bigfloat
bigfloat copied to clipboard
mdickinson
Python wrapper for MPFR, providing high-precision floating-point arithmetic
Is it possible to convert a bigfloat number into its underlying byte pattern? I'm trying to convert to/from quad precision numbers between fortran/python (with bigfloat storing things python side) and...
The following code will print about 21'000 digits (which is what you expect given precision 70'000 and given that one decimal digits is about 3.3 binary digits). ```py from bigfloat...
`BigFloat` objects can't currently be pickled. For multiprocessing applications, it would be useful to be able to pickle a `BigFloat`.
We (libre-riscv.org) are thinking of using bigfloat to test a hardware floating-point implementation. It would be useful to be able to extract the exact value as a python2 long/python3 int...
[from #79] Currently, we cythonize the .pyx sources to .c files at package install time, and then include those .c files in the uploaded sdist. That causes install-time issues for...
When using the package in a coroutine, the context should be local to that coroutine. [PEP 567](https://www.python.org/dev/peps/pep-0567/) provides the mechanisms for achieving this.
`repr` output for `BigFloat` objects should show the hex form of the float instead, since recreation is then unaffected by context. `str` can continue to show a decimal string.
MPFR sets the nan flag even when propagating NaNs; this behaviour is different from the IEEE 754 specification, where propagating NaNs doesn't set the NaN flag. We should consider changing...
1. It's too many ways to do it. (We already have the with statement and the ability to set contexts directly.) 2. It needless complicates the signature of every single...
Check and document current behaviour; decide whether anything needs to change.
Metadata
Owner
Metadata
Python wrapper for MPFR, providing high-precision floating-point arithmetic