This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[PATCH 0/3] math: Update s_sincosf
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: libc-alpha at sourceware dot org
- Date: Tue, 11 Dec 2018 12:24:31 -0800
- Subject: [PATCH 0/3] math: Update s_sincosf
On x86-64, the generic s_sincosf.c is faster than s_sincosf-sse2.S
as well as the previous s_sincosf-fma.c. This set of patches updates
s_sincosf.h to use generic vector computations, use generic s_sincosf.c
for s_sincosf-fma.c and removes s_sincosf-sse2.S.
H.J. Lu (3):
Use generic vector computations in s_sincosf.h
Regenerate sysdeps/x86_64/fpu/libm-test-ulps
x86-64: Remove s_sincosf-sse2.S
sysdeps/ieee754/flt-32/s_sincosf.h | 51 +-
sysdeps/ieee754/flt-32/s_sincosf_data.c | 18 +-
sysdeps/ieee754/flt-32/v2df_to_sf.h | 24 +
sysdeps/powerpc/Makefile | 4 +
sysdeps/x86/fpu/v2df_to_sf.h | 31 +
sysdeps/x86_64/fpu/libm-test-ulps | 6 +
sysdeps/x86_64/fpu/multiarch/s_sincosf-fma.c | 271 +--------
sysdeps/x86_64/fpu/multiarch/s_sincosf-sse2.S | 2 -
sysdeps/x86_64/fpu/multiarch/s_sincosf-sse2.c | 2 +
sysdeps/x86_64/fpu/s_sincosf.S | 566 ------------------
10 files changed, 103 insertions(+), 872 deletions(-)
create mode 100644 sysdeps/ieee754/flt-32/v2df_to_sf.h
create mode 100644 sysdeps/x86/fpu/v2df_to_sf.h
delete mode 100644 sysdeps/x86_64/fpu/multiarch/s_sincosf-sse2.S
create mode 100644 sysdeps/x86_64/fpu/multiarch/s_sincosf-sse2.c
delete mode 100644 sysdeps/x86_64/fpu/s_sincosf.S
--
2.19.2