq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000440, (Qm), (Qn)) )
|
|
#define vshlq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100440, (Qm), (Qn)) )
|
|
#define vshlq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200440, (Qm), (Qn)) )
|
|
#define vshlq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300440, (Qm), (Qn)) )
|
|
#define vshlq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000440, (Qm), (Qn)) )
|
|
|
|
// VSHLL (shift_amount != size)
|
|
#define __internal_vshll_n_t1_s16(Dm, shift_amount) ( __neon_QdDm( 0xf2900a10 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define __internal_vshll_n_t1_s32(Dm, shift_amount) ( __neon_QdDm( 0xf2a00a10 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define __internal_vshll_n_t1_s8(Dm, shift_amount) ( __neon_QdDm( 0xf2880a10 | _NENC_18_16(shift_amount), (Dm)) )
|
|
#define __internal_vshll_n_t1_u16(Dm, shift_amount) ( __neon_QdDm( 0xf3900a10 | _NENC_19_16(shift_amount), (Dm)) )
|
|
#define __internal_vshll_n_t1_u32(Dm, shift_amount) ( __neon_QdDm( 0xf3a00a10 | _NENC_20_16(shift_amount), (Dm)) )
|
|
#define __internal_vshll_n_t1_u8(Dm, shift_amount) ( __neon_QdDm( 0xf3880a10 | _NENC_18_16(shift_amount), (Dm)) )
|
|
|
|
// VSHLL (shift_amount == size)
|
|
#define __internal_vshll_n_t2_s16(Dm) ( __neon_QdDm( 0xf3b60300, (Dm)) )
|
|
#define __internal_vshll_n_t2_s32(Dm) ( __neon_QdDm( 0xf3ba0300, (Dm)) )
|
|
#define __internal_vshll_n_t2_s8(Dm) ( __neon_QdDm( 0xf3b20300, (Dm)) )
|
|
#define __internal_vshll_n_t2_u16(Dm) ( __neon_QdDm( 0xf3b60300, (Dm)) )
|
|
#define __internal_vshll_n_t2_u32(Dm) ( __neon_QdDm( 0xf3ba0300, (Dm)) )
|
|
#define __internal_vshll_n_t2_u8(Dm) ( __neon_QdDm( 0xf3b20300, (Dm)) )
|
|
|
|
// VSHR, VRSHR (immediate)
|
|
#define vrshr_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm( 0xf2900210 | _NENC_19_16(16 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm( 0xf2a00210 | _NENC_20_16(32 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_s64(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm( 0xf2800290 | _NENC_21_16(64 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm( 0xf2880210 | _NENC_18_16(8 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_u16(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm( 0xf3900210 | _NENC_19_16(16 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_u32(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm( 0xf3a00210 | _NENC_20_16(32 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_u64(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm( 0xf3800290 | _NENC_21_16(64 - (shift_amount)), (Dm)) )
|
|
#define vrshr_n_u8(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm( 0xf3880210 | _NENC_18_16(8 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_s16(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm( 0xf2900010 | _NENC_19_16(16 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_s32(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm( 0xf2a00010 | _NENC_20_16(32 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_s64(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm( 0xf2800090 | _NENC_21_16(64 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_s8(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm( 0xf2880010 | _NENC_18_16(8 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_u16(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm( 0xf3900010 | _NENC_19_16(16 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_u32(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm( 0xf3a00010 | _NENC_20_16(32 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_u64(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm( 0xf3800090 | _NENC_21_16(64 - (shift_amount)), (Dm)) )
|
|
#define vshr_n_u8(Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm( 0xf3880010 | _NENC_18_16(8 - (shift_amount)), (Dm)) )
|
|
#define vrshrq_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm( 0xf2900250 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm( 0xf2a00250 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm( 0xf28002d0 | _NENC_21_16(64 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_s8(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm( 0xf2880250 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm( 0xf3900250 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm( 0xf3a00250 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm( 0xf38002d0 | _NENC_21_16(64 - (shift_amount)), (Qm)) )
|
|
#define vrshrq_n_u8(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm( 0xf3880250 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm( 0xf2900050 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm( 0xf2a00050 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm( 0xf28000d0 | _NENC_21_16(64 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_s8(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm( 0xf2880050 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm( 0xf3900050 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm( 0xf3a00050 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm( 0xf38000d0 | _NENC_21_16(64 - (shift_amount)), (Qm)) )
|
|
#define vshrq_n_u8(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm( 0xf3880050 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
|
|
// VSHRN, VRSHRN (immediate)
|
|
#define vrshrn_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880850 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vrshrn_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900850 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vrshrn_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00850 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vrshrn_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880850 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vrshrn_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900850 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vrshrn_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00850 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_s16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880810 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_s32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900810 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_s64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00810 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_u16(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdQm( 0xf2880810 | _NENC_18_16(8 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_u32(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdQm( 0xf2900810 | _NENC_19_16(16 - (shift_amount)), (Qm)) )
|
|
#define vshrn_n_u64(Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdQm( 0xf2a00810 | _NENC_20_16(32 - (shift_amount)), (Qm)) )
|
|
|
|
// VSLI (immediate)
|
|
#define vsli_n_p16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900510 | _NENC_19_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_p8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880510 | _NENC_18_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_s16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900510 | _NENC_19_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_s32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00510 | _NENC_20_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_s64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800590 | _NENC_21_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_s8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880510 | _NENC_18_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_u16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900510 | _NENC_19_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_u32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00510 | _NENC_20_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_u64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800590 | _NENC_21_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsli_n_u8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880510 | _NENC_18_16(shift_amount), (Dd), (Dm)) )
|
|
#define vsliq_n_p16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900550 | _NENC_19_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_p8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880550 | _NENC_18_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_s16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900550 | _NENC_19_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_s32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00550 | _NENC_20_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_s64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38005d0 | _NENC_21_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_s8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880550 | _NENC_18_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_u16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900550 | _NENC_19_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_u32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00550 | _NENC_20_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_u64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38005d0 | _NENC_21_16(shift_amount), (Qd), (Qm)) )
