Math Functions

Modules
	Bit Twiddling
	Utilities for bit-twiddling.
	Floating Point Math with Rounding Modes
	These functions provide convenient access to the floating point instructions with special rounding modes.
	Integer Division
	Tools for doing integer division quickly on GPUS.
	Polynomials
	Trigonometry-related functions.
	Trigonometry
	Trigonometry-related functions.

Enumerations
enum class	sp::RoundMode { RoundMode::Nearest = 0 , RoundMode::Down = 1 , RoundMode::Up = 2 , RoundMode::TowardZero = 3 }
	Rounding mode (usually for floating point). More...

Functions
template<int Rank>
sp::Vec< int, Rank >	sp::offsetToCoords (const sp::Vec< int, Rank > &dims, int offset)
	Convert a 1D offset to n-dimensional co-ordinates, for given dimensions. More...
template<int Rank>
bool	sp::incrementCoords (const sp::Vec< int, Rank > &dims, sp::Vec< int, Rank > &coords)
	Increment co-ordinates in a Rank-dimensional space of size dims, initially at coords. More...
template<int Rank>
bool	sp::advanceCoords (const sp::Vec< int, Rank > &dims, sp::Vec< int, Rank > &coords, int amount)

Detailed Description

Enumeration Type Documentation

RoundMode

enum class sp::RoundMode

strong

Rounding mode (usually for floating point).

Enumerator
Nearest	Round to the nearest representation. Ties are rounded up or down, depending in which results in a zero in the LSB of the mantissa. This is the IEEE 754 default.
Down	Round towards negative infinity.
Up	Round towards positive infinity.
TowardZero	Round towards zero.

Function Documentation

incrementCoords()

template<int Rank>

bool sp::incrementCoords	(	const sp::Vec< int, Rank > &	dims,
		sp::Vec< int, Rank > &	coords
	)

Increment co-ordinates in a Rank-dimensional space of size dims, initially at coords.

offsetToCoords()

template<int Rank>

sp::Vec< int, Rank > sp::offsetToCoords	(	const sp::Vec< int, Rank > &	dims,
		int	offset
	)

Convert a 1D offset to n-dimensional co-ordinates, for given dimensions.