summaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorzimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4>2016-05-28 04:59:39 +0000
committerzimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4>2016-05-28 04:59:39 +0000
commitcf11a3eb08d62789a6fdc3347da236c3c389743f (patch)
treebee39cc3ec584c505e659c2b6a6fc2af81d16ec9 /doc
parented11b5b94687c1a0cefe59f2e899c6e6f08ff589 (diff)
downloadmpfr-cf11a3eb08d62789a6fdc3347da236c3c389743f.tar.gz
start implementing RNDF (work in progress)
git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/branches/faithful@10380 280ebfd0-de03-0410-8827-d642c229c3f4
Diffstat (limited to 'doc')
-rw-r--r--doc/mpfr.texi3
1 files changed, 2 insertions, 1 deletions
diff --git a/doc/mpfr.texi b/doc/mpfr.texi
index f5fa0037c..6492967ad 100644
--- a/doc/mpfr.texi
+++ b/doc/mpfr.texi
@@ -2840,7 +2840,8 @@ In other words, if the error on @var{b} is bounded by two to the power
you should give @var{err}=@var{prec}@minus{}@var{k}.
This function @strong{does not modify} its arguments.
-If @var{rnd1} is @code{MPFR_RNDN}, the error is considered to be either
+If @var{rnd1} is @code{MPFR_RNDN} or @code{MPFR_RNDF},
+the error is considered to be either
positive or negative, thus the possible range
is twice as large as with a directed rounding for @var{rnd1} (with the
same value of @var{err}).