|
|
#define vsliq_n_u8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 0 && (shift_amount) < 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880550 | _NENC_18_16(shift_amount), (Qd), (Qm)) )
|
|
|
|
// VSRA, VRSRA (immediate)
|
|
#define vrsra_n_s16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf2900310 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_s32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf2a00310 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_s64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf2800390 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_s8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf2880310 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_u16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900310 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_u32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00310 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_u64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800390 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsra_n_u8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880310 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_s16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf2900110 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_s32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf2a00110 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_s64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf2800190 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_s8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf2880110 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_u16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900110 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_u32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00110 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_u64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800190 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsra_n_u8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880110 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vrsraq_n_s16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf2900350 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_s32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf2a00350 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_s64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf28003d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_s8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf2880350 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_u16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900350 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_u32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00350 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_u64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38003d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vrsraq_n_u8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880350 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_s16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf2900150 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_s32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf2a00150 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_s64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf28001d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_s8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf2880150 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_u16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900150 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_u32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00150 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_u64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38001d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsraq_n_u8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880150 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
|
|
// VSRI (immediate)
|
|
#define vsri_n_p16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900410 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_p8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880410 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_s16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900410 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_s32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00410 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_s64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800490 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_s8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880410 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_u16(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_DdDm_acc( 0xf3900410 | _NENC_19_16(16 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_u32(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_DdDm_acc( 0xf3a00410 | _NENC_20_16(32 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_u64(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_DdDm_acc( 0xf3800490 | _NENC_21_16(64 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsri_n_u8(Dd, Dm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_DdDm_acc( 0xf3880410 | _NENC_18_16(8 - (shift_amount)), (Dd), (Dm)) )
|
|
#define vsriq_n_p16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900450 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_p8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880450 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_s16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900450 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_s32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00450 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_s64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38004d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_s8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880450 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_u16(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 16, "invalid shift amount"), __neon_QdQm_acc( 0xf3900450 | _NENC_19_16(16 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_u32(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 32, "invalid shift amount"), __neon_QdQm_acc( 0xf3a00450 | _NENC_20_16(32 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_u64(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 64, "invalid shift amount"), __neon_QdQm_acc( 0xf38004d0 | _NENC_21_16(64 - (shift_amount)), (Qd), (Qm)) )
|
|
#define vsriq_n_u8(Qd, Qm, shift_amount) ( __static_assert((shift_amount) >= 1 && (shift_amount) <= 8, "invalid shift amount"), __neon_QdQm_acc( 0xf3880450 | _NENC_18_16(8 - (shift_amount)), (Qd), (Qm)) )
|
|
|
|
// VST1 (multiple single elements)
|
|
#define vst1_f32(pD, D) ( __neon_AdrD1( 0xf400078f, __float32ToN64(pD), (D)) )
|
|
#define vst1_p16(pD, D) ( __neon_AdrD1( 0xf400074f, __poly16ToN64(pD), (D)) )
|
|
#define vst1_p8(pD, D) ( __neon_AdrD1( 0xf400070f, __poly8ToN64(pD), (D)) )
|
|
#define vst1_s16(pD, D) ( __neon_AdrD1( 0xf400074f, __int16ToN64(pD), (D)) )
|
|
#define vst1_s32(pD, D) ( __neon_AdrD1( 0xf400078f, __int32ToN64(pD), (D)) )
|
|
#define vst1_s64(pD, D) ( __neon_AdrD1( 0xf40007cf, __int64ToN64(pD), (D)) )
|
|
#define vst1_s8(pD, D) ( __neon_AdrD1( 0xf400070f, __int8ToN64(pD), (D)) )
|
|
#define vst1_u16(pD, D) ( __neon_AdrD1( 0xf400074f, __uint16ToN64(pD), (D)) )
|
|
#define vst1_u32(pD, D) ( __neon_AdrD1( 0xf400078f, __uint32ToN64(pD), (D)) )
|
|
#define vst1_u64(pD, D) ( __neon_AdrD1( 0xf40007cf, __uint64ToN64(pD), (D)) )
|
|
#define vst1_u8(pD, D) ( __neon_AdrD1( 0xf400070f, __uint8ToN64(pD), (D)) )
|
|
#define vst1_f32_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400078f | _NENC_5_4(_NEON_ALIGN64(align)), __float32ToN64(pD), (D)) )
|
|
#define vst1_p16_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400074f | _NENC_5_4(_NEON_ALIGN64(align)), __poly16ToN64(pD), (D)) )
|
|
#define vst1_p8_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400070f | _NENC_5_4(_NEON_ALIGN64(align)), __poly8ToN64(pD), (D)) )
|
|
#define vst1_s16_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400074f | _NENC_5_4(_NEON_ALIGN64(align)), __int16ToN64(pD), (D)) )
|
|
#define vst1_s32_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400078f | _NENC_5_4(_NEON_ALIGN64(align)), __int32ToN64(pD), (D)) )
|
|
#define vst1_s64_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf40007cf | _NENC_5_4(_NEON_ALIGN64(align)), __int64ToN64(pD), (D)) )
|
|
#define vst1_s8_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400070f | _NENC_5_4(_NEON_ALIGN64(align)), __int8ToN64(pD), (D)) )
|
|
#define vst1_u16_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400074f | _NENC_5_4(_NEON_ALIGN64(align)), __uint16ToN64(pD), (D)) )
|
|
#define vst1_u32_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400078f | _NENC_5_4(_NEON_ALIGN64(align)), __uint32ToN64(pD), (D)) )
|
|
