[Bug math/19790] New: [ldbl-128ibm] nearbyintl incorrect in non-default rounding modes

["jsm28 at gcc dot gnu.org" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=19790

            Bug ID: 19790
           Summary: [ldbl-128ibm] nearbyintl incorrect in non-default
                    rounding modes
           Product: glibc
           Version: 2.23
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: jsm28 at gcc dot gnu.org
  Target Milestone: ---

The ldbl-128ibm implementation of nearbyintl is incorrect in non-default
rounding modes, resulting in many failures such as:

Failure: Test: nearbyint_downward (3.5)
Result:
 is:          4.00000000000000000000000000000000e+00  
0x1.000000000000000000000000000p+2
 should be:   3.00000000000000000000000000000000e+00  
0x1.800000000000000000000000000p+1
 difference:  1.00000000000000000000000000000000e+00  
0x1.000000000000000000000000000p+0
 ulp       :  20282409603651670423947251286016.0000
 max.ulp   :  0.0000

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