[Bug libc/14163] New: Software implementation of cosf is not decreasing on range [0,pi/2]

[Bug libc/14163] New: Software implementation of cosf is not decreasing on range [0,pi/2]

[martin.brain at cs dot ox.ac.uk]

http://sourceware.org/bugzilla/show_bug.cgi?id=14163

             Bug #: 14163
           Summary: Software implementation of cosf is not decreasing on
                    range [0,pi/2]
           Product: glibc
           Version: 2.15
            Status: NEW
          Severity: normal
          Priority: P2
         Component: libc
        AssignedTo: unassigned@sourceware.org
        ReportedBy: martin.brain@cs.ox.ac.uk
                CC: drepper.fsp@gmail.com
    Classification: Unclassified


Created attachment 6421
  --> http://sourceware.org/bugzilla/attachment.cgi?id=6421
An example of one of the glitches

While verifying parts of the software trigonometric functions using CBMC I
found that __kernel_cosf ( glibc-2.15/sysdeps/ieee754/flt-32/k_cosf.c ) has
5530 points in the range [0.1,pi/2] where inaccuracies in the last bit mean
that the function is not descreasing.  I.E. x < y but cos(x) < cos(y).  The
attached code demonstrates this effect on x86_64 (x86 uses the x87 sin and cos
hardware and so is not effected).  Interestingly the corresponding code for sin
does not have this issue.

-- 
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.

[Bug math/14163] Software implementation of cosf is not decreasing on range [0,pi/2]

[jsm28 at gcc dot gnu.org]

http://sourceware.org/bugzilla/show_bug.cgi?id=14163

Joseph Myers <jsm28 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|libc                        |math

-- 
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.

[Bug math/14163] Software implementation of cosf is not decreasing on range [0,pi/2]

[jsm28 at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=14163

Joseph Myers <jsm28 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |INVALID

--- Comment #1 from Joseph Myers <jsm28 at gcc dot gnu.org> ---
glibc accuracy goals are now documented in the manual and the FAQ and
monotonicity is explicitly excluded - without ruling out the possibility that
in future we could add correctly rounding functions under names such as crcosf,
or other functions with well-defined error bounds, probably automatically
generated with tools such as metalibm.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug math/14163] Software implementation of cosf is not decreasing on range [0,pi/2]

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=14163

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fweimer at redhat dot com
              Flags|                            |security-

-- 
You are receiving this mail because:
You are on the CC list for the bug.