#define vst1_u64_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf40007cf | _NENC_5_4(_NEON_ALIGN64(align)), __uint64ToN64(pD), (D)) )
|
|
#define vst1_u8_ex(pD, D, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrD1( 0xf400070f | _NENC_5_4(_NEON_ALIGN64(align)), __uint8ToN64(pD), (D)) )
|
|
#define vst1q_f32(pD, Q) ( __neon_AdrQ1( 0xf4000a8f, __float32ToN64(pD), (Q)) )
|
|
#define vst1q_p16(pD, Q) ( __neon_AdrQ1( 0xf4000a4f, __poly16ToN64(pD), (Q)) )
|
|
#define vst1q_p8(pD, Q) ( __neon_AdrQ1( 0xf4000a0f, __poly8ToN64(pD), (Q)) )
|
|
#define vst1q_s16(pD, Q) ( __neon_AdrQ1( 0xf4000a4f, __int16ToN64(pD), (Q)) )
|
|
#define vst1q_s32(pD, Q) ( __neon_AdrQ1( 0xf4000a8f, __int32ToN64(pD), (Q)) )
|
|
#define vst1q_s64(pD, Q) ( __neon_AdrQ1( 0xf4000acf, __int64ToN64(pD), (Q)) )
|
|
#define vst1q_s8(pD, Q) ( __neon_AdrQ1( 0xf4000a0f, __int8ToN64(pD), (Q)) )
|
|
#define vst1q_u16(pD, Q) ( __neon_AdrQ1( 0xf4000a4f, __uint16ToN64(pD), (Q)) )
|
|
#define vst1q_u32(pD, Q) ( __neon_AdrQ1( 0xf4000a8f, __uint32ToN64(pD), (Q)) )
|
|
#define vst1q_u64(pD, Q) ( __neon_AdrQ1( 0xf4000acf, __uint64ToN64(pD), (Q)) )
|
|
#define vst1q_u8(pD, Q) ( __neon_AdrQ1( 0xf4000a0f, __uint8ToN64(pD), (Q)) )
|
|
#define vst1q_f32_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64(pD), (Q)) )
|
|
#define vst1q_p16_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64(pD), (Q)) )
|
|
#define vst1q_p8_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64(pD), (Q)) )
|
|
#define vst1q_s16_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64(pD), (Q)) )
|
|
#define vst1q_s32_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64(pD), (Q)) )
|
|
#define vst1q_s64_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __int64ToN64(pD), (Q)) )
|
|
#define vst1q_s8_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64(pD), (Q)) )
|
|
#define vst1q_u16_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a4f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64(pD), (Q)) )
|
|
#define vst1q_u32_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a8f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64(pD), (Q)) )
|
|
#define vst1q_u64_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint64ToN64(pD), (Q)) )
|
|
#define vst1q_u8_ex(pD, Q, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf4000a0f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64(pD), (Q)) )
|
|
|
|
// VST1 (single element from one lane)
|
|
#define vst1_lane_f32(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrD1( 0xf480080f | _NENC_7(lane), __float32ToN64(pD), (D)) )
|
|
#define vst1_lane_p16(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane), __poly16ToN64(pD), (D)) )
|
|
#define vst1_lane_p8(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrD1( 0xf480000f | _NENC_7_5(lane), __poly8ToN64(pD), (D)) )
|
|
#define vst1_lane_s16(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane), __int16ToN64(pD), (D)) )
|
|
#define vst1_lane_s32(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrD1( 0xf480080f | _NENC_7(lane), __int32ToN64(pD), (D)) )
|
|
#define vst1_lane_s8(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrD1( 0xf480000f | _NENC_7_5(lane), __int8ToN64(pD), (D)) )
|
|
#define vst1_lane_u16(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane), __uint16ToN64(pD), (D)) )
|
|
#define vst1_lane_u32(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrD1( 0xf480080f | _NENC_7(lane), __uint32ToN64(pD), (D)) )
|
|
#define vst1_lane_u8(pD, D, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrD1( 0xf480000f | _NENC_7_5(lane), __uint8ToN64(pD), (D)) )
|
|
#define vst1q_lane_f32(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __float32ToN64(pD), (Q)) )
|
|
#define vst1q_lane_p16(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __poly16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_p8(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_AdrQ1( 0xf480000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), __poly8ToN64(pD), (Q)) )
|
|
#define vst1q_lane_s16(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __int16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_s32(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __int32ToN64(pD), (Q)) )
|
|
#define vst1q_lane_s8(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_AdrQ1( 0xf480000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), __int8ToN64(pD), (Q)) )
|
|
#define vst1q_lane_u16(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __uint16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_u32(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __uint32ToN64(pD), (Q)) )
|
|
#define vst1q_lane_u8(pD, Q, lane) ( __static_assert((lane) >= 0 && (lane) < 16, "invalid lane index"), __neon_AdrQ1( 0xf480000f | _NENC_7_5((lane) % 8) | _NENC_12((lane) >= 8 ? 1 : 0), __uint8ToN64(pD), (Q)) )
|
|
|
|
// VST1 (single element from one lane, aligned)
|
|
#define vst1_lane_f32_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __float32ToN64(pD), (D)) )
|
|
#define vst1_lane_p16_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), __poly16ToN64(pD), (D)) )
|
|
#define vst1_lane_s16_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), __int16ToN64(pD), (D)) )
|
|
#define vst1_lane_s32_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __int32ToN64(pD), (D)) )
|
|
#define vst1_lane_u16_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480040f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN16(align)), __uint16ToN64(pD), (D)) )
|
|
#define vst1_lane_u32_ex(pD, D, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrD1( 0xf480080f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __uint32ToN64(pD), (D)) )
|
|
#define vst1q_lane_f32_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __float32ToN64(pD), (Q)) )
|
|
#define vst1q_lane_p16_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), __poly16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_s16_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), __int16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_s32_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __int32ToN64(pD), (Q)) )
|
|
#define vst1q_lane_u16_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480040f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN16(align)), __uint16ToN64(pD), (Q)) )
|
|
#define vst1q_lane_u32_ex(pD, Q, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQ1( 0xf480080f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN32(align) > 0 ? 3 : 0), __uint32ToN64(pD), (Q)) )
|
|
|
|
// VST2 (multiple 2-element structures)
|
|
#define vst2_f32(pD, D2) ( __neon_AdrDx2( 0xf400088f, __float32ToN64(pD), (D2)) )
|
|
#define vst2_p16(pD, D2) ( __neon_AdrDx2( 0xf400084f, __poly16ToN64(pD), (D2)) )
|
|
#define vst2_p8(pD, D2) ( __neon_AdrDx2( 0xf400080f, __poly8ToN64(pD), (D2)) )
|
|
#define vst2_s16(pD, D2) ( __neon_AdrDx2( 0xf400084f, __int16ToN64(pD), (D2)) )
|
|
#define vst2_s32(pD, D2) ( __neon_AdrDx2( 0xf400088f, __int32ToN64(pD), (D2)) )
|
|
#define vst2_s8(pD, D2) ( __neon_AdrDx2( 0xf400080f, __int8ToN64(pD), (D2)) )
|
|
#define vst2_u16(pD, D2) ( __neon_AdrDx2( 0xf400084f, __uint16ToN64(pD), (D2)) )
|
|
#define vst2_u32(pD, D2) ( __neon_AdrDx2( 0xf400088f, __uint32ToN64(pD), (D2)) )
|
|
#define vst2_u8(pD, D2) ( __neon_AdrDx2( 0xf400080f, __uint8ToN64(pD), (D2)) )
|
|
#define vst2_s64(pD, D2) ( __neon_AdrDx2( 0xf4000acf, __int64ToN64(pD), (D2)) )
|
|
#define vst2_u64(pD, D2) ( __neon_AdrDx2( 0xf4000acf, __uint64ToN64(pD), (D2)) )
|
|
#define vst2_s64_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf4000acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __int64ToN64(pD), (D2)) )
|
|
#define vst2_u64_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf4000acf | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint64ToN64(pD), (D2)) )
|
|
#define vst2_f32_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64(pD), (D2)) )
|
|
#define vst2_p16_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64(pD), (D2)) )
|
|
#define vst2_p8_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64(pD), (D2)) )
|
|
#define vst2_s16_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64(pD), (D2)) )
|
|
#define vst2_s32_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64(pD), (D2)) )
|
|
#define vst2_s8_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64(pD), (D2)) )
|
|
#define vst2_u16_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400084f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64(pD), (D2)) )
|
|
#define vst2_u32_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400088f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64(pD), (D2)) )
|
|
#define vst2_u8_ex(pD, D2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx2( 0xf400080f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64(pD), (D2)) )
|
|
#define vst2q_f32(pD, Q2) ( __neon_AdrQx2( 0xf400098f, __float32ToN64(pD), (Q2)) )
|
|
#define vst2q_p16(pD, Q2) ( __neon_AdrQx2( 0xf400094f, __poly16ToN64(pD), (Q2)) )
|
|
#define vst2q_p8(pD, Q2) ( __neon_AdrQx2( 0xf400090f, __poly8ToN64(pD), (Q2)) )
|
|
#define vst2q_s16(pD, Q2) ( __neon_AdrQx2( 0xf400094f, __int16ToN64(pD), (Q2)) )
|
|
