Re: make check fails on ppc

[Andreas Jaeger]

address@hidden writes:

> (sorry about line_breaks)
> address@hidden:/tmp/glibc-compile/math] cat test-ifloat.out 
> testing float (inline functions)
> Failure: Test: Imaginary part of: conj (0.0 + 0.0 i) == 0.0 - 0 i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:  -0.00000000000000000000e+00  -0x0.00000000000000000000p+0
>  difference:  0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  ulp       :  0.0000
>  max.ulp   :  0.0000

This result shows that your compiler is somehow broken.

> Failure: Test: Real part of: conj (0.0 - 0 i) == 0.0 + 0.0 i
> Result:
>  is:          2.60324098722933895786e-29   0x1.08000000000000000000p-95
>  should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  difference:  2.60324098722933895786e-29   0x1.08000000000000000000p-95
>  ulp       :  0.0000
>  max.ulp   :  0.0000

another compiler error.

> Failure: Test: Imaginary part of: conj (0.0 - 0 i) == 0.0 + 0.0 i
> Result:
>  is:          2.41593346678174309295e-29   0x1.ea027000000000000000p-96
>  should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  difference:  2.41593346678174309295e-29   0x1.ea027000000000000000p-96
>  ulp       :  0.0000
>  max.ulp   :  0.0000
> Failure: Test: Real part of: conj (NaN + NaN i) == NaN + NaN i
> Result:
>  is:          5.18722576174817644632e-38   0x1.1a6b6000000000000000p-124
>  should be:   nan   nan
> Failure: Test: Imaginary part of: conj (NaN + NaN i) == NaN + NaN i
> Result:
>  is:          5.18754862091435688417e-38   0x1.1a6fe000000000000000p-124
>  should be:   nan   nan
> Failure: Test: Real part of: conj (inf - inf i) == inf + inf i
> Result:
>  is:          1.40129846432481707092e-45   0x1.00000000000000000000p-149
>  should be:   inf   inf
> Failure: Test: Imaginary part of: conj (inf - inf i) == inf + inf i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:   inf   inf
> Failure: Test: Real part of: conj (inf + inf i) == inf - inf i
> Result:
>  is:          nan   nan
>  should be:   inf   inf
> Failure: Test: Imaginary part of: conj (inf + inf i) == inf - inf i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:  -inf  -inf
> Failure: Test: Real part of: conj (1.0 + 2.0 i) == 1.0 - 2.0 i
> Result:
>  is:          5.19072340271513118973e-38   0x1.1a9c2000000000000000p-124
>  should be:   1.00000000000000000000e+00   0x1.00000000000000000000p+0
>  difference:  1.00000000000000000000e+00   0x1.00000000000000000000p+0
>  ulp       :  8388608.0000
>  max.ulp   :  0.0000
> Failure: Test: Imaginary part of: conj (1.0 + 2.0 i) == 1.0 - 2.0 i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:  -2.00000000000000000000e+00  -0x1.00000000000000000000p+1
>  difference:  2.00000000000000000000e+00   0x1.00000000000000000000p+1
>  ulp       :  8388608.0000
>  max.ulp   :  0.0000
> Failure: Test: Real part of: conj (3.0 - 4.0 i) == 3.0 + 4.0 i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:   3.00000000000000000000e+00   0x1.80000000000000000000p+1
>  difference:  3.00000000000000000000e+00   0x1.80000000000000000000p+1
>  ulp       :  12582912.0000
>  max.ulp   :  0.0000
> Failure: Test: Imaginary part of: conj (3.0 - 4.0 i) == 3.0 + 4.0 i
> Result:
>  is:          0.00000000000000000000e+00   0x0.00000000000000000000p+0
>  should be:   4.00000000000000000000e+00   0x1.00000000000000000000p+2
>  difference:  4.00000000000000000000e+00   0x1.00000000000000000000p+2
>  ulp       :  8388608.0000
>  max.ulp   :  0.0000
> Maximal error of real part of: conj
>  is      :  12582912.0000 ulp
>  accepted:  0.0000 ulp
> Maximal error of imaginary part of: conj
>  is      :  8388608.0000 ulp
>  accepted:  0.0000 ulp
> 
> Test suite completed:
>   2410 test cases plus 2208 tests for exception flags executed.
>   14 errors occurred.


All of these errors look like compiler errors.  Please try to
investigate yourself why those functions fail, they should return the
right result.

Andreas
-- 
 Andreas Jaeger
  SuSE Labs address@hidden
   private address@hidden
    http://www.suse.de/~aj