Library for double-double arithmetic calculation

Extension module for numpy providing the ddouble data type.

Loading this module registers an additional scalar data type ddouble with numpy implementing double-double arithmetic. You can use use the data type by passing dtype=xprec.ddouble to numpy functions.

The xprec.linalg module provides some linear algebra subroutines, in particular QR, RRQR, SVD and truncated SVD.

Installation

$ pip install xprec

Quickstart

import numpy as np
x = np.linspace(0, np.pi)

# import double-double precision data type
from xprec import ddouble
x = x.astype(ddouble)
y = x * x + 1
z = np.sin(x)

# do some linalg
import xprec.linalg
A = np.vander(np.linspace(-1, 1, 80, dtype=ddouble), 150)
U, s, VT = xprec.linalg.svd(A)

Trouble shooting

• icc
  You may suffer from a long runtime when xprec is built with icc. If you encounter this problem, please try the following:

CFLAGS="-fp-model=precise" pip install xprec

Licence

The xprec library is Copyright (C) 2021 Markus Wallerberger. Licensed under the MIT license (see LICENSE.txt).

Contains code from the QD library, which is Copyright (C) 2012 Yozo Hida, Xiaoye S. Li, David H. Bailey. Released under a modified BSD license (see QD-LICENSE.txt).