[Bug math/18197] New: atan inaccuracy

[Bug math/18197] New: atan inaccuracy

[jsm28 at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=18197

            Bug ID: 18197
           Summary: atan inaccuracy
           Product: glibc
           Version: 2.21
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: jsm28 at gcc dot gnu.org

The dbl-64 version of atan is inaccurate in non-default rounding modes, with
errors exceeding the 9ulp maximum considered acceptable by libm-test.inc.  For
example (on x86_64):

Failure: Test: atan_upward (-0x3.b02d84p-4)
Result:
 is:         -2.26554758731449840114e-01  -0x1.cffbf0fc09b910000000p-3
 should be:  -2.26554758731449507047e-01  -0x1.cffbf0fc09b850000000p-3
 difference:  3.33066907387546962128e-16   0x1.80000000000000000000p-52
 ulp       :  12.0000
 max.ulp   :  1.0000

Failure: Test: atan_towardzero (-0x3.3fb708p-4)
Result:
 is:         -2.00331721927011846551e-01  -0x1.9a478490391a30000000p-3
 should be:  -2.00331721927011541239e-01  -0x1.9a478490391980000000p-3
 difference:  3.05311331771918048616e-16   0x1.60000000000000000000p-52
 ulp       :  11.0000
 max.ulp   :  1.0000

Failure: Test: atan_downward (-0x3.3fb708p-4)
Result:
 is:         -2.00331721927011846552e-01  -0x1.9a478490391a30000000p-3
 should be:  -2.00331721927011568996e-01  -0x1.9a478490391990000000p-3
 difference:  2.77555756156289135105e-16   0x1.40000000000000000000p-52
 ulp       :  10.0000
 max.ulp   :  1.0000

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

[Bug math/18197] atan inaccuracy

[jsm28 at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=18197

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

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

--- Comment #2 from Joseph Myers <jsm28 at gcc dot gnu.org> ---
Fixed for 2.22.

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

[Bug math/18197] atan inaccuracy

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=18197

--- Comment #1 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU C Library master sources".

The branch, master has been updated
       via  ae63c7ebedcaa57017df4cb8ff0494a4705321ee (commit)
      from  d5856d06c3ef62907f2865c502ada4baf3c9efa1 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=ae63c7ebedcaa57017df4cb8ff0494a4705321ee

commit ae63c7ebedcaa57017df4cb8ff0494a4705321ee
Author: Joseph Myers <joseph@codesourcery.com>
Date:   Wed Apr 8 17:14:12 2015 +0000

    Fix dbl-64 atan in non-default rounding modes (bug 18197).

    The dbl-64 implementation of atan does computations that expect to run
    in round-to-nearest mode, and in other modes the errors can accumulate
    to more than the maximum accepted 9ulp.  This patch makes it use
    FE_TONEAREST internally, similar to other functions with such issues.

    Tested for x86_64 and x86; no ulps updates needed.

        [BZ #18197]
        * sysdeps/ieee754/dbl-64/s_atan.c: Include <fenv.h>.
        (atan): Set FE_TONEAREST mode for internal computations.
        * math/auto-libm-test-in: Add more tests of atan.
        * math/auto-libm-test-out: Regenerated.

-----------------------------------------------------------------------

Summary of changes:
 ChangeLog                       |    8 ++++++
 NEWS                            |    2 +-
 math/auto-libm-test-in          |    2 +
 math/auto-libm-test-out         |   50 +++++++++++++++++++++++++++++++++++++++
 sysdeps/ieee754/dbl-64/s_atan.c |    2 +
 5 files changed, 63 insertions(+), 1 deletions(-)

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