[Bug math/30664] New: FAIL: math/test-double-fma

[Bug math/30664] New: FAIL: math/test-double-fma

[danglin at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

            Bug ID: 30664
           Summary: FAIL: math/test-double-fma
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: danglin at gcc dot gnu.org
  Target Milestone: ---
              Host: hppa*-*-linux*
            Target: hppa*-*-linux*
             Build: hppa*-*-linux*

This occurs with gcc version 13.1.0 (Debian 13.1.0-8).

dave@mx3210:~/gnu/glibc/objdir$ make test t=math/test-double-fma
make -r PARALLELMFLAGS="" -C ../glibc objdir=`pwd` test
make[1]: Entering directory '/home/dave/gnu/glibc/glibc'
make subdir=math -C math/ ..=../
/home/dave/gnu/glibc/objdir/math/test-double-fma.out
make[2]: Entering directory '/home/dave/gnu/glibc/glibc/math'
gcc -o /home/dave/gnu/glibc/objdir/math/test-double-fma -nostdlib -nostartfiles
    -Wl,-z,relro  /home/dave/gnu/glibc/objdir/csu/crt1.o
/home/dave/gnu/glibc/objdir/csu/crti.o `gcc  --print-file-name=crtbegin.o`
/home/dave/gnu/glibc/objdir/math/test-double-fma.o
/home/dave/gnu/glibc/objdir/support/libsupport_nonshared.a
/home/dave/gnu/glibc/objdir/math/libm-test-support-double.o
/home/dave/gnu/glibc/objdir/math/libm.so.6  -Wl,-dynamic-linker=/lib/ld.so.1
-Wl,-rpath-link=/home/dave/gnu/glibc/objdir:/home/dave/gnu/glibc/objdir/math:/home/dave/gnu/glibc/objdir/elf:/home/dave/gnu/glibc/objdir/dlfcn:/home/dave/gnu/glibc/objdir/nss:/home/dave/gnu/glibc/objdir/nis:/home/dave/gnu/glibc/objdir/rt:/home/dave/gnu/glibc/objdir/resolv:/home/dave/gnu/glibc/objdir/mathvec:/home/dave/gnu/glibc/objdir/support:/home/dave/gnu/glibc/objdir/crypt:/home/dave/gnu/glibc/objdir/nptl
-lgcc -Wl,--as-needed -lgcc_s  -Wl,--no-as-needed
/home/dave/gnu/glibc/objdir/libc.so.6
/home/dave/gnu/glibc/objdir/libc_nonshared.a -Wl,--as-needed
/home/dave/gnu/glibc/objdir/elf/ld.so -Wl,--no-as-needed -lgcc -Wl,--as-needed
-lgcc_s  -Wl,--no-as-needed `gcc  --print-file-name=crtend.o`
/home/dave/gnu/glibc/objdir/csu/crtn.o
env GCONV_PATH=/home/dave/gnu/glibc/objdir/iconvdata
LOCPATH=/home/dave/gnu/glibc/objdir/localedata LC_ALL=C  
/home/dave/gnu/glibc/objdir/elf/ld.so.1 --library-path
/home/dave/gnu/glibc/objdir:/home/dave/gnu/glibc/objdir/math:/home/dave/gnu/glibc/objdir/elf:/home/dave/gnu/glibc/objdir/dlfcn:/home/dave/gnu/glibc/objdir/nss:/home/dave/gnu/glibc/objdir/nis:/home/dave/gnu/glibc/objdir/rt:/home/dave/gnu/glibc/objdir/resolv:/home/dave/gnu/glibc/objdir/mathvec:/home/dave/gnu/glibc/objdir/support:/home/dave/gnu/glibc/objdir/crypt:/home/dave/gnu/glibc/objdir/nptl
/home/dave/gnu/glibc/objdir/math/test-double-fma  >
/home/dave/gnu/glibc/objdir/math/test-double-fma.out; \
../scripts/evaluate-test.sh math/test-double-fma $? false false >
/home/dave/gnu/glibc/objdir/math/test-double-fma.test-result
make[2]: Leaving directory '/home/dave/gnu/glibc/glibc/math'
FAIL: math/test-double-fma
original exit status 1
testing double (without inline functions)
Failure: fma (-0x7.ffffffffffffp-1024, 0x8.0000000000008p-4, -0x4p-1076):
Exception "Underflow" set
Failure: fma (0x7.ffffffffffffp-1024, 0x8.0000000000008p-4, 0x4p-1076):
Exception "Underflow" set
Failure: fma_downward (-0x4p-1076, 0x8.8p-4, -0x3.ffffffffffffcp-1024):
Exception "Underflow" set
Failure: fma_downward (-0x7.ffffffffffffp-1024, 0x8.0000000000008p-4,
-0x4p-1076): Exception "Underflow" set
Failure: Test: fma_upward (-0x3.ffffffffffffep-712, 0x3.ffffffffffffep-276,
0x3.fffffc0000ffep-984)
Result:
 is:          1.8348707892449242e-296   0x1.7ffffe00007ffp-983
 should be:   1.8348707892449245e-296   0x1.7ffffe0000800p-983
 difference:  2.7161546124355486e-312   0x0.0008000000000p-1022
 ulp       :  1.0000
 max.ulp   :  0.0000
Failure: fma_upward (0x4p-1076, 0x8.8p-4, 0x3.ffffffffffffcp-1024): Exception
"Underflow" set
Failure: fma_upward (0x7.ffffffffffffp-1024, 0x8.0000000000008p-4, 0x4p-1076):
Exception "Underflow" set

