On Mit, 2011-10-05 at 15:02 -0700, Jeremy Huddleston wrote:
Co-authored-by: Jeremy Huddleston jerem...@apple.com
Signed-off-by: Peter Hutterer peter.hutte...@who-t.net
[...]
+frac = ldexp((double)(in 0x), -16);
The compiler can't optimize away these ldexp() calls.
+tmp = (in
Co-authored-by: Jeremy Huddleston jerem...@apple.com
Signed-off-by: Peter Hutterer peter.hutte...@who-t.net
---
Attached is a self-contained version of this called test.c
Changes from v2:
1UUL instead of ldexp
Cleanup to make style match in both cases
Improved testing
Fixed a bug with negatives
On Wed, Oct 05, 2011 at 03:02:52PM -0700, Jeremy Huddleston wrote:
Co-authored-by: Jeremy Huddleston jerem...@apple.com
Signed-off-by: Peter Hutterer peter.hutte...@who-t.net
---
Attached is a self-contained version of this called test.c
Changes from v2:
1UUL instead of ldexp
Cleanup to
