[Bug math/26982] New: [x86_64] x86_64 lgamma has too large ULP error

[Bug math/26982] New: [x86_64] x86_64 lgamma has too large ULP error

[zimmerma+gcc at loria dot fr]

https://sourceware.org/bugzilla/show_bug.cgi?id=26982

            Bug ID: 26982
           Summary: [x86_64] x86_64 lgamma has too large ULP error
           Product: glibc
           Version: 2.32
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: zimmerma+gcc at loria dot fr
                CC: joseph at codesourcery dot com
  Target Milestone: ---

$ cat test.c
#include <stdio.h>
#include <math.h>

int main()
{
  double x = -0x1.f60c969a239f2p+1, y;
  y = lgamma (x);
  printf ("x=%a y=%a\n", x, y);
}
$ gcc -fno-builtin test.c -lm
$ ./a.out
x=-0x1.f60c969a239f2p+1 y=-0x1.fd3d359162ae4p-2

The correctly rounded result is -0x1.fd3d359162adap-2, thus the error is 10
ulps, which is larger than the maximum accepted limit of 9 ulps.

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

[Bug math/26982] [x86_64] x86_64 lgamma has too large ULP error

[woodard at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26982

Ben Woodard <woodard at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |woodard at redhat dot com

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

[Bug math/26982] [x86_64] x86_64 lgamma has too large ULP error

[zimmerma+gcc at loria dot fr]

https://sourceware.org/bugzilla/show_bug.cgi?id=26982

--- Comment #1 from Paul Zimmermann <zimmerma+gcc at loria dot fr> ---
I also get an error of 11.8884 ulps in the binary128 format for
x=-0x3.ec1bd0fc6a9c6bf9b04c0bcd1a96p+0

(from https://members.loria.fr/PZimmermann/papers/accuracy.pdf)

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

[Bug math/26982] [x86_64] x86_64 lgamma has too large ULP error

[zimmerma+gcc at loria dot fr]

https://sourceware.org/bugzilla/show_bug.cgi?id=26982

--- Comment #2 from Paul Zimmermann <zimmerma+gcc at loria dot fr> ---
for long double on x86_64 (ldbl-96) I get an error of 11.4817 ulps
for x = -0x3.ec953aa5efab0258p+0l with glibc 2.32.

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