rack::simd::Vector< float, 4 > Struct Reference

Wrapper for __m128 representing an aligned vector of 4 single-precision float values. More...

#include <Vector.hpp>

Collaboration diagram for rack::simd::Vector< float, 4 >:

Public Types
using	type = float

Public Member Functions
	Vector ()=default
	Constructs an uninitialized vector.
	Vector (__m128 v)
	Constructs a vector from a native __m128 type.
	Vector (float x)
	Constructs a vector with all elements set to x.
	Vector (float x1, float x2, float x3, float x4)
	Constructs a vector from four scalars.
void	store (float *x)
	Writes an array of 4 values.
float &	operator[] (int i)
	Accessing vector elements individually is slow and defeats the purpose of vectorizing.
const float &	operator[] (int i) const
	Vector (Vector< int32_t, 4 > a)

Static Public Member Functions
static Vector	zero ()
	Returns a vector with all 0 bits.
static Vector	mask ()
	Returns a vector with all 1 bits.
static Vector	load (const float *x)
	Reads an array of 4 values.
static Vector	cast (Vector< int32_t, 4 > a)

Public Attributes
union rack::simd::Vector< float, 4 >:: { ... }

Static Public Attributes
static constexpr int	size = 4

Detailed Description

Wrapper for __m128 representing an aligned vector of 4 single-precision float values.

Member Typedef Documentation

type

using rack::simd::Vector< float, 4 >::type = float

Constructor & Destructor Documentation

Vector() [1/5]

rack::simd::Vector< float, 4 >::Vector ( )

default

Constructs an uninitialized vector.

Vector() [2/5]

rack::simd::Vector< float, 4 >::Vector ( __m128 v )

inline

Constructs a vector from a native __m128 type.

Vector() [3/5]

rack::simd::Vector< float, 4 >::Vector ( float x )

inline

Constructs a vector with all elements set to x.

Vector() [4/5]

rack::simd::Vector< float, 4 >::Vector ( float x1, float x2, float x3, float x4 )

inline

Constructs a vector from four scalars.

Vector() [5/5]

rack::simd::Vector< float, 4 >::Vector ( Vector< int32_t, 4 > a )

inline

Member Function Documentation

zero()

static Vector rack::simd::Vector< float, 4 >::zero ( )

inlinestatic

Returns a vector with all 0 bits.

mask()

static Vector rack::simd::Vector< float, 4 >::mask ( )

inlinestatic

Returns a vector with all 1 bits.

load()

static Vector rack::simd::Vector< float, 4 >::load ( const float * x )

inlinestatic

Reads an array of 4 values.

On little-endian machines (e.g. x86_64), the order is reversed, so x[0] corresponds to vector.s[3].

store()

void rack::simd::Vector< float, 4 >::store ( float * x )

inline

Writes an array of 4 values.

On little-endian machines (e.g. x86_64), the order is reversed, so x[0] corresponds to vector.s[3].

operator[]() [1/2]

float & rack::simd::Vector< float, 4 >::operator[] ( int i )

inline

Accessing vector elements individually is slow and defeats the purpose of vectorizing.

However, this operator is convenient when writing simple serial code in a non-bottlenecked section.

operator[]() [2/2]

const float & rack::simd::Vector< float, 4 >::operator[] ( int i ) const

inline

cast()

Vector< float, 4 > rack::simd::Vector< float, 4 >::cast ( Vector< int32_t, 4 > a )

inlinestatic

Member Data Documentation

size

constexpr int rack::simd::Vector< float, 4 >::size = 4

staticconstexpr

v

__m128 rack::simd::Vector< float, 4 >::v

s

float rack::simd::Vector< float, 4 >::s[4]

Accessing this array of scalars is slow and defeats the purpose of vectorizing.

[]

union rack::simd::Vector< float, 4 > { ... } rack::simd::Vector< float, 4 >

---

The documentation for this struct was generated from the following file:

• /home/vortico/src/vcv/Rack2/include/simd/Vector.hpp