BigFloat¶

Arbitrary-precision floating point number built on mpf_class.

Header: unumber/float/BigFloat.h

Precision model¶

Each instance has a precision in bits. Operations pick max(lhs.precision, rhs.precision) for the result.

mp_bitcnt_t getPrecision() const noexcept;
void setPrecision(mp_bitcnt_t prec);

Construction¶

BigFloat();                                      // 0, 256 bits
BigFloat(const std::string& s, mp_bitcnt_t p=256);
BigFloat(const char* s, mp_bitcnt_t p=256);
BigFloat(double v, mp_bitcnt_t p=256);
BigFloat(const mpf_class& v);
BigFloat(const BigInteger& v, mp_bitcnt_t p=256);

Formatting¶

std::string toString() const;                    // fixed with digits derived from precision
std::string toStringScientific() const;          // mantissa + e10
std::string toStringFixed(int digits) const;     // fixed digits after decimal

Operators¶

+ - * /, compounds, comparisons, stream << >>.

Notes:

toString() is intended as a reasonable default; use toStringScientific() for stable output across magnitudes.
Input parsing uses base-10 strings via mpf_set_str.

Example¶

using usub::mp::BigFloat;

BigFloat x("1.0", 512);
BigFloat y("3.0", 512);
auto z = x / y;

std::cout << z.toStringScientific() << "\n";
std::cout << z.toStringFixed(50) << "\n";