arm_simd.h File Reference

Support functions for ARM and vector operations. More...

Go to the source code of this file.

Functions
CRC32 checksum
uint32_t	CRC32B (uint32_t crc, uint8_t val)
	CRC32 checksum.
uint32_t	CRC32W (uint32_t crc, uint32_t val)
	CRC32 checksum.
uint32_t	CRC32Wx4 (uint32_t crc, const uint32_t vals[4])
	CRC32 checksum.
CRC32-C checksum
uint32_t	CRC32CB (uint32_t crc, uint8_t val)
	CRC32-C checksum.
uint32_t	CRC32CW (uint32_t crc, uint32_t val)
	CRC32-C checksum.
uint32_t	CRC32CWx4 (uint32_t crc, const uint32_t vals[4])
	CRC32-C checksum.
Polynomial multiplication
uint64x2_t	PMULL_00 (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
uint64x2_t	PMULL_01 (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
uint64x2_t	PMULL_10 (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
uint64x2_t	PMULL_11 (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
uint64x2_t	PMULL (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
uint64x2_t	PMULL_HIGH (const uint64x2_t a, const uint64x2_t b)
	Polynomial multiplication.
template<unsigned int C>
uint64x2_t	VEXT_U8 (uint64x2_t a, uint64x2_t b)
	Vector extraction.
ARMv8.2 operations
uint64x2_t	VEOR3 (uint64x2_t a, uint64x2_t b, uint64x2_t c)
	Three-way XOR.
uint64x2_t	VXAR (uint64x2_t a, uint64x2_t b, const int c)
	XOR and rotate.
template<unsigned int C>
uint64x2_t	VXAR (uint64x2_t a, uint64x2_t b)
	XOR and rotate.
uint64x2_t	VRAX1 (uint64x2_t a, uint64x2_t b)
	XOR and rotate.

Detailed Description

Support functions for ARM and vector operations.

Definition in file arm_simd.h.

Function Documentation

CRC32B()

uint32_t CRC32B ( uint32_t crc, uint8_t val )

inline

CRC32 checksum.

Parameters

crc	the starting crc value
val	the value to checksum

Returns

CRC32 value

Since

Crypto++ 8.6

Definition at line 30 of file arm_simd.h.

CRC32W()

uint32_t CRC32W ( uint32_t crc, uint32_t val )

inline

CRC32 checksum.

Parameters

crc	the starting crc value
val	the value to checksum

Returns

CRC32 value

Since

Crypto++ 8.6

Definition at line 46 of file arm_simd.h.

CRC32Wx4()

uint32_t CRC32Wx4 ( uint32_t crc, const uint32_t vals[4] )

inline

CRC32 checksum.

Parameters

crc	the starting crc value
vals	the values to checksum

Returns

CRC32 value

Since

Crypto++ 8.6

Definition at line 62 of file arm_simd.h.

CRC32CB()

uint32_t CRC32CB ( uint32_t crc, uint8_t val )

inline

CRC32-C checksum.

Parameters

crc	the starting crc value
val	the value to checksum

Returns

CRC32-C value

Since

Crypto++ 8.6

Definition at line 86 of file arm_simd.h.

CRC32CW()

uint32_t CRC32CW ( uint32_t crc, uint32_t val )

inline

CRC32-C checksum.

Parameters

crc	the starting crc value
val	the value to checksum

Returns

CRC32-C value

Since

Crypto++ 8.6

Definition at line 102 of file arm_simd.h.

CRC32CWx4()

uint32_t CRC32CWx4 ( uint32_t crc, const uint32_t vals[4] )

inline

CRC32-C checksum.

Parameters

crc	the starting crc value
vals	the values to checksum

Returns

CRC32-C value

Since

Crypto++ 8.6

Definition at line 118 of file arm_simd.h.

PMULL_00()

uint64x2_t PMULL_00 ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL_00() performs polynomial multiplication and presents the result like Intel's c = _mm_clmulepi64_si128(a, b, 0x00). The 0x00 indicates the low 64-bits of a and b are multiplied.

Note

An Intel XMM register is composed of 128-bits. The leftmost bit is MSB and numbered 127, while the rightmost bit is LSB and numbered 0.

Since

Crypto++ 8.0

Definition at line 152 of file arm_simd.h.