Test suite completed:
  2524 test cases plus 2520 tests for exception flags and
    2520 tests for errno executed.
  7 errors occurred.
make[1]: Leaving directory '/home/dave/gnu/glibc/glibc'

Test doesn't fail with gcc-12.

Similar fails:
FAIL: math/test-double-ldouble-fma
FAIL: math/test-float32x-float64-fma
FAIL: math/test-float32x-fma
FAIL: math/test-float64-fma
FAIL: math/test-ldouble-fma

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

[Bug math/30664] FAIL: math/test-double-fma

[danglin at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

--- Comment #1 from John David Anglin <danglin at gcc dot gnu.org> ---
If sysdeps/ieee754/dbl-64/s_fma.c is built using gcc-12, the fails don't occur.
This is likely bug in gcc-13.

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

[Bug math/30664] FAIL: math/test-double-fma

[danglin at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

--- Comment #2 from John David Anglin <danglin at gcc dot gnu.org> ---
GCC bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111709

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

[Bug math/30664] FAIL: math/test-double-fma

[sam at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

Sam James <sam at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sam at gentoo dot org

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

[Bug math/30664] FAIL: math/test-double-fma

[sam at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

Sam James <sam at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=111709

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

[Bug math/30664] FAIL: math/test-double-fma

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |NOTABUG
             Status|NEW                         |RESOLVED
                 CC|                            |adhemerval.zanella at linaro dot o
                   |                            |rg

--- Comment #3 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
As per previous comment.

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

[Bug math/30664] FAIL: math/test-double-fma

[dave.anglin at bell dot net]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

--- Comment #4 from dave.anglin at bell dot net ---
On 2023-10-11 2:43 p.m., adhemerval.zanella at linaro dot org wrote:
> Resolution|--- |NOTABUG 
Although this was most certainly introduced by a specific gcc change, it's far
from
clear that it's not a bug in s_fma.c. See:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111709#c8

We will learn more when I finish bisecting gcc.

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

[Bug math/30664] FAIL: math/test-double-fma

[matoro_bugzilla_glibc at matoro dot tk]

https://sourceware.org/bugzilla/show_bug.cgi?id=30664

matoro <matoro_bugzilla_glibc at matoro dot tk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |matoro_bugzilla_glibc@mator
                   |                            |o.tk

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