EBGeometry::SIMD Namespace Reference

SIMD register wrappers and the scalar operations that mirror them. More...

Classes
struct	Pack
	N lanes of T in one SIMD register. Defined only for the (T, N) pairs the compiler's flags provide (see NativeWidth()), and never in a device pass. More...
struct	Scalar
	The scalar type of a scalar or pack type: T for T, and T for Pack<T, N>. More...
struct	Scalar< Pack< T, N > >
	The scalar type of Pack<T, N>. More...

Functions
template<class T >
T	select (const bool a_mask, const T a_ifTrue, const T a_ifFalse) noexcept
	Pick a_ifTrue where a_mask holds, else a_ifFalse (scalar form).
bool	both (const bool a_lhs, const bool a_rhs) noexcept
	Whether both conditions hold (scalar form).
float	sqrt (const float a_x) noexcept
	Square root (scalar form).
double	sqrt (const double a_x) noexcept
	Square root (scalar form).
float	abs (const float a_x) noexcept
	Absolute value (scalar form).
double	abs (const double a_x) noexcept
	Absolute value (scalar form).
template<class T >
constexpr size_t	WidestPack () noexcept
	The widest SIMD register, in lanes of T, that this compilation pass has: 16 floats or 8 doubles under AVX-512F, 8 or 4 under AVX, 4 or 2 under SSE4.1, and 1 otherwise and in every device pass.
template<class T , size_t W>
constexpr size_t	NativeWidth () noexcept
	The lane count to evaluate a group of W values of T with: the widest register this compilation pass has whose lane count divides W, or 1 (the scalar path) if none does.

Detailed Description

SIMD register wrappers and the scalar operations that mirror them.

Function Documentation

abs() [1/2]

double EBGeometry::SIMD::abs ( const double  a_x )

inlinenoexcept

Absolute value (scalar form).

Parameters

[in]

a_x

Argument.

Returns

The absolute value.

abs() [2/2]

float EBGeometry::SIMD::abs ( const float  a_x )

inlinenoexcept

Absolute value (scalar form).

Parameters

[in]

a_x

Argument.

Returns

The absolute value.

both()

bool EBGeometry::SIMD::both ( const bool  a_lhs, const bool  a_rhs  )

inlinenoexcept

Whether both conditions hold (scalar form).

Parameters

[in]	a_lhs	One condition.
[in]	a_rhs	The other.

Returns

a_lhs && a_rhs.

NativeWidth()

template<class T , size_t W>

constexpr size_t EBGeometry::SIMD::NativeWidth ( )

constexprnoexcept

The lane count to evaluate a group of W values of T with: the widest register this compilation pass has whose lane count divides W, or 1 (the scalar path) if none does.

Template Parameters

T	Floating-point type.
W	Group width.

Returns

A lane count N for which Pack<T, N> exists, or 1.

select()

template<class T >

T EBGeometry::SIMD::select ( const bool  a_mask, const T  a_ifTrue, const T  a_ifFalse  )

inlinenoexcept

Pick a_ifTrue where a_mask holds, else a_ifFalse (scalar form).

Template Parameters

T	Floating-point type.

Parameters

[in]	a_mask	Condition.
[in]	a_ifTrue	Value where the condition holds.
[in]	a_ifFalse	Value where it does not.

Returns

The chosen value.

sqrt() [1/2]

double EBGeometry::SIMD::sqrt ( const double  a_x )

inlinenoexcept

Square root (scalar form).

Parameters

[in]

a_x

Argument.

Returns

The square root.

sqrt() [2/2]

float EBGeometry::SIMD::sqrt ( const float  a_x )

inlinenoexcept

Square root (scalar form).

Parameters

[in]

a_x

Argument.

Returns

The square root.

WidestPack()

template<class T >

constexpr size_t EBGeometry::SIMD::WidestPack ( )

constexprnoexcept

The widest SIMD register, in lanes of T, that this compilation pass has: 16 floats or 8 doubles under AVX-512F, 8 or 4 under AVX, 4 or 2 under SSE4.1, and 1 otherwise and in every device pass.

Template Parameters

T	Floating-point type.

Returns

The lane count.