Vector types

 

libsimdpp

Types

Capabilities

Operations

Conversions

Bitwise operations

Floating-point operations

Integer operations

Memory access operations

Shuffle operations

Miscellaneous operations

 

Types

Overview

Vector type promotion

 

This page outlines the available vector types in the library. Several higher-level descriptions are available in the following pages:

• Overview

• Vector type promotion

Regular vector types

Regular vector types are any vector type with void Expr template parameter. Vector types correspond to one or more physical SIMD registers.

uint8<N,void>	vector containing unsigned 8-bit integers (class template)
int8<N,void>	vector containing signed 8-bit integers (class template)
mask_int8<N,void>	vector containing masks for 8-bit integers (class template)
uint16<N,void>	vector containing unsigned 16-bit integers (class template)
int16<N,void>	vector containing signed 16-bit integers (class template)
mask_int16<N,void>	vector containing masks for 16-bit integers (class template)
uint32<N,void>	vector containing unsigned 32-bit integers (class template)
int32<N,void>	vector containing signed 32-bit integers (class template)
mask_int32<N,void>	vector containing masks for 32-bit integers (class template)
uint64<N,void>	vector containing unsigned 64-bit integers (class template)
int64<N,void>	vector containing signed 64-bit integers (class template)
mask_int64<N,void>	vector containing masks for 64-bit integers (class template)
float32<N,void>	vector containing 32-bit floating-point numbers (class template)
mask_float32<N,void>	vector containing masks for 32-bit floating-point numbers (class template)
float64<N,void>	vector containing 64-bit floating-point numbers (class template)
mask_float64<N,void>	vector containing masks for 64-bit floating-point numbers (class template)

Vector expressions

Vector expressions are any vector type with non-void Expr template parameter. Vector expression stores information needed to evaluate a certain expression.

Vector expressions are not meant to be created by the library user. The user may only evaluate them to a regular vector type.

uint8<N,E>	vector containing unsigned 8-bit integers (class template)
int8<N,E>	vector containing signed 8-bit integers (class template)
mask_int8<N,E>	vector containing masks for 8-bit integers (class template)
uint16<N,E>	vector containing unsigned 16-bit integers (class template)
int16<N,E>	vector containing signed 16-bit integers (class template)
mask_int16<N,E>	vector containing masks for 16-bit integers (class template)
uint32<N,E>	vector containing unsigned 32-bit integers (class template)
int32<N,E>	vector containing signed 32-bit integers (class template)
mask_int32<N,E>	vector containing masks for 32-bit integers (class template)
uint64<N,E>	vector containing unsigned 64-bit integers (class template)
int64<N,E>	vector containing signed 64-bit integers (class template)
mask_int64<N,E>	vector containing masks for 64-bit integers (class template)
float32<N,E>	vector containing 32-bit floating-point numbers (class template)
mask_float32<N,E>	vector containing masks for 32-bit floating-point numbers (class template)
float64<N,E>	vector containing 64-bit floating-point numbers (class template)
mask_float64<N,E>	vector containing masks for 64-bit floating-point numbers (class template)

Categorization types

any_int8	represents either of int8, uint8 or mask_int8 (class template)
any_int16	represents either of int16, uint16 or mask_int16 (class template)
any_int32	represents either of int32, uint32 or mask_int32 (class template)
any_int64	represents either of int64, uint64 or mask_int64 (class template)
any_float32	represents either of float32 or mask_float32 (class template)
any_float64	represents either of float64 or mask_float64 (class template)
any_vec8	represents any_int8 (class template)
any_vec16	represents any_int16 (class template)
any_vec32	represents either any_int32 or any_float32 (class template)
any_vec64	represents either any_int64 or any_float64 (class template)
any_vec	represents any vector (class template)

Helper typedefs

The following helper typedefs are provided:

