From c70e4e9c9efff9df4c847dd7cfd81bae674219ab Mon Sep 17 00:00:00 2001 From: "H.J. Lu" Date: Mon, 8 Jan 2018 08:04:26 -0800 Subject: x86-64: Add sincosf with vector FMA Since the x86-64 assembly version of sincosf is higly optimized with vector instructions, there isn't much room for improvement. However s_sincosf.c written in C with vector math and intrinsics can be optimized by GCC with FMA. On Skylake, bench-sincosf reports performance improvement: Assembly FMA improvement max 104.042 101.008 3% min 9.426 8.586 10% mean 20.6209 18.2238 13% * sysdeps/x86_64/fpu/multiarch/Makefile (libm-sysdep_routines): Add s_sincosf-sse2 and s_sincosf-fma. (CFLAGS-s_sincosf-fma.c): New. * sysdeps/x86_64/fpu/multiarch/s_sincosf-fma.c: New file. * sysdeps/x86_64/fpu/multiarch/s_sincosf-sse2.S: Likewise. * sysdeps/x86_64/fpu/multiarch/s_sincosf.c: Likewise. * sysdeps/x86_64/fpu/s_sincosf.S: Don't add alias if __sincosf is defined. --- NEWS | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'NEWS') diff --git a/NEWS b/NEWS index 0d6f826..75bf467 100644 --- a/NEWS +++ b/NEWS @@ -19,8 +19,8 @@ Major new features: supported on i386, x86_64, x32 and aarch64. * Optimized x86-64 asin, atan2, exp, expf, log, pow, atan, sin, cosf, - sinf and tan with FMA, contributed by Arjan van de Ven and H.J. Lu - from Intel. + sinf, sincosf and tan with FMA, contributed by Arjan van de Ven and + H.J. Lu from Intel. * Optimized x86-64 trunc and truncf for processors with SSE4.1. -- cgit v1.1