PMULL_01()

uint64x2_t PMULL_01 ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL_01 performs() polynomial multiplication and presents the result like Intel's c = _mm_clmulepi64_si128(a, b, 0x01). The 0x01 indicates the low 64-bits of a and high 64-bits of b are multiplied.

Note

An Intel XMM register is composed of 128-bits. The leftmost bit is MSB and numbered 127, while the rightmost bit is LSB and numbered 0.

Since

Crypto++ 8.0

Definition at line 182 of file arm_simd.h.

PMULL_10()

uint64x2_t PMULL_10 ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL_10() performs polynomial multiplication and presents the result like Intel's c = _mm_clmulepi64_si128(a, b, 0x10). The 0x10 indicates the high 64-bits of a and low 64-bits of b are multiplied.

Note

An Intel XMM register is composed of 128-bits. The leftmost bit is MSB and numbered 127, while the rightmost bit is LSB and numbered 0.

Since

Crypto++ 8.0

Definition at line 212 of file arm_simd.h.

PMULL_11()

uint64x2_t PMULL_11 ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL_11() performs polynomial multiplication and presents the result like Intel's c = _mm_clmulepi64_si128(a, b, 0x11). The 0x11 indicates the high 64-bits of a and b are multiplied.

Note

An Intel XMM register is composed of 128-bits. The leftmost bit is MSB and numbered 127, while the rightmost bit is LSB and numbered 0.

Since

Crypto++ 8.0

Definition at line 242 of file arm_simd.h.

PMULL()

uint64x2_t PMULL ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL() performs vmull_p64(). PMULL is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

Since

Crypto++ 8.0

Definition at line 267 of file arm_simd.h.

PMULL_HIGH()

uint64x2_t PMULL_HIGH ( const uint64x2_t a, const uint64x2_t b )

inline

Polynomial multiplication.

Parameters

a	the first value
b	the second value

Returns

vector product

PMULL_HIGH() performs vmull_high_p64(). PMULL_HIGH is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

Since

Crypto++ 8.0

Definition at line 292 of file arm_simd.h.

VEXT_U8()

template<unsigned int C>

uint64x2_t VEXT_U8 ( uint64x2_t a, uint64x2_t b )

inline

Vector extraction.

Template Parameters

C	the byte count

Parameters

a	the first value
b	the second value

Returns

vector

VEXT_U8() extracts the first C bytes of vector a and the remaining bytes in b. VEXT_U8 is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

Since

Crypto++ 8.0

Definition at line 320 of file arm_simd.h.

VEOR3()

uint64x2_t VEOR3 ( uint64x2_t a, uint64x2_t b, uint64x2_t c )

inline

Three-way XOR.

Parameters

a	the first value
b	the second value
c	the third value

Returns

three-way exclusive OR of the values

VEOR3() performs veor3q_u64(). VEOR3 is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

VEOR3 requires ARMv8.2.

Since

Crypto++ 8.6

Definition at line 350 of file arm_simd.h.

VXAR() [1/2]

uint64x2_t VXAR ( uint64x2_t a, uint64x2_t b, const int c )

inline

XOR and rotate.

Parameters

a	the first value
b	the second value
c	the third value

Returns

two-way exclusive OR of the values, then rotated by c

VXARQ() performs vxarq_u64(). VXARQ is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

VXARQ requires ARMv8.2.

Since

Crypto++ 8.6

Definition at line 371 of file arm_simd.h.

VXAR() [2/2]

template<unsigned int C>

uint64x2_t VXAR ( uint64x2_t a, uint64x2_t b )

inline

XOR and rotate.

Template Parameters

C	the rotate amount

Parameters

a	the first value
b	the second value

Returns

two-way exclusive OR of the values, then rotated by C

VXARQ() performs vxarq_u64(). VXARQ is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

VXARQ requires ARMv8.2.

Since

Crypto++ 8.6

Definition at line 393 of file arm_simd.h.

VRAX1()

uint64x2_t VRAX1 ( uint64x2_t a, uint64x2_t b )

inline

XOR and rotate.

Parameters

a	the first value
b	the second value

Returns

two-way exclusive OR of the values, then rotated 1-bit

VRAX1() performs vrax1q_u64(). VRAX1 is provided as GCC inline assembly due to Clang and lack of support for the intrinsic.

VRAX1 requires ARMv8.2.

Since

Crypto++ 8.6

Definition at line 413 of file arm_simd.h.