Vector sincosf for x86_64 and tests.

Here is implementation of vectorized sincosf containing SSE, AVX,
AVX2 and AVX512 versions according to Vector ABI
<https://groups.google.com/forum/#!topic/x86-64-abi/LmppCfN1rZ4>.

    * NEWS: Mention addition of x86_64 vector sincosf.
    * math/test-float-vlen16.h: Added wrapper for sincosf tests.
    * math/test-float-vlen4.h: Likewise.
    * math/test-float-vlen8.h: Likewise.
    * sysdeps/unix/sysv/linux/x86_64/libmvec.abilist: New symbols added.
    * sysdeps/x86/fpu/bits/math-vector.h: Added sincosf SIMD declaration.
    * sysdeps/x86_64/fpu/Makefile (libmvec-support): Added new files.
    * sysdeps/x86_64/fpu/Versions: New versions added.
    * sysdeps/x86_64/fpu/libm-test-ulps: Regenerated.
    * sysdeps/x86_64/fpu/multiarch/Makefile (libmvec-sysdep_routines):
    Added build of SSE, AVX2 and AVX512 IFUNC versions.
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core.S
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core.S
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core_sse4.S
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core.S
    * sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core_avx2.S
    * sysdeps/x86_64/fpu/svml_s_sincosf16_core.S
    * sysdeps/x86_64/fpu/svml_s_sincosf4_core.S
    * sysdeps/x86_64/fpu/svml_s_sincosf8_core.S
    * sysdeps/x86_64/fpu/svml_s_sincosf8_core_avx.S
    * sysdeps/x86_64/fpu/svml_s_sincosf_data.S: New file.
    * sysdeps/x86_64/fpu/svml_s_sincosf_data.h: New file.
    * sysdeps/x86_64/fpu/svml_s_wrapper_impl.h: Added 3 argument wrappers.
    * sysdeps/x86_64/fpu/test-float-vlen16.c: : Vector sincosf tests.
    * sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen4.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen8-avx2.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c: Likewise.
    * sysdeps/x86_64/fpu/test-float-vlen8.c: Likewise.

1 files changed, 268 insertions, 0 deletions

diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core_sse4.S
new file mode 100644
index 0000000000..8c51e44988
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core_sse4.S
@@ -0,0 +1,268 @@
+/* Function sincosf vectorized with SSE4.
+   Copyright (C) 2014-2015 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <http://www.gnu.org/licenses/>.  */
+
+#include <sysdep.h>
+#include "svml_s_sincosf_data.h"
+
+	.text
+ENTRY (_ZGVbN4vvv_sincosf_sse4)
+/*
+   ALGORITHM DESCRIPTION:
+
+     1) Range reduction to [-Pi/4; +Pi/4] interval
+        a) Grab sign from source argument and save it.
+        b) Remove sign using AND operation
+        c) Getting octant Y by 2/Pi multiplication
+        d) Add "Right Shifter" value
+        e) Treat obtained value as integer S for destination sign setting.
+           SS = ((S-S&1)&2)<<30; For sin part
+           SC = ((S+S&1)&2)<<30; For cos part
+        f) Change destination sign if source sign is negative
+           using XOR operation.
+        g) Subtract "Right Shifter" (0x4B000000) value
+        h) Subtract Y*(PI/2) from X argument, where PI/2 divided to 4 parts:
+           X = X - Y*PI1 - Y*PI2 - Y*PI3 - Y*PI4;
+     2) Polynomial (minimax for sin within  [-Pi/4; +Pi/4] interval)
+        a) Calculate X^2 = X * X
+        b) Calculate 2 polynomials for sin and cos:
+           RS = X * ( A0 + X^2 * (A1 + x^2 * (A2 + x^2 * (A3))));
+           RC = B0 + X^2 * (B1 + x^2 * (B2 + x^2 * (B3 + x^2 * (B4))));
+        c) Swap RS & RC if if first bit of obtained value after
+           Right Shifting is set to 1. Using And, Andnot & Or operations.