#define vst2q_s32(pD, Q2) ( __neon_AdrQx2( 0xf400098f, __int32ToN64(pD), (Q2)) )
|
|
#define vst2q_s8(pD, Q2) ( __neon_AdrQx2( 0xf400090f, __int8ToN64(pD), (Q2)) )
|
|
#define vst2q_u16(pD, Q2) ( __neon_AdrQx2( 0xf400094f, __uint16ToN64(pD), (Q2)) )
|
|
#define vst2q_u32(pD, Q2) ( __neon_AdrQx2( 0xf400098f, __uint32ToN64(pD), (Q2)) )
|
|
#define vst2q_u8(pD, Q2) ( __neon_AdrQx2( 0xf400090f, __uint8ToN64(pD), (Q2)) )
|
|
#define vst2q_f32_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64(pD), (Q2)) )
|
|
#define vst2q_p16_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly16ToN64(pD), (Q2)) )
|
|
#define vst2q_p8_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __poly8ToN64(pD), (Q2)) )
|
|
#define vst2q_s16_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int16ToN64(pD), (Q2)) )
|
|
#define vst2q_s32_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64(pD), (Q2)) )
|
|
#define vst2q_s8_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __int8ToN64(pD), (Q2)) )
|
|
#define vst2q_u16_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400094f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint16ToN64(pD), (Q2)) )
|
|
#define vst2q_u32_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400098f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64(pD), (Q2)) )
|
|
#define vst2q_u8_ex(pD, Q2, align) ( __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx2( 0xf400090f | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint8ToN64(pD), (Q2)) )
|
|
|
|
// VST2 (single 2-element structure from one lane)
|
|
#define vst2_lane_f32(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane), __float32ToN64(pD), (D2)) )
|
|
#define vst2_lane_p16(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane), __poly16ToN64(pD), (D2)) )
|
|
#define vst2_lane_p8(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane), __poly8ToN64(pD), (D2)) )
|
|
#define vst2_lane_s16(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane), __int16ToN64(pD), (D2)) )
|
|
#define vst2_lane_s32(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane), __int32ToN64(pD), (D2)) )
|
|
#define vst2_lane_s8(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane), __int8ToN64(pD), (D2)) )
|
|
#define vst2_lane_u16(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane), __uint16ToN64(pD), (D2)) )
|
|
#define vst2_lane_u32(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane), __uint32ToN64(pD), (D2)) )
|
|
#define vst2_lane_u8(pD, D2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane), __uint8ToN64(pD), (D2)) )
|
|
#define vst2q_lane_f32(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __float32ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_p16(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __poly16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_s16(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __int16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_s32(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __int32ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_u16(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __uint16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_u32(pD, Q2, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __uint32ToN64(pD), (Q2)) )
|
|
|
|
// VST2 (single 2-element structure from one lane, aligned)
|
|
#define vst2_lane_f32_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), __float32ToN64(pD), (D2)) )
|
|
#define vst2_lane_p16_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), __poly16ToN64(pD), (D2)) )
|
|
#define vst2_lane_p8_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), __poly8ToN64(pD), (D2)) )
|
|
#define vst2_lane_s16_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), __int16ToN64(pD), (D2)) )
|
|
#define vst2_lane_s32_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), __int32ToN64(pD), (D2)) )
|
|
#define vst2_lane_s8_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), __int8ToN64(pD), (D2)) )
|
|
#define vst2_lane_u16_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480050f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN32(align)), __uint16ToN64(pD), (D2)) )
|
|
#define vst2_lane_u32_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480090f | _NENC_7(lane) | _NENC_4(_NEON_ALIGN64(align)), __uint32ToN64(pD), (D2)) )
|
|
#define vst2_lane_u8_ex(pD, D2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN16(align) >= 0, "invalid align"), __neon_AdrDx2x( 0xf480010f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN16(align)), __uint8ToN64(pD), (D2)) )
|
|
#define vst2q_lane_f32_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __float32ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_p16_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), __poly16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_s16_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), __int16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_s32_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __int32ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_u16_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480052f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN32(align)), __uint16ToN64(pD), (Q2)) )
|
|
#define vst2q_lane_u32_ex(pD, Q2, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx2x( 0xf480094f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __uint32ToN64(pD), (Q2)) )
|
|
|
|
// VST3 (multiple 3-element structures)
|
|
#define vst3_f32(pD, D3) ( __neon_AdrDx3( 0xf400048f, __float32ToN64(pD), (D3)) )
|
|
#define vst3_p16(pD, D3) ( __neon_AdrDx3( 0xf400044f, __poly16ToN64(pD), (D3)) )
|
|
#define vst3_p8(pD, D3) ( __neon_AdrDx3( 0xf400040f, __poly8ToN64(pD), (D3)) )
|
|
#define vst3_s16(pD, D3) ( __neon_AdrDx3( 0xf400044f, __int16ToN64(pD), (D3)) )
|
|
#define vst3_s32(pD, D3) ( __neon_AdrDx3( 0xf400048f, __int32ToN64(pD), (D3)) )
|
|
#define vst3_s8(pD, D3) ( __neon_AdrDx3( 0xf400040f, __int8ToN64(pD), (D3)) )
|
|
#define vst3_u16(pD, D3) ( __neon_AdrDx3( 0xf400044f, __uint16ToN64(pD), (D3)) )
|
|
#define vst3_u32(pD, D3) ( __neon_AdrDx3( 0xf400048f, __uint32ToN64(pD), (D3)) )
|
|
#define vst3_u8(pD, D3) ( __neon_AdrDx3( 0xf400040f, __uint8ToN64(pD), (D3)) )
|
|
#define vst3_s64(pD, D3) ( __neon_AdrDx3( 0xf40006cf, __int64ToN64(pD), (D3)) )
|
|
#define vst3_u64(pD, D3) ( __neon_AdrDx3( 0xf40006cf, __uint64ToN64(pD), (D3)) )
|
|
#define vst3_s64_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf40006cf | _NENC_4(_NEON_ALIGN64(align)), __int64ToN64(pD), (D3)) )
|
|
#define vst3_u64_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf40006cf | _NENC_4(_NEON_ALIGN64(align)), __uint64ToN64(pD), (D3)) )
|
|
#define vst3_f32_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400048f | _NENC_4(_NEON_ALIGN64(align)), __float32ToN64(pD), (D3)) )
|
|
#define vst3_p16_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400044f | _NENC_4(_NEON_ALIGN64(align)), __poly16ToN64(pD), (D3)) )
|
|
#define vst3_p8_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400040f | _NENC_4(_NEON_ALIGN64(align)), __poly8ToN64(pD), (D3)) )
|
|
#define vst3_s16_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400044f | _NENC_4(_NEON_ALIGN64(align)), __int16ToN64(pD), (D3)) )
|
|
#define vst3_s32_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400048f | _NENC_4(_NEON_ALIGN64(align)), __int32ToN64(pD), (D3)) )
|
|
#define vst3_s8_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400040f | _NENC_4(_NEON_ALIGN64(align)), __int8ToN64(pD), (D3)) )
|
|
#define vst3_u16_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400044f | _NENC_4(_NEON_ALIGN64(align)), __uint16ToN64(pD), (D3)) )
|
|
#define vst3_u32_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400048f | _NENC_4(_NEON_ALIGN64(align)), __uint32ToN64(pD), (D3)) )
|
|
#define vst3_u8_ex(pD, D3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx3( 0xf400040f | _NENC_4(_NEON_ALIGN64(align)), __uint8ToN64(pD), (D3)) )
|
|
#define vst3q_f32(pD, Q3) ( __neon_AdrQx3( 0xf400058f, __float32ToN64(pD), (Q3)) )
|
|
#define vst3q_p16(pD, Q3) ( __neon_AdrQx3( 0xf400054f, __poly16ToN64(pD), (Q3)) )
|
|
#define vst3q_p8(pD, Q3) ( __neon_AdrQx3( 0xf400050f, __poly8ToN64(pD), (Q3)) )
|
|
#define vst3q_s16(pD, Q3) ( __neon_AdrQx3( 0xf400054f, __int16ToN64(pD), (Q3)) )
|
|
#define vst3q_s32(pD, Q3) ( __neon_AdrQx3( 0xf400058f, __int32ToN64(pD), (Q3)) )
|
|
#define vst3q_s8(pD, Q3) ( __neon_AdrQx3( 0xf400050f, __int8ToN64(pD), (Q3)) )
|
|
#define vst3q_u16(pD, Q3) ( __neon_AdrQx3( 0xf400054f, __uint16ToN64(pD), (Q3)) )
|
|
#define vst3q_u32(pD, Q3) ( __neon_AdrQx3( 0xf400058f, __uint32ToN64(pD), (Q3)) )
|
|
#define vst3q_u8(pD, Q3) ( __neon_AdrQx3( 0xf400050f, __uint8ToN64(pD), (Q3)) )
|
|
#define vst3q_f32_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400058f | _NENC_4(_NEON_ALIGN64(align)), __float32ToN64(pD), (Q3)) )
|
|
#define vst3q_p16_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400054f | _NENC_4(_NEON_ALIGN64(align)), __poly16ToN64(pD), (Q3)) )
|
|
#define vst3q_p8_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400050f | _NENC_4(_NEON_ALIGN64(align)), __poly8ToN64(pD), (Q3)) )
|
|
