diff options
author | zimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4> | 2021-01-20 09:58:34 +0000 |
---|---|---|
committer | zimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4> | 2021-01-20 09:58:34 +0000 |
commit | 66cd6919fcf7f2a0cd33727383c89da3f9f232f6 (patch) | |
tree | 498e3910a2a79173563a1e9265e129583226c1f0 /TODO | |
parent | f49c7da0e6f4fccc41556932c259a0dd81397fd8 (diff) | |
download | mpfr-66cd6919fcf7f2a0cd33727383c89da3f9f232f6.tar.gz |
[TODO] update about powr, rootn and rsqrt
git-svn-id: https://scm.gforge.inria.fr/anonscm/svn/mpfr/trunk@14262 280ebfd0-de03-0410-8827-d642c229c3f4
Diffstat (limited to 'TODO')
-rw-r--r-- | TODO | 9 |
1 files changed, 6 insertions, 3 deletions
@@ -118,9 +118,12 @@ Table of contents: compoundn = (1+x)^n exp10m1, exp2m1, log10p1, log2p1 pown (but there is mpfr_pow_si with n long instead of intmax_t) - powr = exp(y*log(x)) - rootn (there is mpfr_rootn_ui but not mpfr_rootn_si) - maybe rename mpfr_rec_sqrt into mpfr_rsqrt? + powr = exp(y*log(x)). The difference with pow = x^y is that when x < 0 + and y is an integer, powr returns NaN, also when x=1 and y=+/Inf, + or x=+/-Inf and y=0. + rootn (there is mpfr_rootn_ui but not mpfr_rootn_si yet) + rsqrt (differs from mpfr_rec_sqrt in -0: mpfr_rec_sqrt gives +Inf whereas + rsqrt gives -Inf) - a function to compute the hash of a floating-point number (suggested by Patrick Pelissier) - implement new functions from the C++17 standard: |