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.

7 files changed, 1130 insertions, 1 deletions

diff --git a/sysdeps/x86_64/fpu/multiarch/Makefile b/sysdeps/x86_64/fpu/multiarch/Makefile
index 9e510db365..86ea473b4f 100644
--- a/sysdeps/x86_64/fpu/multiarch/Makefile
+++ b/sysdeps/x86_64/fpu/multiarch/Makefile
@@ -69,5 +69,6 @@ libmvec-sysdep_routines += svml_d_cos2_core_sse4 svml_d_cos4_core_avx2 \
 			   svml_s_expf16_core_avx512 svml_d_pow2_core_sse4 \
 			   svml_d_pow4_core_avx2 svml_d_pow8_core_avx512 \
 			   svml_s_powf4_core_sse4 svml_s_powf8_core_avx2 \
-			   svml_s_powf16_core_avx512
+			   svml_s_powf16_core_avx512 svml_s_sincosf4_core_sse4 \
+			   svml_s_sincosf8_core_avx2 svml_s_sincosf16_core_avx512
 endif
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core.S
new file mode 100644
index 0000000000..0a1753eab7
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core.S
@@ -0,0 +1,39 @@
+/* Multiple versions of vectorized sincosf.
+   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 <init-arch.h>
+
+	.text
+ENTRY (_ZGVeN16vvv_sincosf)
+        .type   _ZGVeN16vvv_sincosf, @gnu_indirect_function
+        cmpl    $0, KIND_OFFSET+__cpu_features(%rip)
+        jne     1
+        call    __init_cpu_features
+1:      leaq    _ZGVeN16vvv_sincosf_skx(%rip), %rax
+        testl   $bit_AVX512DQ_Usable, __cpu_features+FEATURE_OFFSET+index_AVX512DQ_Usable(%rip)
+        jnz     3
+2:      leaq    _ZGVeN16vvv_sincosf_knl(%rip), %rax
+        testl   $bit_AVX512F_Usable, __cpu_features+FEATURE_OFFSET+index_AVX512F_Usable(%rip)
+        jnz     3
+        leaq    _ZGVeN16vvv_sincosf_avx2_wrapper(%rip), %rax
+3:      ret
+END (_ZGVeN16vvv_sincosf)
+
+#define _ZGVeN16vvv_sincosf _ZGVeN16vvv_sincosf_avx2_wrapper
+#include "../svml_s_sincosf16_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S
new file mode 100644
index 0000000000..cae49f63a6
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S
@@ -0,0 +1,504 @@
+/* Function sincosf vectorized with AVX-512. KNL and SKX versions.
+   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"
+#include "svml_s_wrapper_impl.h"
+
+/*
+   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 ).  */
+
+	.text
+ENTRY (_ZGVeN16vvv_sincosf_knl)
+#ifndef HAVE_AVX512_ASM_SUPPORT
+WRAPPER_IMPL_AVX512_fFF _ZGVdN8vvv_sincosf
+#else
+        pushq     %rbp
+        cfi_adjust_cfa_offset (8)
+        cfi_rel_offset (%rbp, 0)
+        movq      %rsp, %rbp
+        cfi_def_cfa_register (%rbp)
+        andq      $-64, %rsp
+        subq      $1344, %rsp
+        movq      __svml_ssincos_data@GOTPCREL(%rip), %rax
+        vmovaps   %zmm0, %zmm2
+        movl      $-1, %edx
+        vmovups __sAbsMask(%rax), %zmm0
+        vmovups __sInvPI(%rax), %zmm3
+
+/* Absolute argument computation */
+        vpandd    %zmm0, %zmm2, %zmm1
+        vmovups __sPI1_FMA(%rax), %zmm5
+        vmovups __sSignMask(%rax), %zmm9
+        vpandnd   %zmm2, %zmm0, %zmm0
+
+/* h) Subtract Y*(PI/2) from X argument, where PI/2 divided to 3 parts:
+      X = X - Y*PI1 - Y*PI2 - Y*PI3 */
+        vmovaps   %zmm1, %zmm6
+        vmovaps   %zmm1, %zmm8
+
+/* c) Getting octant Y by 2/Pi multiplication
+   d) Add "Right Shifter" value */
+        vfmadd213ps __sRShifter(%rax), %zmm1, %zmm3
+        vmovups __sPI3_FMA(%rax), %zmm7
+
+/* g) Subtract "Right Shifter" (0x4B000000) value */
+        vsubps __sRShifter(%rax), %zmm3, %zmm12
