∣x∣y=et, t=ylog∣x∣, for the mantissas x=(xd,xn)2xe−xnFLINT_BITS (∣xe∣<2FLINT_BITS−3) and y=(−1)ysgn(yd,yn)2ye−ynFLINT_BITS: sets Y=(y,nk+1) (y needs nk + 2 limbs) and err such that ∣x∣y=Y2K−FLINT_BITSnk within err units, with the exponent K=(−1)returned sign(q,qn), q needing ⌈tbmax/FLINT_BITS⌉+2 limbs (qn = 0 for K=0). Instead of the sign it returns MP_REAL_POW_ONE for ∣x∣=1, MP_REAL_POW_TINY when ∣t∣<2−(FLINT_BITSnk+4), so that ∣x∣y=1+δ with 0<∣δ∣<2−(FLINT_BITSnk+3) and δ of the sign tsgn (1 for negative), or MP_REAL_POW_HUGE when ∣t∣≥2tbmax (tbmax≥FLINT_BITS−3), with the sign of t in tsgn. The logarithm runs at nk+⌈(tb+12)/FLINT_BITS⌉ limbs for the bound 2tb on ∣t∣; err is that of the exponential plus a few units.
Cite this page
The FLINT team, “_mp_real_pow_fp (mp_real.h),” FLINT: Fast Library for Number Theory, documentation, version 3.7.0-dev, 2026. https://flintlib.org/doc/mp_real/_mp_real_pow_fp.html (generated 2026-11-17).
@misc{flint-doc-mp-real-mp-real-pow-fp,
author = {{The FLINT team}},
title = {{\_mp\_real\_pow\_fp (mp\_real.h)}},
howpublished = {FLINT: Fast Library for Number Theory, documentation, version 3.7.0-dev},
year = {2026},
url = {https://flintlib.org/doc/mp_real/_mp_real_pow_fp.html},
note = {Generated 2026-11-17}
}