Platform specific constexpr information. More...
#include <eve/arch/platform.hpp>
Provides constexpr information about the status (supported or not) of various platform computing capability. Some of those information are tied to the current target architecture, some from compiler options, and some from explicit preprocessor options.
Static Public Attributes | |
static constexpr bool | supports_nans = true |
Indicates if current platform supports NaN values If this value evaluates to true , implementation of functions will not take any special provision to handle NaN input. By default, it evaluates to true . This can be changed by using -ffast-math or -DEVE_NO_NANS . | |
static constexpr bool | supports_infinites = true |
Indicates if current platform supports infinites values If this value evaluates to true , implementation of functions will not take any special provision to handle infinites input. By default, it evaluates to true . This can be changed by using -ffast-math or -DEVE_NO_INFINITIES . | |
static constexpr bool | supports_invalids = supports_infinites && supports_nans |
Indicates if current platform supports invalids values If this value evaluates to true , implementation of functions will not take any special provision to handle neither NaN nor infinites input. By default, it evaluates to true . This can be changed by using -ffast-math , -DEVE_NO_INFINITIES -DEVE_NO_NANS or directly-DEVE_NO_INVALIDS . | |
static constexpr bool | supports_negative_zero = true |
Indicates if current platform supports negative 0 If this value evaluates to true , implementation of functions will not take any special provision to handle negative 0 input. By default, it evaluates to true . This can be changed by using -ffast-math or -DEVE_NO_MINUSZERO . | |
static constexpr bool | supports_denormals = true |
Indicates if current platform supports denormal values If this value evaluates to true , implementation of functions will not take any special provision to handle denormal values input. By default, it evaluates to true . This can be changed by using -ffast-math or -DEVE_NO_DENORMALS . | |