+     3) Destination sign setting
+        a) Set shifted destination sign using XOR operation:
+           R1 = XOR( RS, SS );
+           R2 = XOR( RC, SC ).  */
+
+        pushq     %rbp
+        cfi_adjust_cfa_offset (8)
+        cfi_rel_offset (%rbp, 0)
+        movq      %rsp, %rbp
+        cfi_def_cfa_register (%rbp)
+        andq      $-64, %rsp
+        subq      $320, %rsp
+        movq      __svml_ssincos_data@GOTPCREL(%rip), %rax
+        movups    %xmm12, 176(%rsp)
+        movups    %xmm9, 160(%rsp)
+        movups __sAbsMask(%rax), %xmm12
+
+/* Absolute argument computation */
+        movaps    %xmm12, %xmm5
+        andnps    %xmm0, %xmm12
+        movups __sInvPI(%rax), %xmm7
+        andps     %xmm0, %xmm5
+
+/* c) Getting octant Y by 2/Pi multiplication
+   d) Add "Right Shifter" value.  */
+        mulps     %xmm5, %xmm7
+        movups    %xmm10, 144(%rsp)
+        movups __sPI1(%rax), %xmm10
+
+/* h) Subtract Y*(PI/2) from X argument, where PI/2 divided to 3 parts:
+      X = X - Y*PI1 - Y*PI2 - Y*PI3.  */
+        movaps    %xmm10, %xmm1
+        addps __sRShifter(%rax), %xmm7
+
+/* e) Treat obtained value as integer S for destination sign setting */
+        movaps    %xmm7, %xmm9
+
+/* g) Subtract "Right Shifter" (0x4B000000) value */
+        subps __sRShifter(%rax), %xmm7
+        mulps     %xmm7, %xmm1
+        pslld     $31, %xmm9
+        movups __sPI2(%rax), %xmm6
+        movups    %xmm13, 112(%rsp)
+        movaps    %xmm5, %xmm13
+        movaps    %xmm6, %xmm2
+        subps     %xmm1, %xmm13
+        mulps     %xmm7, %xmm2
+        movups __sSignMask(%rax), %xmm3
+        movaps    %xmm5, %xmm1
+        movups __sOneHalf(%rax), %xmm4
+        subps     %xmm2, %xmm13
+        cmpnleps __sRangeReductionVal(%rax), %xmm5
+        movaps    %xmm3, %xmm2
+        andps     %xmm13, %xmm2
+        xorps     %xmm2, %xmm4
+
+/* Result sign calculations */
+        xorps     %xmm2, %xmm3
+        xorps     %xmm9, %xmm3
+
+/* Add correction term 0.5 for cos() part */
+        addps     %xmm7, %xmm4
+        movmskps  %xmm5, %ecx
+        mulps     %xmm4, %xmm10
+        mulps     %xmm4, %xmm6
+        subps     %xmm10, %xmm1
+        movups __sPI3(%rax), %xmm10
+        subps     %xmm6, %xmm1
+        movaps    %xmm10, %xmm6
+        mulps     %xmm7, %xmm6
+        mulps     %xmm4, %xmm10
+        subps     %xmm6, %xmm13
+        subps     %xmm10, %xmm1
+        movups __sPI4(%rax), %xmm6
+        mulps     %xmm6, %xmm7
+        mulps     %xmm6, %xmm4
+        subps     %xmm7, %xmm13
+        subps     %xmm4, %xmm1
+        xorps     %xmm9, %xmm13
+        xorps     %xmm3, %xmm1
+        movaps    %xmm13, %xmm4
+        movaps    %xmm1, %xmm2
+        mulps     %xmm13, %xmm4
+        mulps     %xmm1, %xmm2
+        movups __sA9(%rax), %xmm7
+
+/* 2) Polynomial (minimax for sin within  [-Pi/4; +Pi/4] interval)
+      a) Calculate X^2 = X * X
+      b) Calculate 2 polynomials for sin and cos:
+         RS = X * ( A0 + X^2 * (A1 + x^2 * (A2 + x^2 * (A3))));
+         RC = B0 + X^2 * (B1 + x^2 * (B2 + x^2 * (B3 + x^2 * (B4)))) */
+        movaps    %xmm7, %xmm3
+        mulps     %xmm4, %xmm3
+        mulps     %xmm2, %xmm7
+        addps __sA7(%rax), %xmm3
+        addps __sA7(%rax), %xmm7
+        mulps     %xmm4, %xmm3
+        mulps     %xmm2, %xmm7
