This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/12588] New: fmaf gives incorrect results on x86-64
- From: "nwhitehe at mailinator dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sources dot redhat dot com
- Date: Wed, 16 Mar 2011 01:52:45 +0000
- Subject: [Bug math/12588] New: fmaf gives incorrect results on x86-64
- Auto-submitted: auto-generated
http://sourceware.org/bugzilla/show_bug.cgi?id=12588
Summary: fmaf gives incorrect results on x86-64
Product: glibc
Version: 2.11
Status: NEW
Severity: normal
Priority: P2
Component: math
AssignedTo: aj@suse.de
ReportedBy: nwhitehe@mailinator.com
Created attachment 5313
--> http://sourceware.org/bugzilla/attachment.cgi?id=5313
simple fmaf() test case
The fmaf() function appears to return incorrect results for some inputs. A
small test case shows that the result of the operation is not computed as if in
infinite precision and then rounded once as the specification requires.
The test case is attached. It computes fmaf(-0x1.6a0a06p-1, 0x1.6a0a06p-1,
0x1.00002ep-1). On my x86-64 Ubuntu 10.04 system I get a result of 0x1p-24.
The correct result is 0x1.51c3eep-25.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.