Fortran::evaluate::value::RealValue Class Reference
#include <flang/Evaluate/real-value.h>
Public Types | |
| using | Word = IntegerValue |
Public Member Functions | |
| RealValue (const RealValue &) | |
| RealValue (RealValue &&) | |
| RealValue & | operator= (const RealValue &) |
| RealValue & | operator= (RealValue &&) |
| RealValue (int kind, const RealValue &v) | |
| RealValue (int kind, RealValue &&v) | |
| RealValue (int kind, const Word &w) | |
| Interpret w as the raw bit pattern for the given runtime kind. | |
| RealValue (int kind, double x) | |
| void | print (llvm::raw_ostream &os) const |
| LLVM_DUMP_METHOD void | dump () const |
| bool | IsNull () const |
| int | kind () const |
| The kind of the value currently stored. | |
| int | bits () const |
| std::size_t | bytesStored () const |
| Number of bytes accessed by FromRawBytes/StoreRawBytes. | |
| bool | operator== (const RealValue &y) const |
| bool | operator!= (const RealValue &y) const |
| bool | IsNegative () const |
| bool | IsNotANumber () const |
| bool | IsSignalingNaN () const |
| bool | IsInfinite () const |
| bool | IsFinite () const |
| bool | IsZero () const |
| bool | IsNormal () const |
| RealValue | ABS () const |
| RealValue | SetSign (bool toNegative) const |
| RealValue | SIGN (const RealValue &x) const |
| RealValue | Negate () const |
| Relation | Compare (const RealValue &y) const |
| ValueWithRealFlags< RealValue > | Add (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | Subtract (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | Multiply (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | Divide (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | SQRT (Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | NEAREST (bool upward) const |
| NEAREST(), IEEE_NEXT_AFTER(), IEEE_NEXT_UP(), and IEEE_NEXT_DOWN() | |
| ValueWithRealFlags< RealValue > | HYPOT (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| ValueWithRealFlags< RealValue > | DIM (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| DIM(X,Y) = MAX(X-Y, 0) | |
| ValueWithRealFlags< RealValue > | MOD (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| MOD(x,y) = x - AINT(x/y)*y (in the standard) | |
| ValueWithRealFlags< RealValue > | MODULO (const RealValue &y, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| MODULO(x,y) = x - FLOOR(x/y)*y (in the standard) | |
| ValueWithRealFlags< RealValue > | KahanSummation (const RealValue &y, RealValue &correction, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| IntegerValue | EXPONENT () const |
| RealValue | RRSPACING () const |
| RealValue | SPACING () const |
| RealValue | SET_EXPONENT (std::int64_t e) const |
| RealValue | FRACTION () const |
| ValueWithRealFlags< RealValue > | SCALE (const IntegerValue &by, Rounding rounding=TargetCharacteristics::defaultRounding) const |
| SCALE(); also known as IEEE_SCALB and (in IEEE-754 '08) ScaleB. | |
| RealValue | FlushSubnormalToZero () const |
| ValueWithRealFlags< RealValue > | ToWholeNumber (common::RoundingMode mode=common::RoundingMode::ToZero) const |
| Conversion to integer in the same real format (AINT(), ANINT()) | |
| ValueWithRealFlags< IntegerValue > | ToInteger (common::RoundingMode mode=common::RoundingMode::ToZero, int toBits=0) const |
| Conversion to an integer (INT(), NINT(), FLOOR(), CEILING()) | |
| Word | RawBits () const |
| int | Exponent () const |
| Extracts "raw" biased exponent field. | |
| std::string | DumpHexadecimal () const |
| llvm::raw_ostream & | AsFortran (llvm::raw_ostream &o, int kind, bool minimal=false) const |
| void | StoreRawBytes (void *dst, size_t size, bool *changed=nullptr) const |
Static Public Member Functions | |
| static RealValue | Zero (int kind) |
| static RealValue | NegativeZero (int kind) |
| Creates a floating-point with value -0.0 of a given kind. | |
| static RealValue | Infinity (int kind, bool negative=false) |
| static constexpr int | bits (int kind) |
| static constexpr std::size_t | bytesStored (int kind) |
| static RealValue | EPSILON (int kind) |
| static RealValue | HUGE (int kind) |
| static RealValue | TINY (int kind) |
| static int | DIGITS (int kind) |
| static int | PRECISION (int kind) |
| static int | RANGE (int kind) |
| static int | MAXEXPONENT (int kind) |
| static int | MINEXPONENT (int kind) |
| static RealValue | NotANumber (int kind) |
| static ValueWithRealFlags< RealValue > | FromInteger (int kind, const IntegerValue &n, bool isUnsigned=false, Rounding rounding=TargetCharacteristics::defaultRounding) |
| static ValueWithRealFlags< RealValue > | Convert (int kind, const RealValue &from, Rounding rounding=TargetCharacteristics::defaultRounding) |
| static ValueWithRealFlags< RealValue > | Read (int kind, const char *&pp, Rounding rounding=TargetCharacteristics::defaultRounding) |
| static RealValue | FromRawBytes (int kind, const void *raw, std::size_t expectedSize) |
Detailed Description
A floating-point value with dynamic precision.
The precision is dynamic, but only a predefined set of Fortran kinds are allowed. It is also kind-aware, i.e. knows which REAL kind it currently represents.
The implementation is hidden from this header using a pImpl-like idiom.
Constructor & Destructor Documentation
◆ RealValue()
| Fortran::evaluate::value::RealValue::RealValue | ( | int | kind, |
| double | x ) |
Creates a floating-point value of a given kind from a host double, rounded to the target kind's precision (per the default rounding mode). Portable: does not assume that the host "double" shares any bit layout with the target kind, only that <cmath>'s frexp()/ldexp() are available.
Member Function Documentation
◆ AsFortran()
| llvm::raw_ostream & Fortran::evaluate::value::RealValue::AsFortran | ( | llvm::raw_ostream & | o, |
| int | kind, | ||
| bool | minimal = false ) const |
Emits a character representation for an equivalent Fortran constant or parenthesized constant expression that produces this value.
◆ HYPOT()
| ValueWithRealFlags< RealValue > Fortran::evaluate::value::RealValue::HYPOT | ( | const RealValue & | y, |
| Rounding | rounding = TargetCharacteristics::defaultRounding ) const |
HYPOT(x,y)=SQRT(x**2 + y**2) computed so as to avoid spurious intermediate overflows.
◆ IsNull()
| bool Fortran::evaluate::value::RealValue::IsNull | ( | ) | const |
Whether this object represents a default-initialized value (zero) or unknown value.
◆ Zero()
|
static |
Creates a floating-point with value +0.0 of a given kind. In contrast, the default ctor creates a null state that represents +0.0 of unknown kind.
The documentation for this class was generated from the following files:
- include/flang/Evaluate/real-value.h
- lib/Evaluate/real-value.cpp
Generated on for FLANG by
1.14.0