#include <Teuchos_ScalarTraits.hpp>
Public Types | |
| typedef T | magnitudeType |
| Madatory typedef for result of magnitude. | |
Static Public Member Functions | |
| magnitudeType | eps () |
| Returns relative machine precision. | |
| magnitudeType | sfmin () |
| Returns safe minimum (sfmin), such that 1/sfmin does not overflow. | |
| magnitudeType | base () |
| Returns the base of the machine. | |
| magnitudeType | prec () |
Returns eps*base. | |
| magnitudeType | t () |
| Returns the number of (base) digits in the mantissa. | |
| magnitudeType | rnd () |
| Returns 1.0 when rounding occurs in addition, 0.0 otherwise. | |
| magnitudeType | emin () |
| Returns the minimum exponent before (gradual) underflow. | |
| magnitudeType | rmin () |
Returns the underflow threshold - base^(emin-1). | |
| magnitudeType | emax () |
| Returns the largest exponent before overflow. | |
| magnitudeType | rmax () |
Overflow theshold - (base^emax)*(1-eps). | |
| magnitudeType | magnitude (T a) |
Returns the magnitudeType of the scalar type a. | |
| T | zero () |
| Returns representation of zero for this scalar type. | |
| T | one () |
| Returns representation of one for this scalar type. | |
| magnitudeType | real (T a) |
Returns the real part of the scalar type a. | |
| magnitudeType | imag (T a) |
Returns the imaginary part of the scalar type a. | |
| T | conjugate (T a) |
Returns the conjugate of the scalar type a. | |
| T | nan () |
| Returns a number that represents NaN. | |
| bool | isnaninf (const T &x) |
Returns true if x is NaN or Inf. | |
| void | seedrandom (unsigned int s) |
Seed the random number generator returned by random(). | |
| T | random () |
| Returns a random number (between -one() and +one()) of this scalar type. | |
| std::string | name () |
| Returns the name of this scalar type. | |
| T | squareroot (T x) |
Returns a number of magnitudeType that is the square root of this scalar type x. | |
| T | pow (T x, T y) |
Returns the result of raising one scalar x to the power y. | |
Static Public Attributes | |
| const bool | isComplex = false |
| Determines if scalar type is std::complex. | |
| const bool | isComparable = false |
| Determines if scalar type supports relational operators such as <, >, <=, >=. | |
| const bool | hasMachineParameters = false |
| Determines if scalar type have machine-specific parameters (i.e. eps(), sfmin(), base(), prec(), t(), rnd(), emin(), rmin(), emax(), rmax() are supported). | |
Scalar traits are an essential part of templated codes. This structure offers the basic traits of the templated scalar type, like defining zero and one, and basic functions on the templated scalar type, like performing a square root.
The functions in the templated base unspecialized struct are designed not to compile (giving a nice compile-time error message) and therefore specializations must be written for Scalar types actually used.
int, float, and double. http://crd.lbl.gov/~dhbailey/mpdist/ ) by configuring Teuchos with --enable-teuchos-arprec and giving the appropriate paths to ARPREC. Then ScalarTraits has the specialization: mp_real. --enable-teuchos-std::complex then ScalarTraits also has a parital specialization for all std::complex numbers of the form std::complex<T>.
|
|||||
|
Madatory typedef for result of magnitude.
|
|
|||||||||
|
Returns relative machine precision.
|
|
|||||||||
|
Returns safe minimum (sfmin), such that 1/sfmin does not overflow.
|
|
|||||||||
|
Returns the base of the machine.
|
|
|||||||||
|
Returns
|
|
|||||||||
|
Returns the number of (base) digits in the mantissa.
|
|
|||||||||
|
Returns 1.0 when rounding occurs in addition, 0.0 otherwise.
|
|
|||||||||
|
Returns the minimum exponent before (gradual) underflow.
|
|
|||||||||
|
Returns the underflow threshold -
|
|
|||||||||
|
Returns the largest exponent before overflow.
|
|
|||||||||
|
Overflow theshold -
|
|
||||||||||
|
Returns the magnitudeType of the scalar type
|
|
|||||||||
|
Returns representation of zero for this scalar type.
|
|
|||||||||
|
Returns representation of one for this scalar type.
|
|
||||||||||
|
Returns the real part of the scalar type
|
|
||||||||||
|
Returns the imaginary part of the scalar type
|
|
||||||||||
|
Returns the conjugate of the scalar type
|
|
|||||||||
|
Returns a number that represents NaN.
|
|
||||||||||
|
Returns
|
|
||||||||||
|
Seed the random number generator returned by
|
|
|||||||||
|
Returns a random number (between -one() and +one()) of this scalar type.
|
|
|||||||||
|
Returns the name of this scalar type.
|
|
||||||||||
|
Returns a number of magnitudeType that is the square root of this scalar type
|
|
||||||||||||||||
|
Returns the result of raising one scalar
|
|
|||||
|
Determines if scalar type is std::complex.
|
|
|||||
|
Determines if scalar type supports relational operators such as <, >, <=, >=.
|
|
|||||
|
Determines if scalar type have machine-specific parameters (i.e. eps(), sfmin(), base(), prec(), t(), rnd(), emin(), rmin(), emax(), rmax() are supported).
|
1.3.9.1