This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/16503] [SH] wrong results on fma_towardzero
- From: "kkojima at rr dot iij4u.or.jp" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Mon, 27 Jan 2014 00:05:13 +0000
- Subject: [Bug math/16503] [SH] wrong results on fma_towardzero
- Auto-submitted: auto-generated
- References: <bug-16503-131 at http dot sourceware dot org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=16503
Kaz Kojima <kkojima at rr dot iij4u.or.jp> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[SH] wrong results on fma |[SH] wrong results on
| |fma_towardzero
--- Comment #1 from Kaz Kojima <kkojima at rr dot iij4u.or.jp> ---
It turned out that these errors are bogus result without -mieee SH
gcc option which is required for proper IEEE arithmetic on this target.
Although the lines for fma went away with enabling -mieee, there are
new fma_towardzero lines:
# fma_towardzero
Test "fma_towardzero (-0x1.ffffffffffffcp-1022, 0x1.0000000000001p-1,
-0x1p-1074)":
double: 1
idouble: 1
Test "fma_towardzero (-0x1p-1074, 0x1.1p-1, -0x0.fffffffffffffp-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1.ffffffffffffcp-1022, 0x1.0000000000001p-1,
0x1p-1074)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, -0x1p-1074, 0x0.fffffffffffffp-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, -0x1p-1074, 0x1p-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, -0x1p-1074, 0x1p-1074)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, -0x1p-1074, 0x1p1023)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, 0x1.1p-1, 0x0.fffffffffffffp-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, 0x1p-1074, -0x0.fffffffffffffp-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, 0x1p-1074, -0x1p-1022)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, 0x1p-1074, -0x1p-1074)":
double: 1
idouble: 1
Test "fma_towardzero (0x1p-1074, 0x1p-1074, -0x1p1023)":
double: 1
idouble: 1
--
You are receiving this mail because:
You are on the CC list for the bug.