This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
sparc long double math failures...
- From: David Miller <davem at davemloft dot net>
- To: libc-alpha at sourceware dot org
- Cc: joseph at codesourcery dot com
- Date: Sun, 17 Mar 2013 12:38:06 -0400 (EDT)
- Subject: sparc long double math failures...
Joseph, this commit:
2013-03-16 Joseph Myers <joseph@codesourcery.com>
[BZ #15283]
* sysdeps/ieee754/ldbl-96/e_j1l.c (__ieee754_y1l): Correct test
for arguments at most half maximum finite value.
* math/libm-test.inc (j0_test): Add more tests.
(j1_test): Likewise.
(y0_test): Likewise.
(y1_test): Likewise.
* sysdeps/i386/fpu/libm-test-ulps: Update.
* sysdeps/x86_64/fpu/libm-test-ulps: Likewise.
[BZ #14155]
* sysdeps/ieee754/ldbl-128/e_j0l.c (__ieee754_j0l): Do not compute
1 / x and functions P and Q for arguments above 0x1p256L.
(__ieee754_y0l): Likewise.
* sysdeps/ieee754/ldbl-128/e_j1l.c (__ieee754_j1l): Likewise.
(__ieee754_y1l): Likewise.
* math/libm-test.inc (j0_test): Do not allow spurious underflows.
(j1_test): Likewise.
(y0_test): Likewise.
(y1_test): Likewise.
Introduced the following long double math test failures on sparc:
test-ldouble:
testing long double (without inline functions)
Failure: j0 (0x1p16383) == 9.5859502826270374691362975419147645151233e-2467: Exception "Invalid operation" se\
t
Failure: j0 (0x1p16383) == 9.5859502826270374691362975419147645151233e-2467: Exception "Overflow" set
Failure: j1 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467: Exception "Invalid operation" s\
et
Failure: j1 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467: Exception "Overflow" set
Failure: Test: j1 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467
Result:
is: nan nan
should be: -3.88955319557660206486e-2467 -0x1.b26f313524de8c1bb07cp-8194
Failure: y0 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467: Exception "Invalid operation" s\
et
Failure: y0 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467: Exception "Overflow" set
Failure: Test: y0 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467
Result:
is: -nan -nan
should be: -3.88955319557660206486e-2467 -0x1.b26f313524de8c1bb07cp-8194
Failure: y1 (0x1p16383) == -9.5859502826270374691362975419147645151233e-2467: Exception "Invalid operation" s\
et
Failure: y1 (0x1p16383) == -9.5859502826270374691362975419147645151233e-2467: Exception "Overflow" set
Test suite completed:
7016 test cases plus 6011 tests for exception flags executed.
10 errors occurred.
test-ildoubl:
testing long double (inline functions)
Failure: Test: j1 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467
Result:
is: nan nan
should be: -3.88955319557660206486e-2467 -0x1.b26f313524de8c1bb07cp-8194
Failure: Test: y0 (0x1p16383) == -3.8895531955766020648617743624167352352217e-2467
Result:
is: -nan -nan
should be: -3.88955319557660206486e-2467 -0x1.b26f313524de8c1bb07cp-8194
Test suite completed:
6949 test cases plus 5944 tests for exception flags executed.
2 errors occurred.