diff options
author | enge <enge@211d60ee-9f03-0410-a15a-8952a2c7a4e4> | 2011-09-02 16:17:43 +0000 |
---|---|---|
committer | enge <enge@211d60ee-9f03-0410-a15a-8952a2c7a4e4> | 2011-09-02 16:17:43 +0000 |
commit | ec4abbab5bfd18c3b1a1a642499873384d7da336 (patch) | |
tree | ae86df0aac647c5a7677848a4f5fc22587d74c7d /TODO | |
parent | 20c497b30de133737ab0814432d621ecd3f69d6c (diff) | |
download | mpc-ec4abbab5bfd18c3b1a1a642499873384d7da336.tar.gz |
div.c: more stringent handling of over- and underflows
div.dat: added tests with intermediate over- or underflows
The result of div may be wrong in such cases.
git-svn-id: svn://scm.gforge.inria.fr/svn/mpc/trunk@1074 211d60ee-9f03-0410-a15a-8952a2c7a4e4
Diffstat (limited to 'TODO')
-rw-r--r-- | TODO | 1 |
1 files changed, 0 insertions, 1 deletions
@@ -3,7 +3,6 @@ return the exact ternary value in mpc_div: - for directed rounding modes it is easy (assuming we are able to detect exact cases) - rounding to nearest to n bits is equivalent to directed rounding to n+1 bits -Also for mpc_abs. From Andreas Enge 31 August 2011: implement mul_karatsuba with three multiplications at precision around p, |