This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [comitted PATCH][AArch64] Regenerate libm-test-ulps


On 2016-07-18 11:45, Szabolcs Nagy wrote:
> 2016-07-18  Szabolcs Nagy  <szabolcs.nagy@arm.com>
> 
> 	* sysdeps/aarch64/libm-test-ulps: Regenerated.

This patch has reduced the aarch64 ULPs and cause the ctan_towardzero to
fail here. Note that I build using GCC 6. Here are the failures:

testing double (without inline functions)
Failure: Test: Imaginary part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          9.1013340476761850e-01   0x1.d1fd0170f2b6fp-1
 should be:   9.1013340476761828e-01   0x1.d1fd0170f2b6dp-1
 difference:  2.2204460492503130e-16   0x1.0000000000000p-52
 ulp       :  2.0000
 max.ulp   :  1.0000
Maximal error of real part of: ctan_towardzero
 is      : 5 ulp
 accepted: 5 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 1 ulp

Test suite completed:
  91530 test cases plus 82998 tests for exception flags and
    82998 tests for errno executed.
  2 errors occurred.

----------

testing double (finite-math-only)
Failure: Test: Imaginary part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          9.1013340476761850e-01   0x1.d1fd0170f2b6fp-1
 should be:   9.1013340476761828e-01   0x1.d1fd0170f2b6dp-1
 difference:  2.2204460492503130e-16   0x1.0000000000000p-52
 ulp       :  2.0000
 max.ulp   :  1.0000
Maximal error of real part of: ctan_towardzero
 is      : 5 ulp
 accepted: 5 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 1 ulp

Test suite completed:
  74558 test cases plus 67154 tests for exception flags and
    0 tests for errno executed.
  2 errors occurred.

----------

testing float (without inline functions)
Failure: Test: Real part of: ctan_towardzero (-0x2p+0 - 0x3p+0 i)
Result:
 is:          3.76402633e-03   0x1.ed5bc4p-9
 should be:   3.76402563e-03   0x1.ed5bbep-9
 difference:  6.98491930e-10   0x1.800000p-31
 ulp       :  3.0000
 max.ulp   :  2.0000
Failure: Test: Real part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          2.44635969e-01   0x1.f503b4p-3
 should be:   2.44635924e-01   0x1.f503aep-3
 difference:  4.47034835e-08   0x1.800000p-25
 ulp       :  3.0000
 max.ulp   :  2.0000
Maximal error of real part of: ctan_towardzero
 is      : 3 ulp
 accepted: 2 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 2 ulp

Test suite completed:
  62838 test cases plus 56226 tests for exception flags and
    56226 tests for errno executed.
  3 errors occurred.

----------

testing float (finite-math-only)
Failure: Test: Real part of: ctan_towardzero (-0x2p+0 - 0x3p+0 i)
Result:
 is:          3.76402633e-03   0x1.ed5bc4p-9
 should be:   3.76402563e-03   0x1.ed5bbep-9
 difference:  6.98491930e-10   0x1.800000p-31
 ulp       :  3.0000
 max.ulp   :  2.0000
Failure: Test: Real part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          2.44635969e-01   0x1.f503b4p-3
 should be:   2.44635924e-01   0x1.f503aep-3
 difference:  4.47034835e-08   0x1.800000p-25
 ulp       :  3.0000
 max.ulp   :  2.0000
Maximal error of real part of: ctan_towardzero
 is      : 3 ulp
 accepted: 2 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 2 ulp

Test suite completed:
  47034 test cases plus 41550 tests for exception flags and
    0 tests for errno executed.
  3 errors occurred.

----------

testing double (inline functions)
Failure: Test: Imaginary part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          9.1013340476761850e-01   0x1.d1fd0170f2b6fp-1
 should be:   9.1013340476761828e-01   0x1.d1fd0170f2b6dp-1
 difference:  2.2204460492503130e-16   0x1.0000000000000p-52
 ulp       :  2.0000
 max.ulp   :  1.0000
Maximal error of real part of: ctan_towardzero
 is      : 5 ulp
 accepted: 5 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 1 ulp

Test suite completed:
  90882 test cases plus 0 tests for exception flags and
    0 tests for errno executed.
  2 errors occurred.

testing float (inline functions)
Failure: Test: Real part of: ctan_towardzero (-0x2p+0 - 0x3p+0 i)
Result:
 is:          3.76402633e-03   0x1.ed5bc4p-9
 should be:   3.76402563e-03   0x1.ed5bbep-9
 difference:  6.98491930e-10   0x1.800000p-31
 ulp       :  3.0000
 max.ulp   :  2.0000
Failure: Test: Real part of: ctan_towardzero (0x8p+124 + 0x1p+0 i)
Result:
 is:          2.44635969e-01   0x1.f503b4p-3
 should be:   2.44635924e-01   0x1.f503aep-3
 difference:  4.47034835e-08   0x1.800000p-25
 ulp       :  3.0000
 max.ulp   :  2.0000
Maximal error of real part of: ctan_towardzero
 is      : 3 ulp
 accepted: 2 ulp
Maximal error of imaginary part of: ctan_towardzero
 is      : 2 ulp
 accepted: 2 ulp

Test suite completed:
  62414 test cases plus 0 tests for exception flags and
    0 tests for errno executed.
  3 errors occurred.


Aurelien

-- 
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
aurelien@aurel32.net                 http://www.aurel32.net


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]