Make the default context equal to IEEE 754 quadruple precision.

mdickinson / bigfloat

Python wrapper for MPFR, providing high-precision floating-point arithmetic

GNU Lesser General Public License v3.0

41 stars 11 forks source link

Closed mdickinson closed 8 years ago

mdickinson commented 8 years ago

This PR replaces the current default context with IEEE 754 quadruple precision, so that:

users get extra precision out of the box (there's little value in the original 53-bit precision, since floats already provide that)
the exponent min and max are consistent across platforms (the previous values were based on the MPFR limits, which may vary across platforms).

Closes #17.