#define vst3q_s16_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400054f | _NENC_4(_NEON_ALIGN64(align)), __int16ToN64(pD), (Q3)) )
|
|
#define vst3q_s32_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400058f | _NENC_4(_NEON_ALIGN64(align)), __int32ToN64(pD), (Q3)) )
|
|
#define vst3q_s8_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400050f | _NENC_4(_NEON_ALIGN64(align)), __int8ToN64(pD), (Q3)) )
|
|
#define vst3q_u16_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400054f | _NENC_4(_NEON_ALIGN64(align)), __uint16ToN64(pD), (Q3)) )
|
|
#define vst3q_u32_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400058f | _NENC_4(_NEON_ALIGN64(align)), __uint32ToN64(pD), (Q3)) )
|
|
#define vst3q_u8_ex(pD, Q3, align) ( __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx3( 0xf400050f | _NENC_4(_NEON_ALIGN64(align)), __uint8ToN64(pD), (Q3)) )
|
|
|
|
// VST3 (single 3-element structure from one lane)
|
|
#define vst3_lane_f32(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx3x( 0xf4800a0f | _NENC_7(lane), __float32ToN64(pD), (D3)) )
|
|
#define vst3_lane_p16(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx3x( 0xf480060f | _NENC_7_6(lane), __poly16ToN64(pD), (D3)) )
|
|
#define vst3_lane_p8(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx3x( 0xf480020f | _NENC_7_5(lane), __poly8ToN64(pD), (D3)) )
|
|
#define vst3_lane_s16(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx3x( 0xf480060f | _NENC_7_6(lane), __int16ToN64(pD), (D3)) )
|
|
#define vst3_lane_s32(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx3x( 0xf4800a0f | _NENC_7(lane), __int32ToN64(pD), (D3)) )
|
|
#define vst3_lane_s8(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx3x( 0xf480020f | _NENC_7_5(lane), __int8ToN64(pD), (D3)) )
|
|
#define vst3_lane_u16(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx3x( 0xf480060f | _NENC_7_6(lane), __uint16ToN64(pD), (D3)) )
|
|
#define vst3_lane_u32(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx3x( 0xf4800a0f | _NENC_7(lane), __uint32ToN64(pD), (D3)) )
|
|
#define vst3_lane_u8(pD, D3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx3x( 0xf480020f | _NENC_7_5(lane), __uint8ToN64(pD), (D3)) )
|
|
#define vst3q_lane_f32(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx3x( 0xf4800a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __float32ToN64(pD), (Q3)) )
|
|
#define vst3q_lane_p16(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx3x( 0xf480062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __poly16ToN64(pD), (Q3)) )
|
|
#define vst3q_lane_s16(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx3x( 0xf480062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __int16ToN64(pD), (Q3)) )
|
|
#define vst3q_lane_s32(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx3x( 0xf4800a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __int32ToN64(pD), (Q3)) )
|
|
#define vst3q_lane_u16(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx3x( 0xf480062f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __uint16ToN64(pD), (Q3)) )
|
|
#define vst3q_lane_u32(pD, Q3, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx3x( 0xf4800a4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __uint32ToN64(pD), (Q3)) )
|
|
|
|
// VST4 (multiple 4-element structures)
|
|
#define vst4_f32(pD, D4) ( __neon_AdrDx4( 0xf400008f, __float32ToN64(pD), (D4)) )
|
|
#define vst4_p16(pD, D4) ( __neon_AdrDx4( 0xf400004f, __poly16ToN64(pD), (D4)) )
|
|
#define vst4_p8(pD, D4) ( __neon_AdrDx4( 0xf400000f, __poly8ToN64(pD), (D4)) )
|
|
#define vst4_s16(pD, D4) ( __neon_AdrDx4( 0xf400004f, __int16ToN64(pD), (D4)) )
|
|
#define vst4_s32(pD, D4) ( __neon_AdrDx4( 0xf400008f, __int32ToN64(pD), (D4)) )
|
|
#define vst4_s8(pD, D4) ( __neon_AdrDx4( 0xf400000f, __int8ToN64(pD), (D4)) )
|
|
#define vst4_u16(pD, D4) ( __neon_AdrDx4( 0xf400004f, __uint16ToN64(pD), (D4)) )
|
|
#define vst4_u32(pD, D4) ( __neon_AdrDx4( 0xf400008f, __uint32ToN64(pD), (D4)) )
|
|
#define vst4_u8(pD, D4) ( __neon_AdrDx4( 0xf400000f, __uint8ToN64(pD), (D4)) )
|
|
#define vst4_s64(pD, D4) ( __neon_AdrDx4( 0xf40002cf, __int64ToN64(pD), (D4)) )
|
|
#define vst4_u64(pD, D4) ( __neon_AdrDx4( 0xf40002cf, __uint64ToN64(pD), (D4)) )
|
|
#define vst4_s64_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf40002cf | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int64ToN64(pD), (D4)) )
|
|
#define vst4_u64_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf40002cf | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint64ToN64(pD), (D4)) )
|
|
#define vst4_f32_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __float32ToN64(pD), (D4)) )
|
|
#define vst4_p16_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly16ToN64(pD), (D4)) )
|
|
#define vst4_p8_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly8ToN64(pD), (D4)) )
|
|
#define vst4_s16_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int16ToN64(pD), (D4)) )
|
|
#define vst4_s32_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int32ToN64(pD), (D4)) )
|
|
#define vst4_s8_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int8ToN64(pD), (D4)) )
|
|
#define vst4_u16_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400004f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint16ToN64(pD), (D4)) )
|
|
#define vst4_u32_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400008f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint32ToN64(pD), (D4)) )
|
|
#define vst4_u8_ex(pD, D4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrDx4( 0xf400000f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint8ToN64(pD), (D4)) )
|
|
#define vst4q_f32(pD, Q4) ( __neon_AdrQx4( 0xf400018f, __float32ToN64(pD), (Q4)) )
|
|
#define vst4q_p16(pD, Q4) ( __neon_AdrQx4( 0xf400014f, __poly16ToN64(pD), (Q4)) )
|
|
#define vst4q_p8(pD, Q4) ( __neon_AdrQx4( 0xf400010f, __poly8ToN64(pD), (Q4)) )
|
|
#define vst4q_s16(pD, Q4) ( __neon_AdrQx4( 0xf400014f, __int16ToN64(pD), (Q4)) )
|
|
#define vst4q_s32(pD, Q4) ( __neon_AdrQx4( 0xf400018f, __int32ToN64(pD), (Q4)) )
|
|
#define vst4q_s8(pD, Q4) ( __neon_AdrQx4( 0xf400010f, __int8ToN64(pD), (Q4)) )
|
|
#define vst4q_u16(pD, Q4) ( __neon_AdrQx4( 0xf400014f, __uint16ToN64(pD), (Q4)) )
|
|
#define vst4q_u32(pD, Q4) ( __neon_AdrQx4( 0xf400018f, __uint32ToN64(pD), (Q4)) )
|
|
#define vst4q_u8(pD, Q4) ( __neon_AdrQx4( 0xf400010f, __uint8ToN64(pD), (Q4)) )
|
|
#define vst4q_f32_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __float32ToN64(pD), (Q4)) )
|
|
#define vst4q_p16_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly16ToN64(pD), (Q4)) )
|
|
#define vst4q_p8_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __poly8ToN64(pD), (Q4)) )
|
|
#define vst4q_s16_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int16ToN64(pD), (Q4)) )
|
|
#define vst4q_s32_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int32ToN64(pD), (Q4)) )
|
|
#define vst4q_s8_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __int8ToN64(pD), (Q4)) )
|
|
#define vst4q_u16_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400014f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint16ToN64(pD), (Q4)) )
|
|
#define vst4q_u32_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400018f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint32ToN64(pD), (Q4)) )
|
|
#define vst4q_u8_ex(pD, Q4, align) ( __static_assert(_NEON_ALIGN64_128_256(align) >= 0, "invalid align"), __neon_AdrQx4( 0xf400010f | _NENC_5_4(_NEON_ALIGN64_128_256(align)), __uint8ToN64(pD), (Q4)) )
|
|
|
|
// VST4 (single 4-element structure from one lane)
|
|
#define vst4_lane_f32(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane), __float32ToN64(pD), (D4)) )
|
|
#define vst4_lane_p16(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane), __poly16ToN64(pD), (D4)) )
|
|
#define vst4_lane_p8(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane), __poly8ToN64(pD), (D4)) )
|
|
#define vst4_lane_s16(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane), __int16ToN64(pD), (D4)) )
|
|
#define vst4_lane_s32(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane), __int32ToN64(pD), (D4)) )
|
|
#define vst4_lane_s8(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane), __int8ToN64(pD), (D4)) )
|
|
#define vst4_lane_u16(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane), __uint16ToN64(pD), (D4)) )
|
|
#define vst4_lane_u32(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane), __uint32ToN64(pD), (D4)) )
|
|
#define vst4_lane_u8(pD, D4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane), __uint8ToN64(pD), (D4)) )
|
|
#define vst4q_lane_f32(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __float32ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_p16(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __poly16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_s16(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __int16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_s32(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __int32ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_u16(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0), __uint16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_u32(pD, Q4, lane) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0), __uint32ToN64(pD), (Q4)) )