+        addps __sA5(%rax), %xmm3
+        addps __sA5(%rax), %xmm7
+        mulps     %xmm4, %xmm3
+        mulps     %xmm2, %xmm7
+        addps __sA3(%rax), %xmm3
+        addps __sA3(%rax), %xmm7
+        mulps     %xmm3, %xmm4
+        mulps     %xmm7, %xmm2
+        mulps     %xmm13, %xmm4
+        mulps     %xmm1, %xmm2
+        addps     %xmm4, %xmm13
+        addps     %xmm2, %xmm1
+        xorps     %xmm12, %xmm13
+        testl     %ecx, %ecx
+        jne       .LBL_1_3
+
+.LBL_1_2:
+        cfi_remember_state
+        movups    160(%rsp), %xmm9
+        movaps    %xmm13, (%rdi)
+        movups    144(%rsp), %xmm10
+        movups    176(%rsp), %xmm12
+        movups    112(%rsp), %xmm13
+        movups    %xmm1, (%rsi)
+        movq      %rbp, %rsp
+        cfi_def_cfa_register (%rsp)
+        popq      %rbp
+        cfi_adjust_cfa_offset (-8)
+        cfi_restore (%rbp)
+        ret
+
+.LBL_1_3:
+        cfi_restore_state
+        movups    %xmm0, 128(%rsp)
+        movups    %xmm13, 192(%rsp)
+        movups    %xmm1, 256(%rsp)
+        je        .LBL_1_2
+
+        xorb      %dl, %dl
+        xorl      %eax, %eax
+        movups    %xmm8, 48(%rsp)
+        movups    %xmm11, 32(%rsp)
+        movups    %xmm14, 16(%rsp)
+        movups    %xmm15, (%rsp)
+        movq      %rsi, 64(%rsp)
+        movq      %r12, 104(%rsp)
+        cfi_offset_rel_rsp (12, 104)
+        movb      %dl, %r12b
+        movq      %r13, 96(%rsp)
+        cfi_offset_rel_rsp (13, 96)
+        movl      %eax, %r13d
+        movq      %r14, 88(%rsp)
+        cfi_offset_rel_rsp (14, 88)
+        movl      %ecx, %r14d
+        movq      %r15, 80(%rsp)
+        cfi_offset_rel_rsp (15, 80)
+        movq      %rbx, 72(%rsp)
+        movq      %rdi, %rbx
+        cfi_remember_state
+
+.LBL_1_6:
+        btl       %r13d, %r14d
+        jc        .LBL_1_13
+
+.LBL_1_7:
+        lea       1(%r13), %esi
+        btl       %esi, %r14d
+        jc        .LBL_1_10
+
+.LBL_1_8:
+        incb      %r12b
+        addl      $2, %r13d
+        cmpb      $16, %r12b
+        jb        .LBL_1_6
+
+        movups    48(%rsp), %xmm8
+        movq      %rbx, %rdi
+        movups    32(%rsp), %xmm11
+        movups    16(%rsp), %xmm14
+        movups    (%rsp), %xmm15
+        movq      64(%rsp), %rsi
+        movq      104(%rsp), %r12
+        cfi_restore (%r12)
+        movq      96(%rsp), %r13
+        cfi_restore (%r13)
+        movq      88(%rsp), %r14
+        cfi_restore (%r14)
+        movq      80(%rsp), %r15
+        cfi_restore (%r15)
+        movq      72(%rsp), %rbx
+        movups    192(%rsp), %xmm13
+        movups    256(%rsp), %xmm1
+        jmp       .LBL_1_2
+
+.LBL_1_10:
+        cfi_restore_state
+        movzbl    %r12b, %r15d
+        movss     132(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        movss     %xmm0, 196(%rsp,%r15,8)
+        movss     132(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        movss     %xmm0, 260(%rsp,%r15,8)
+        jmp       .LBL_1_8
+
+.LBL_1_13:
+        movzbl    %r12b, %r15d
+        movss     128(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        movss     %xmm0, 192(%rsp,%r15,8)
+        movss     128(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        movss     %xmm0, 256(%rsp,%r15,8)
+        jmp       .LBL_1_7
+
+END (_ZGVbN4vvv_sincosf_sse4)