diff options
author | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2015-02-27 14:15:01 +0000 |
---|---|---|
committer | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2015-02-27 14:15:01 +0000 |
commit | b1f648279080e9a5c8ea4e5de8dbce0f534a6cce (patch) | |
tree | 2d7cdd2cacca4ca6894017d915dd93dbc3a297d0 | |
parent | ce134ef25aecccdf75c37ae22402596057a54ad3 (diff) | |
download | mpfr-b1f648279080e9a5c8ea4e5de8dbce0f534a6cce.tar.gz |
[doc/sum.txt] Updated TODO.
git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/branches/new-sum@9332 280ebfd0-de03-0410-8827-d642c229c3f4
-rw-r--r-- | doc/sum.txt | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/doc/sum.txt b/doc/sum.txt index 5a8201f8a..baca8d577 100644 --- a/doc/sum.txt +++ b/doc/sum.txt @@ -39,6 +39,8 @@ sum_raw (obtained from code refactoring for steps 3 to 6, and 8) should be introduced and described. Pre-rounding in (7) and rounding in (8) have been merged, and this [previously pre-rounding in (7)] is still done at the same time as sign handling with a single mpn operation. +The table concerning the correction term is also incorrect (the initial +truncation also has an influence, but it isn't taken into account yet). The general ideas of the algorithm: @@ -608,5 +610,3 @@ is regarded as important (in all cases). Moreover, while the use of two accumulators does not take much more memory (since both accumulators can occupy the same area, with a flexible limit between them), it probably makes the code a bit more complex, and noticeably slower if n is small. - -$Id: sum 75579 2014-12-17 15:59:34Z vinc17/ypig $ |