|
|
|
|
// VST4 (single 4-element structure from one lane, aligned)
|
|
#define vst4_lane_f32_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64(pD), (D4)) )
|
|
#define vst4_lane_p16_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), __poly16ToN64(pD), (D4)) )
|
|
#define vst4_lane_p8_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), __poly8ToN64(pD), (D4)) )
|
|
#define vst4_lane_s16_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), __int16ToN64(pD), (D4)) )
|
|
#define vst4_lane_s32_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64(pD), (D4)) )
|
|
#define vst4_lane_s8_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), __int8ToN64(pD), (D4)) )
|
|
#define vst4_lane_u16_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480070f | _NENC_7_6(lane) | _NENC_4(_NEON_ALIGN64(align)), __uint16ToN64(pD), (D4)) )
|
|
#define vst4_lane_u32_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 2, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf4800b0f | _NENC_7(lane) | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64(pD), (D4)) )
|
|
#define vst4_lane_u8_ex(pD, D4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN32(align) >= 0, "invalid align"), __neon_AdrDx4x( 0xf480030f | _NENC_7_5(lane) | _NENC_4(_NEON_ALIGN32(align)), __uint8ToN64(pD), (D4)) )
|
|
#define vst4q_lane_f32_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), __float32ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_p16_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __poly16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_s16_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __int16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_s32_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), __int32ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_u16_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 8, "invalid lane index"), __static_assert(_NEON_ALIGN64(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf480072f | _NENC_7_6((lane) % 4) | _NENC_12((lane) >= 4 ? 1 : 0) | _NENC_4(_NEON_ALIGN64(align)), __uint16ToN64(pD), (Q4)) )
|
|
#define vst4q_lane_u32_ex(pD, Q4, lane, align) ( __static_assert((lane) >= 0 && (lane) < 4, "invalid lane index"), __static_assert(_NEON_ALIGN64_128(align) >= 0, "invalid align"), __neon_AdrQx4x( 0xf4800b4f | _NENC_7((lane) % 2) | _NENC_12((lane) >= 2 ? 1 : 0) | _NENC_5_4(_NEON_ALIGN64_128(align)), __uint32ToN64(pD), (Q4)) )
|
|
|
|
// VSUB
|
|
#define vsub_f32(Dn, Dm) ( __neon_DdDnDm( 0xf2200d00, (Dn), (Dm)) )
|
|
#define vsub_s16(Dn, Dm) ( __neon_DdDnDm( 0xf3100800, (Dn), (Dm)) )
|
|
#define vsub_s32(Dn, Dm) ( __neon_DdDnDm( 0xf3200800, (Dn), (Dm)) )
|
|
#define vsub_s64(Dn, Dm) ( __neon_DdDnDm( 0xf3300800, (Dn), (Dm)) )
|
|
#define vsub_s8(Dn, Dm) ( __neon_DdDnDm( 0xf3000800, (Dn), (Dm)) )
|
|
#define vsub_u16(Dn, Dm) ( __neon_DdDnDm( 0xf3100800, (Dn), (Dm)) )
|
|
#define vsub_u32(Dn, Dm) ( __neon_DdDnDm( 0xf3200800, (Dn), (Dm)) )
|
|
#define vsub_u64(Dn, Dm) ( __neon_DdDnDm( 0xf3300800, (Dn), (Dm)) )
|
|
#define vsub_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3000800, (Dn), (Dm)) )
|
|
#define vsubq_f32(Qn, Qm) ( __neon_QdQnQm( 0xf2200d40, (Qn), (Qm)) )
|
|
#define vsubq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf3100840, (Qn), (Qm)) )
|
|
#define vsubq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf3200840, (Qn), (Qm)) )
|
|
#define vsubq_s64(Qn, Qm) ( __neon_QdQnQm( 0xf3300840, (Qn), (Qm)) )
|
|
#define vsubq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf3000840, (Qn), (Qm)) )
|
|
#define vsubq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf3100840, (Qn), (Qm)) )
|
|
#define vsubq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf3200840, (Qn), (Qm)) )
|
|
#define vsubq_u64(Qn, Qm) ( __neon_QdQnQm( 0xf3300840, (Qn), (Qm)) )
|
|
#define vsubq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf3000840, (Qn), (Qm)) )
|
|
|
|
// VSUBHN, VRSUBHN
|
|
#define vrsubhn_s16(Qn, Qm) ( __neon_DdQnQm( 0xf3800600, (Qn), (Qm)) )
|
|
#define vrsubhn_s32(Qn, Qm) ( __neon_DdQnQm( 0xf3900600, (Qn), (Qm)) )
|
|
#define vrsubhn_s64(Qn, Qm) ( __neon_DdQnQm( 0xf3a00600, (Qn), (Qm)) )
|
|
#define vrsubhn_u16(Qn, Qm) ( __neon_DdQnQm( 0xf3800600, (Qn), (Qm)) )
|
|
#define vrsubhn_u32(Qn, Qm) ( __neon_DdQnQm( 0xf3900600, (Qn), (Qm)) )
|
|
#define vrsubhn_u64(Qn, Qm) ( __neon_DdQnQm( 0xf3a00600, (Qn), (Qm)) )
|
|
#define vsubhn_s16(Qn, Qm) ( __neon_DdQnQm( 0xf2800600, (Qn), (Qm)) )
|
|
#define vsubhn_s32(Qn, Qm) ( __neon_DdQnQm( 0xf2900600, (Qn), (Qm)) )
|
|
#define vsubhn_s64(Qn, Qm) ( __neon_DdQnQm( 0xf2a00600, (Qn), (Qm)) )
|
|
#define vsubhn_u16(Qn, Qm) ( __neon_DdQnQm( 0xf2800600, (Qn), (Qm)) )
|
|
#define vsubhn_u32(Qn, Qm) ( __neon_DdQnQm( 0xf2900600, (Qn), (Qm)) )
|
|
#define vsubhn_u64(Qn, Qm) ( __neon_DdQnQm( 0xf2a00600, (Qn), (Qm)) )
|
|
|
|
// VSUBL, VSUBW
|
|
#define vsubl_s16(Dn, Dm) ( __neon_QdDnDm( 0xf2900200, (Dn), (Dm)) )
|
|
#define vsubl_s32(Dn, Dm) ( __neon_QdDnDm( 0xf2a00200, (Dn), (Dm)) )
|
|
#define vsubl_s8(Dn, Dm) ( __neon_QdDnDm( 0xf2800200, (Dn), (Dm)) )
|
|
#define vsubl_u16(Dn, Dm) ( __neon_QdDnDm( 0xf3900200, (Dn), (Dm)) )
|
|
#define vsubl_u32(Dn, Dm) ( __neon_QdDnDm( 0xf3a00200, (Dn), (Dm)) )
|
|
#define vsubl_u8(Dn, Dm) ( __neon_QdDnDm( 0xf3800200, (Dn), (Dm)) )
|
|
#define vsubw_s16(Qn, Dm) ( __neon_QdQnDm( 0xf2900300, (Qn), (Dm)) )
|
|
#define vsubw_s32(Qn, Dm) ( __neon_QdQnDm( 0xf2a00300, (Qn), (Dm)) )
|
|
#define vsubw_s8(Qn, Dm) ( __neon_QdQnDm( 0xf2800300, (Qn), (Dm)) )
|
|
#define vsubw_u16(Qn, Dm) ( __neon_QdQnDm( 0xf3900300, (Qn), (Dm)) )
|
|
#define vsubw_u32(Qn, Dm) ( __neon_QdQnDm( 0xf3a00300, (Qn), (Dm)) )
|
|
#define vsubw_u8(Qn, Dm) ( __neon_QdQnDm( 0xf3800300, (Qn), (Dm)) )
|
|
|
|
// VTBL, VTBX
|
|
#define vtbl2_p8(D2, Dm) ( __neon_DdDx2Dm( 0xf3b00900, (D2), (Dm)) )
|
|
#define vtbl2_s8(D2, Dm) ( __neon_DdDx2Dm( 0xf3b00900, (D2), (Dm)) )
|
|
#define vtbl2_u8(D2, Dm) ( __neon_DdDx2Dm( 0xf3b00900, (D2), (Dm)) )
|
|
#define vtbx2_p8(Dd, D2, Dm) ( __neon_DdDx2Dm_acc( 0xf3b00940, (Dd), (D2), (Dm)) )
|
|
#define vtbx2_s8(Dd, D2, Dm) ( __neon_DdDx2Dm_acc( 0xf3b00940, (Dd), (D2), (Dm)) )
|
|
#define vtbx2_u8(Dd, D2, Dm) ( __neon_DdDx2Dm_acc( 0xf3b00940, (Dd), (D2), (Dm)) )
|
|
#define vtbl3_p8(D3, Dm) ( __neon_DdDx3Dm( 0xf3b00a00, (D3), (Dm)) )
|
|
#define vtbl3_s8(D3, Dm) ( __neon_DdDx3Dm( 0xf3b00a00, (D3), (Dm)) )
|
|
#define vtbl3_u8(D3, Dm) ( __neon_DdDx3Dm( 0xf3b00a00, (D3), (Dm)) )
|
|
#define vtbx3_p8(Dd, D3, Dm) ( __neon_DdDx3Dm_acc( 0xf3b00a40, (Dd), (D3), (Dm)) )
|
|
#define vtbx3_s8(Dd, D3, Dm) ( __neon_DdDx3Dm_acc( 0xf3b00a40, (Dd), (D3), (Dm)) )
|
|
#define vtbx3_u8(Dd, D3, Dm) ( __neon_DdDx3Dm_acc( 0xf3b00a40, (Dd), (D3), (Dm)) )
|
|
#define vtbl4_p8(D4, Dm) ( __neon_DdDx4Dm( 0xf3b00b00, (D4), (Dm)) )
|
|
#define vtbl4_s8(D4, Dm) ( __neon_DdDx4Dm( 0xf3b00b00, (D4), (Dm)) )
|
|
#define vtbl4_u8(D4, Dm) ( __neon_DdDx4Dm( 0xf3b00b00, (D4), (Dm)) )
|
|
#define vtbx4_p8(Dd, D4, Dm) ( __neon_DdDx4Dm_acc( 0xf3b00b40, (Dd), (D4), (Dm)) )
|
|
#define vtbx4_s8(Dd, D4, Dm) ( __neon_DdDx4Dm_acc( 0xf3b00b40, (Dd), (D4), (Dm)) )
|
|
#define vtbx4_u8(Dd, D4, Dm) ( __neon_DdDx4Dm_acc( 0xf3b00b40, (Dd), (D4), (Dm)) )
|
|
#define vtbl1_p8(Dn, Dm) ( __neon_DdDnDm( 0xf3b00800, (Dn), (Dm)) )
|
|
#define vtbl1_s8(Dn, Dm) ( __neon_DdDnDm( 0xf3b00800, (Dn), (Dm)) )
|
|
#define vtbl1_u8(Dn, Dm) ( __neon_DdDnDm( 0xf3b00800, (Dn), (Dm)) )
|
|
#define vtbx1_p8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3b00840, (Dd), (Dn), (Dm)) )
|
|
#define vtbx1_s8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3b00840, (Dd), (Dn), (Dm)) )
|
|
#define vtbx1_u8(Dd, Dn, Dm) ( __neon_DdDnDm_acc( 0xf3b00840, (Dd), (Dn), (Dm)) )
|
|
|
|
// VTRN
|
|
#define vtrn_f32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vtrn_p16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60080, (Dd), (Dm)) )
|
|
#define vtrn_p8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20080, (Dd), (Dm)) )
|
|
#define vtrn_s16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60080, (Dd), (Dm)) )
|
|
#define vtrn_s32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vtrn_s8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20080, (Dd), (Dm)) )
|
|
#define vtrn_u16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60080, (Dd), (Dm)) )
|
|
#define vtrn_u32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vtrn_u8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20080, (Dd), (Dm)) )
|
|
#define vtrnq_f32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba00c0, (Qd), (Qm)) )
|
|
#define vtrnq_p16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b600c0, (Qd), (Qm)) )
|
|
#define vtrnq_p8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b200c0, (Qd), (Qm)) )
|
|
#define vtrnq_s16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b600c0, (Qd), (Qm)) )
|
|
#define vtrnq_s32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba00c0, (Qd), (Qm)) )
|
|
#define vtrnq_s8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b200c0, (Qd), (Qm)) )
|
|
#define vtrnq_u16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b600c0, (Qd), (Qm)) )
|
|
