diff options
| author | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-05-06 11:16:12 +0000 |
|---|---|---|
| committer | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-05-06 11:16:12 +0000 |
| commit | d247dd40dcf7a9dc4209a22fd89c1f47734d5ff4 (patch) | |
| tree | 483b388636c358ebbe71e9f7cabdce7e6b00209b /algorithms.tex | |
| parent | fb4847d38b82e640b08811debad6f1595b5a3cab (diff) | |
| download | mpfr-d247dd40dcf7a9dc4209a22fd89c1f47734d5ff4.tar.gz | |
My latest change was not completely correct...
git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/trunk@4447 280ebfd0-de03-0410-8827-d642c229c3f4
Diffstat (limited to 'algorithms.tex')
| -rw-r--r-- | algorithms.tex | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/algorithms.tex b/algorithms.tex index 3d02ad4bb..c9625ca73 100644 --- a/algorithms.tex +++ b/algorithms.tex @@ -826,7 +826,7 @@ $64$-bit machine). Whatever the input $x$ and $y$, it should be noted that if $\ulp(x) \geq \ulp(y)$, then $x - q y$ is always -exactly representable in the precision of $y$ if its exponent is larger +exactly representable in the precision of $y$ unless its exponent is smaller than the minimum exponent. To see this, let $\ulp(y) = 2^{-k}$; multiplying $x$ and $y$ by $2^k$ we get $X = 2^k x$ and |