Type	Definition
int8x16	int8<16>
int8x32	int8<32>
uint8x16	uint8<16>
uint8x32	uint8<32>
mask_int8x16	mask_int8<16>
mask_int8x32	mask_int8<32>
int16x8	int16<8>
int16x16	int16<16>
uint16x8	uint16<8>
uint16x16	uint16<16>
mask_int16x8	mask_int16<8>
mask_int16x16	mask_int16<16>
int32x4	int32<4>
int32x8	int32<8>
uint32x4	uint32<4>
uint32x8	uint32<8>
mask_int32x4	mask_int32<4>
mask_int32x8	mask_int32<8>
int64x2	int64<2>
int64x4	int64<4>
uint64x2	uint64<2>
uint64x4	uint64<4>
mask_int64x2	mask_int64<2>
mask_int64x4	mask_int64<4>
float32v	float32<SIMDPP_FAST_FLOAT32_SIZE>
mask_float32v	mask_float32<SIMDPP_FAST_FLOAT32_SIZE>
float64v	float64<SIMDPP_FAST_FLOAT64_SIZE>
mask_float64v	mask_float64<SIMDPP_FAST_FLOAT64_SIZE>
int8v	int8<SIMDPP_FAST_INT8_SIZE>
uint8v	uint8<SIMDPP_FAST_INT8_SIZE>
mask_int8v	mask_int8<SIMDPP_FAST_INT8_SIZE>
int16v	int16<SIMDPP_FAST_INT16_SIZE>
uint16v	uint16<SIMDPP_FAST_INT16_SIZE>
mask_int16v	mask_int16<SIMDPP_FAST_INT16_SIZE>
int32v	int32<SIMDPP_FAST_INT32_SIZE>
uint32v	uint32<SIMDPP_FAST_INT32_SIZE>
mask_int32v	mask_int32<SIMDPP_FAST_INT32_SIZE>
int64v	int64<SIMDPP_FAST_INT64_SIZE>
uint64v	uint64<SIMDPP_FAST_INT64_SIZE>
mask_int64v	mask_int64<SIMDPP_FAST_INT64_SIZE>
mask_float32v2	mask_float32<SIMDPP_FAST_FLOAT32_SIZE*2>
float64v2	float64<SIMDPP_FAST_FLOAT64_SIZE*2>
mask_float64v2	mask_float64<SIMDPP_FAST_FLOAT64_SIZE*2>
int8v2	int8<SIMDPP_FAST_INT8_SIZE*2>
uint8v2	uint8<SIMDPP_FAST_INT8_SIZE*2>
mask_int8v2	mask_int8<SIMDPP_FAST_INT8_SIZE*2>
int16v2	int16<SIMDPP_FAST_INT16_SIZE*2>
uint16v2	uint16<SIMDPP_FAST_INT16_SIZE*2>
mask_int16v2	mask_int16<SIMDPP_FAST_INT16_SIZE*2>
int32v2	int32<SIMDPP_FAST_INT32_SIZE*2>
uint32v2	uint32<SIMDPP_FAST_INT32_SIZE*2>
mask_int32v2	mask_int32<SIMDPP_FAST_INT32_SIZE*2>
int64v2	int64<SIMDPP_FAST_INT64_SIZE*2>
uint64v2	uint64<SIMDPP_FAST_INT64_SIZE*2>
mask_int64v2	mask_int64<SIMDPP_FAST_INT64_SIZE*2>
float32v4	float32<SIMDPP_FAST_FLOAT32_SIZE*4>
mask_float32v4	mask_float32<SIMDPP_FAST_FLOAT32_SIZE*4>
float64v4	float64<SIMDPP_FAST_FLOAT64_SIZE*4>
mask_float64v4	mask_float64<SIMDPP_FAST_FLOAT64_SIZE*4>
int8v4	int8<SIMDPP_FAST_INT8_SIZE*4>
uint8v4	uint8<SIMDPP_FAST_INT8_SIZE*4>
mask_int8v4	mask_int8<SIMDPP_FAST_INT8_SIZE*4>
int16v4	int16<SIMDPP_FAST_INT16_SIZE*4>
uint16v4	uint16<SIMDPP_FAST_INT16_SIZE*4>
mask_int16v4	mask_int16<SIMDPP_FAST_INT16_SIZE*4>
int32v4	int32<SIMDPP_FAST_INT32_SIZE*4>
uint32v4	uint32<SIMDPP_FAST_INT32_SIZE*4>
mask_int32v4	mask_int32<SIMDPP_FAST_INT32_SIZE*4>
int64v4	int64<SIMDPP_FAST_INT64_SIZE*4>
uint64v4	uint64<SIMDPP_FAST_INT64_SIZE*4>
mask_int64v4	mask_int64<SIMDPP_FAST_INT64_SIZE*4>