#define vtrnq_u32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba00c0, (Qd), (Qm)) )
|
|
#define vtrnq_u8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b200c0, (Qd), (Qm)) )
|
|
|
|
// VTRNQ64
|
|
#define vtrnq_s64(Qd, Qm) ( __neon_QdQm_acc3( 0x00000000, (Qd), (Qm)) )
|
|
#define vtrnq_u64(Qd, Qm) ( __neon_QdQm_acc3( 0x00000000, (Qd), (Qm)) )
|
|
|
|
// VTST
|
|
#define vtst_p8(Dn, Dm) ( __neon_DdDnDm( 0xf2000810, (Dn), (Dm)) )
|
|
#define vtst_s16(Dn, Dm) ( __neon_DdDnDm( 0xf2100810, (Dn), (Dm)) )
|
|
#define vtst_s32(Dn, Dm) ( __neon_DdDnDm( 0xf2200810, (Dn), (Dm)) )
|
|
#define vtst_s8(Dn, Dm) ( __neon_DdDnDm( 0xf2000810, (Dn), (Dm)) )
|
|
#define vtst_u16(Dn, Dm) ( __neon_DdDnDm( 0xf2100810, (Dn), (Dm)) )
|
|
#define vtst_u32(Dn, Dm) ( __neon_DdDnDm( 0xf2200810, (Dn), (Dm)) )
|
|
#define vtst_u8(Dn, Dm) ( __neon_DdDnDm( 0xf2000810, (Dn), (Dm)) )
|
|
#define vtstq_p8(Qn, Qm) ( __neon_QdQnQm( 0xf2000850, (Qn), (Qm)) )
|
|
#define vtstq_s16(Qn, Qm) ( __neon_QdQnQm( 0xf2100850, (Qn), (Qm)) )
|
|
#define vtstq_s32(Qn, Qm) ( __neon_QdQnQm( 0xf2200850, (Qn), (Qm)) )
|
|
#define vtstq_s8(Qn, Qm) ( __neon_QdQnQm( 0xf2000850, (Qn), (Qm)) )
|
|
#define vtstq_u16(Qn, Qm) ( __neon_QdQnQm( 0xf2100850, (Qn), (Qm)) )
|
|
#define vtstq_u32(Qn, Qm) ( __neon_QdQnQm( 0xf2200850, (Qn), (Qm)) )
|
|
#define vtstq_u8(Qn, Qm) ( __neon_QdQnQm( 0xf2000850, (Qn), (Qm)) )
|
|
|
|
// VUZP
|
|
#define vuzp_p16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60100, (Dd), (Dm)) )
|
|
#define vuzp_p8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20100, (Dd), (Dm)) )
|
|
#define vuzp_s16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60100, (Dd), (Dm)) )
|
|
#define vuzp_s8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20100, (Dd), (Dm)) )
|
|
#define vuzp_u16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60100, (Dd), (Dm)) )
|
|
#define vuzp_u8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20100, (Dd), (Dm)) )
|
|
#define vuzp_f32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vuzp_s32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vuzp_u32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vuzpq_f32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba0140, (Qd), (Qm)) )
|
|
#define vuzpq_p16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b60140, (Qd), (Qm)) )
|
|
#define vuzpq_p8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b20140, (Qd), (Qm)) )
|
|
#define vuzpq_s16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b60140, (Qd), (Qm)) )
|
|
#define vuzpq_s32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba0140, (Qd), (Qm)) )
|
|
#define vuzpq_s8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b20140, (Qd), (Qm)) )
|
|
#define vuzpq_u16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b60140, (Qd), (Qm)) )
|
|
#define vuzpq_u32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba0140, (Qd), (Qm)) )
|
|
#define vuzpq_u8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b20140, (Qd), (Qm)) )
|
|
|
|
// VZIP
|
|
#define vzip_p16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60180, (Dd), (Dm)) )
|
|
#define vzip_p8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20180, (Dd), (Dm)) )
|
|
#define vzip_s16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60180, (Dd), (Dm)) )
|
|
#define vzip_s8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20180, (Dd), (Dm)) )
|
|
#define vzip_u16(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b60180, (Dd), (Dm)) )
|
|
#define vzip_u8(Dd, Dm) ( __neon_DdDm_acc2( 0xf3b20180, (Dd), (Dm)) )
|
|
#define vzip_f32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vzip_s32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vzip_u32(Dd, Dm) ( __neon_DdDm_acc2( 0xf3ba0080, (Dd), (Dm)) )
|
|
#define vzipq_f32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba01c0, (Qd), (Qm)) )
|
|
#define vzipq_p16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b601c0, (Qd), (Qm)) )
|
|
#define vzipq_p8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b201c0, (Qd), (Qm)) )
|
|
#define vzipq_s16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b601c0, (Qd), (Qm)) )
|
|
#define vzipq_s32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba01c0, (Qd), (Qm)) )
|
|
#define vzipq_s8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b201c0, (Qd), (Qm)) )
|
|
#define vzipq_u16(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b601c0, (Qd), (Qm)) )
|
|
#define vzipq_u32(Qd, Qm) ( __neon_QdQm_acc2( 0xf3ba01c0, (Qd), (Qm)) )
|
|
#define vzipq_u8(Qd, Qm) ( __neon_QdQm_acc2( 0xf3b201c0, (Qd), (Qm)) )
|
|
|
|
// } +++ auto-generated code ends (Neon macros)
|
|
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////
|
|
//
|
|
// { +++ auto-generated code begins (vreinterpret macros)
|
|
|
|
#define vreinterpret_f32_s8(a) (a)
|
|
#define vreinterpret_f32_s16(a) (a)
|
|
#define vreinterpret_f32_s32(a) (a)
|
|
#define vreinterpret_f32_s64(a) (a)
|
|
#define vreinterpret_f32_p8(a) (a)
|
|
#define vreinterpret_f32_p16(a) (a)
|
|
#define vreinterpret_f32_u8(a) (a)
|
|
#define vreinterpret_f32_u16(a) (a)
|
|
#define vreinterpret_f32_u32(a) (a)
|
|
#define vreinterpret_f32_u64(a) (a)
|
|
#define vreinterpret_s8_f32(a) (a)
|
|
#define vreinterpret_s8_s16(a) (a)
|
|
#define vreinterpret_s8_s32(a) (a)
|
|
#define vreinterpret_s8_s64(a) (a)
|
|
#define vreinterpret_s8_p8(a) (a)
|
|
#define vreinterpret_s8_p16(a) (a)
|
|
#define vreinterpret_s8_u8(a) (a)
|
|
#define vreinterpret_s8_u16(a) (a)
|
|
#define vreinterpret_s8_u32(a) (a)
|
|
#define vreinterpret_s8_u64(a) (a)
|
|
#define vreinterpret_s16_f32(a) (a)
|
|
#define vreinterpret_s16_s8(a) (a)
|
|
#define vreinterpret_s16_s32(a) (a)
|
|
#define vreinterpret_s16_s64(a) (a)
|
|
#define vreinterpret_s16_p8(a) (a)
|
|
#define vreinterpret_s16_p16(a) (a)
|
|
#define vreinterpret_s16_u8(a) (a)
|
|
#define vreinterpret_s16_u16(a) (a)
|
|
#define vreinterpret_s16_u32(a) (a)
|
|
#define vreinterpret_s16_u64(a) (a)
|
|
#define vreinterpret_s32_f32(a) (a)
|
|
#define vreinterpret_s32_s8(a) (a)
|
|
#define vreinterpret_s32_s16(a) (a)
|
|
#define vreinterpret_s32_s64(a) (a)
|
|
#define vreinterpret_s32_p8(a) (a)
|
|
#define vreinterpret_s32_p16(a) (a)
|
|
#define vreinterpret_s32_u8(a) (a)
|
|
#define vreinterpret_s32_u16(a) (a)
|
|
#define vreinterpret_s32_u32(a) (a)
|
|
#define vreinterpret_s32_u64(a) (a)
|
|
#define vreinterpret_s64_f32(a) (a)
|
|
#define vreinterpret_s64_s8(a) (a)
|
|
#define vreinterpret_s64_s16(a) (a)
|
|
#define vreinterpret_s64_s32(a) (a)
|
|
#define vreinterpret_s64_p8(a) (a)
|
|
#define vreinterpret_s64_p16(a) (a)
|
|
#define vreinterpret_s64_u8(a) (a)
|
|
#define vreinterpret_s64_u16(a) (a)
|
|
#define vreinterpret_s64_u32(a) (a)
|
|
#define vreinterpret_s64_u64(a) (a)
|
|
#define vreinterpret_p8_f32(a) (a)
|
|
#define vreinterpret_p8_s8(a) (a)
|
|
#define vreinterpret_p8_s16(a) (a)
|
|
#define vreinterpret_p8_s32(a) (a)
|
|
#define vreinterpret_p8_s64(a) (a)
|
|
#define vreinterpret_p8_p16(a) (a)
|
|
#define vreinterpret_p8_u8(a) (a)
|
|
#define vreinterpret_p8_u16(a) (a)
|
|
#define vreinterpret_p8_u32(a) (a)
|
|
#define vreinterpret_p8_u64(a) (a)
|
|
#define vreinterpret_p16_f32(a) (a)
|
|
#define vreinterpret_p16_s8(a) (a)
|
|
#define vreinterpret_p16_s16(a) (a)
|
|
#define vreinterpret_p16_s32(a) (a)
|
|
#define vreinterpret_p16_s64(a) (a)
|
|
#define vreinterpret_p16_p8(a) (a)
|
|
#define vreinterpret_p16_u8(a) (a)
|
|
#define vreinterpret_p16_u16(a) (a)
|
|
#define vreinterpret_p16_u32(a) (a)
|
|
#define vreinterpret_p16_u64(a) (a)
|
|
#define vreinterpret_u8_f32(a) (a)
|
|
#define vreinterpret_u8_s8(a) (a)
|
|
#define vreinterpret_u8_s16(a) (a)
|
|
#define vreinterpret_u8_s32(a) (a)
|
|
#define vreinterpret_u8_s64(a) (a)
|
|
#define vreinterpret_u8_p8(a) (a)
|
|
#define vreinterpret_u8_p16(a) (a)
|
|
#define vreinterpret_u8_u16(a) (a)
|
|
#define vreinterpret_u8_u32(a) (a)
|
|
#define vreinterpret_u8_u64(a) (a)
|
|
#define vreinterpret_u16_f32(a) (a)
|
|
#define vreinterpret_u16_s8(a) (a)
|
|
#define vreinterpret_u16_s16(a) (a)
|
|
#define vreinterpret_u16_s32(a) (a)
|
|
#define vreinterpret_u16_s64(a) (a)
|
|
#define vreinterpret_u16_p8(a) (a)
|
|
#define vreinterpret_u16_p16(a) (a)
|
|
#define vreinterpret_u16_u8(a) (a)
|
|
#define vreinterpret_u16_u32(a) (a)
|
|
#define vreinterpret_u16_u64(a) (a)
|
|
#define vreinterpret_u32_f32(a) (a)
|
|
#define vreinterpret_u32_s8(a) (a)
|
|
#define vreinterpret_u32_s16(a) (a)
|
|
#define vreinterpret_u32_s32(a) (a)
|
|
#define vreinterpret_u32_s64(a) (a)
|
|
#define vreinterpret_u32_p8(a) (a)
|
|
#define vreinterpret_u32_p16(a) (a)
|
|
#define vreinterpret_u32_u8(a) (a)
|
|
#define vreinterpret_u32_u16(a) (a)
|
|
#define vreinterpret_u32_u64(a) (a)
|
|
#define vreinterpret_u64_f32(a) (a)
|
|
#define vreinterpret_u64_s8(a) (a)
|
|
#define vreinterpret_u64_s16(a) (a)
|
|
#define vreinterpret_u64_s32(a) (a)
|
|
#define vreinterpret_u64_s64(a) (a)
|
|
#define vreinterpret_u64_p8(a) (a)
|
|
#define vreinterpret_u64_p16(a) (a)
|
|
#define vreinterpret_u64_u8(a) (a)
|
|
#define vreinterpret_u64_u16(a) (a)
|
|
#define vreinterpret_u64_u32(a) (a)
|
|
#define vreinterpretq_f32_s8(a) (a)
|
|
#define vreinterpretq_f32_s16(a) (a)
|
|
#define vreinterpretq_f32_s32(a) (a)
|
|
#define vreinterpretq_f32_s64(a) (a)
|
|
#define vreinterpretq_f32_p8(a) (a)
|
|
#define vreinterpretq_f32_p16(a) (a)
|
|
#define vreinterpretq_f32_u8(a) (a)
|
|
#define vreinterpretq_f32_u16(a) (a)
|
|
#define vreinterpretq_f32_u32(a) (a)
|
|
#define vreinterpretq_f32_u64(a) (a)
|
|
#define vreinterpretq_s8_f32(a) (a)
|
|
#define vreinterpretq_s8_s16(a) (a)
|
|
#define vreinterpretq_s8_s32(a) (a)
|
|
#define vreinterpretq_s8_s64(a) (a)
|
|
#define vreinterpretq_s8_p8(a) (a)
|
|
#define vreinterpretq_s8_p16(a) (a)
|
|
#define vreinterpretq_s8_u8(a) (a)
|
|
#define vreinterpretq_s8_u16(a) (a)
|
|
#define vreinterpretq_s8_u32(a) (a)
|
|
#define vreinterpretq_s8_u64(a) (a)
|
|
#define vreinterpretq_s16_f32(a) (a)
|
|
#define vreinterpretq_s16_s8(a) (a)
|
|
#define vreinterpretq_s16_s32(a) (a)
|
|
#define vreinterpretq_s16_s64(a) (a)
|
|
#define vreinterpretq_s16_p8(a) (a)
|
|
#define vreinterpretq_s16_p16(a) (a)
|
|
#define vreinterpretq_s16_u8(a) (a)
|
|
#define vreinterpretq_s16_u16(a) (a)
|
|
#define vreinterpretq_s16_u32(a) (a)
|
|
#define vreinterpretq_s16_u64(a) (a)
|
|
#define vreinterpretq_s32_f32(a) (a)
|
|
#define vreinterpretq_s32_s8(a) (a)
|
|
#define vreinterpretq_s32_s16(a) (a)
|
|
#define vreinterpretq_s32_s64(a) (a)
|
|
#define vreinterpretq_s32_p8(a) (a)
|
|
#define vreinterpretq_s32_p16(a) (a)
|
|
#define vreinterpretq_s32_u8(a) (a)
|
|
#define vreinterpretq_s32_u16(a) (a)
|
|
#define vreinterpretq_s32_u32(a) (a)
|
|
#define vreinterpretq_s32_u64(a) (a)
|
|
#define vreinterpretq_s64_f32(a) (a)
|
|
#define vreinterpretq_s64_s8(a) (a)
|
|
#define vreinterpretq_s64_s16(a) (a)
|
|
#define vreinterpretq_s64_s32(a) (a)
|
|
#define vreinterpretq_s64_p8(a) (a)
|
|
#define vreinterpretq_s64_p16(a) (a)
|
|
#define vreinterpretq_s64_u8(a) (a)
|
|
#define vreinterpretq_s64_u16(a) (a)
|
|
#define vreinterpretq_s64_u32(a) (a)
|
|
#define vreinterpretq_s64_u64(a) (a)
|
|
#define vreinterpretq_p8_f32(a) (a)
|
|
#define vreinterpretq_p8_s8(a) (a)
|
|
#define vreinterpretq_p8_s16(a) (a)
|
|
#define vreinterpretq_p8_s32(a) (a)
|
|
#define vreinterpretq_p8_s64(a) (a)
|
|
#define vreinterpretq_p8_p16(a) (a)
|
|
#define vreinterpretq_p8_u8(a) (a)
|
|
#define vreinterpretq_p8_u16(a) (a)
|
|
#define vreinterpretq_p8_u32(a) (a)
|
|
#define vreinterpretq_p8_u64(a) (a)
|
|
#define vreinterpretq_p16_f32(a) (a)
|
|
#define vreinterpretq_p16_s8(a) (a)
|
|
#define vreinterpretq_p16_s16(a) (a)
|
|
#define vreinterpretq_p16_s32(a) (a)
|
|
#define vreinterpretq_p16_s64(a) (a)
|
|
#define vreinterpretq_p16_p8(a) (a)
|
|
#define vreinterpretq_p16_u8(a) (a)
|
|
#define vreinterpretq_p16_u16(a) (a)
|
|
#define vreinterpretq_p16_u32(a) (a)
|
|
#define vreinterpretq_p16_u64(a) (a)
|
|
#define vreinterpretq_u8_f32(a) (a)
|
|
#define vreinterpretq_u8_s8(a) (a)
|
|
#define vreinterpretq_u8_s16(a) (a)
|
|
#define vreinterpretq_u8_s32(a) (a)
|
|
#define vreinterpretq_u8_s64(a) (a)
|
|
#define vreinterpretq_u8_p8(a) (a)
|
|
#define vreinterpretq_u8_p16(a) (a)
|
|
#define vreinterpretq_u8_u16(a) (a)
|
|
#define vreinterpretq_u8_u32(a) (a)
|
|
#define vreinterpretq_u8_u64(a) (a)
|
|
#define vreinterpretq_u16_f32(a) (a)
|
|
#define vreinterpretq_u16_s8(a) (a)
|
|
#define vreinterpretq_u16_s16(a) (a)
|
|
#define vreinterpretq_u16_s32(a) (a)
|
|
#define vreinterpretq_u16_s64(a) (a)
|
|
#define vreinterpretq_u16_p8(a) (a)
|
|
#define vreinterpretq_u16_p16(a) (a)
|
|
#define vreinterpretq_u16_u8(a) (a)
|
|
#define vreinterpretq_u16_u32(a) (a)
|
|
#define vreinterpretq_u16_u64(a) (a)
|
|
#define vreinterpretq_u32_f32(a) (a)
|
|
#define vreinterpretq_u32_s8(a) (a)
|
|
#define vreinterpretq_u32_s16(a) (a)
|
|
#define vreinterpretq_u32_s32(a) (a)
|
|
#define vreinterpretq_u32_s64(a) (a)
|
|
#define vreinterpretq_u32_p8(a) (a)
|
|
#define vreinterpretq_u32_p16(a) (a)
|
|
#define vreinterpretq_u32_u8(a) (a)
|
|
#define vreinterpretq_u32_u16(a) (a)
|
|
#define vreinterpretq_u32_u64(a) (a)
|
|
#define vreinterpretq_u64_f32(a) (a)
|
|
#define vreinterpretq_u64_s8(a) (a)
|
|
#define vreinterpretq_u64_s16(a) (a)
|
|
#define vreinterpretq_u64_s32(a) (a)
|
|
#define vreinterpretq_u64_s64(a) (a)
|
|
#define vreinterpretq_u64_p8(a) (a)
|
|
#define vreinterpretq_u64_p16(a) (a)
|
|
#define vreinterpretq_u64_u8(a) (a)
|
|
#define vreinterpretq_u64_u16(a) (a)
|
|
#define vreinterpretq_u64_u32(a) (a)
|
|
|
|
// } +++ auto-generated code ends (vreinterpret macros)
|
|
|
|
// { +++ auto-generated code begins (Pseudo intrinsics)
|
|
|
|
// Multiply by scalar
|
|
#define vmul_n_s16(Vd, Rt) vmul_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vmul_n_s32(Vd, Rt) vmul_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vmul_n_u16(Vd, Rt) vmul_lane_u16((Vd), vmov_n_u16(Rt), 0)
|
|
#define vmul_n_u32(Vd, Rt) vmul_lane_u32((Vd), vmov_n_u32(Rt), 0)
|
|
#define vmulq_n_s16(Vd, Rt) vmulq_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vmulq_n_s32(Vd, Rt) vmulq_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vmulq_n_u16(Vd, Rt) vmulq_lane_u16((Vd), vmov_n_u16(Rt), 0)
|
|
#define vmulq_n_u32(Vd, Rt) vmulq_lane_u32((Vd), vmov_n_u32(Rt), 0)
|
|
#define vmull_n_s16(Vd, Rt) vmull_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vmull_n_s32(Vd, Rt) vmull_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vmull_n_u16(Vd, Rt) vmull_lane_u16((Vd), vmov_n_u16(Rt), 0)
|
|
#define vmull_n_u32(Vd, Rt) vmull_lane_u32((Vd), vmov_n_u32(Rt), 0)
|
|
#define vqdmulh_n_s16(Vd, Rt) vqdmulh_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vqdmulh_n_s32(Vd, Rt) vqdmulh_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vqdmulhq_n_s16(Vd, Rt) vqdmulhq_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vqdmulhq_n_s32(Vd, Rt) vqdmulhq_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vqdmull_n_s16(Vd, Rt) vqdmull_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vqdmull_n_s32(Vd, Rt) vqdmull_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vqrdmulh_n_s16(Vd, Rt) vqrdmulh_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vqrdmulh_n_s32(Vd, Rt) vqrdmulh_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
#define vqrdmulhq_n_s16(Vd, Rt) vqrdmulhq_lane_s16((Vd), vmov_n_s16(Rt), 0)
|
|
#define vqrdmulhq_n_s32(Vd, Rt) vqrdmulhq_lane_s32((Vd), vmov_n_s32(Rt), 0)
|
|
|
|
// Multiply by scalar with accumulate
|
|
#define vmla_n_s16(Vd, Vn, Rt) vmla_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmla_n_s32(Vd, Vn, Rt) vmla_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmla_n_u16(Vd, Vn, Rt) vmla_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmla_n_u32(Vd, Vn, Rt) vmla_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vmlaq_n_s16(Vd, Vn, Rt) vmlaq_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmlaq_n_s32(Vd, Vn, Rt) vmlaq_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmlaq_n_u16(Vd, Vn, Rt) vmlaq_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmlaq_n_u32(Vd, Vn, Rt) vmlaq_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vmlal_n_s16(Vd, Vn, Rt) vmlal_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmlal_n_s32(Vd, Vn, Rt) vmlal_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmlal_n_u16(Vd, Vn, Rt) vmlal_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmlal_n_u32(Vd, Vn, Rt) vmlal_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vmls_n_s16(Vd, Vn, Rt) vmls_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmls_n_s32(Vd, Vn, Rt) vmls_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmls_n_u16(Vd, Vn, Rt) vmls_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmls_n_u32(Vd, Vn, Rt) vmls_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vmlsq_n_s16(Vd, Vn, Rt) vmlsq_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmlsq_n_s32(Vd, Vn, Rt) vmlsq_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmlsq_n_u16(Vd, Vn, Rt) vmlsq_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmlsq_n_u32(Vd, Vn, Rt) vmlsq_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vmlsl_n_s16(Vd, Vn, Rt) vmlsl_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vmlsl_n_s32(Vd, Vn, Rt) vmlsl_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vmlsl_n_u16(Vd, Vn, Rt) vmlsl_lane_u16((Vd), (Vn), vmov_n_u16(Rt), 0)
|
|
#define vmlsl_n_u32(Vd, Vn, Rt) vmlsl_lane_u32((Vd), (Vn), vmov_n_u32(Rt), 0)
|
|
#define vqdmlal_n_s16(Vd, Vn, Rt) vqdmlal_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vqdmlal_n_s32(Vd, Vn, Rt) vqdmlal_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
#define vqdmlsl_n_s16(Vd, Vn, Rt) vqdmlsl_lane_s16((Vd), (Vn), vmov_n_s16(Rt), 0)
|
|
#define vqdmlsl_n_s32(Vd, Vn, Rt) vqdmlsl_lane_s32((Vd), (Vn), vmov_n_s32(Rt), 0)
|
|
|
|
// VDUP.64 (scalar)
|
|
#define vdup_lane_s64(Dn, lane) ( __static_assert(lane == 0, "invalid lane index"), (Dn) )
|
|
#define vdup_lane_u64(Dn, lane) ( __static_assert(lane == 0, "invalid lane index"), (Dn) )
|
|
|
|
// VDUP.W.64 (scalar)
|
|
#define vdupq_lane_s64(Dn, lane) ( __static_assert(lane == 0, "invalid lane index"), vcombine_s64((Dn), (Dn)) )
|
|
#define vdupq_lane_u64(Dn, lane) ( __static_assert(lane == 0, "invalid lane index"), vcombine_u64((Dn), (Dn)) )
|
|
|
|
// } +++ auto-generated code ends (Pseudo intrinsics)
|