+
+/* e) Treat obtained value as integer S for destination sign setting */
+        vpslld    $31, %zmm3, %zmm13
+        vmovups __sA7_FMA(%rax), %zmm14
+        vfnmadd231ps %zmm12, %zmm5, %zmm6
+
+/* 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)))) */
+        vmovaps   %zmm14, %zmm15
+        vmovups __sA9_FMA(%rax), %zmm3
+        vcmpps    $22, __sRangeReductionVal(%rax), %zmm1, %k1
+        vpbroadcastd %edx, %zmm1{%k1}{z}
+        vfnmadd231ps __sPI2_FMA(%rax), %zmm12, %zmm6
+        vptestmd  %zmm1, %zmm1, %k0
+        vpandd    %zmm6, %zmm9, %zmm11
+        kmovw     %k0, %ecx
+        vpxord __sOneHalf(%rax), %zmm11, %zmm4
+
+/* Result sign calculations */
+        vpternlogd $150, %zmm13, %zmm9, %zmm11
+
+/* Add correction term 0.5 for cos() part */
+        vaddps    %zmm4, %zmm12, %zmm10
+        vfnmadd213ps %zmm6, %zmm7, %zmm12
+        vfnmadd231ps %zmm10, %zmm5, %zmm8
+        vpxord    %zmm13, %zmm12, %zmm13
+        vmulps    %zmm13, %zmm13, %zmm12
+        vfnmadd231ps __sPI2_FMA(%rax), %zmm10, %zmm8
+        vfmadd231ps __sA9_FMA(%rax), %zmm12, %zmm15
+        vfnmadd213ps %zmm8, %zmm7, %zmm10
+        vfmadd213ps __sA5_FMA(%rax), %zmm12, %zmm15
+        vpxord    %zmm11, %zmm10, %zmm5
+        vmulps    %zmm5, %zmm5, %zmm4
+        vfmadd213ps __sA3(%rax), %zmm12, %zmm15
+        vfmadd213ps %zmm14, %zmm4, %zmm3
+        vmulps    %zmm12, %zmm15, %zmm14
+        vfmadd213ps __sA5_FMA(%rax), %zmm4, %zmm3
+        vfmadd213ps %zmm13, %zmm13, %zmm14
+        vfmadd213ps __sA3(%rax), %zmm4, %zmm3
+        vpxord    %zmm0, %zmm14, %zmm0
+        vmulps    %zmm4, %zmm3, %zmm3
+        vfmadd213ps %zmm5, %zmm5, %zmm3
+        testl     %ecx, %ecx
+        jne       .LBL_1_3
+
+.LBL_1_2:
+        cfi_remember_state
+        vmovups   %zmm0, (%rdi)
+        vmovups   %zmm3, (%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
+        vmovups   %zmm2, 1152(%rsp)
+        vmovups   %zmm0, 1216(%rsp)
+        vmovups   %zmm3, 1280(%rsp)
+        je        .LBL_1_2
+
+        xorb      %dl, %dl
+        kmovw     %k4, 1048(%rsp)
+        xorl      %eax, %eax
+        kmovw     %k5, 1040(%rsp)
+        kmovw     %k6, 1032(%rsp)
+        kmovw     %k7, 1024(%rsp)
+        vmovups   %zmm16, 960(%rsp)
+        vmovups   %zmm17, 896(%rsp)
+        vmovups   %zmm18, 832(%rsp)
+        vmovups   %zmm19, 768(%rsp)
+        vmovups   %zmm20, 704(%rsp)
+        vmovups   %zmm21, 640(%rsp)
+        vmovups   %zmm22, 576(%rsp)
+        vmovups   %zmm23, 512(%rsp)
+        vmovups   %zmm24, 448(%rsp)
+        vmovups   %zmm25, 384(%rsp)
+        vmovups   %zmm26, 320(%rsp)
+        vmovups   %zmm27, 256(%rsp)
+        vmovups   %zmm28, 192(%rsp)
+        vmovups   %zmm29, 128(%rsp)
+        vmovups   %zmm30, 64(%rsp)
+        vmovups   %zmm31, (%rsp)
+        movq      %rsi, 1056(%rsp)
+        movq      %r12, 1096(%rsp)
+        cfi_offset_rel_rsp (12, 1096)
+        movb      %dl, %r12b
+        movq      %r13, 1088(%rsp)
+        cfi_offset_rel_rsp (13, 1088)
+        movl      %eax, %r13d
+        movq      %r14, 1080(%rsp)
+        cfi_offset_rel_rsp (14, 1080)
+        movl      %ecx, %r14d
+        movq      %r15, 1072(%rsp)
+        cfi_offset_rel_rsp (15, 1072)
+        movq      %rbx, 1064(%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:
+        addb      $1, %r12b
+        addl      $2, %r13d
+        cmpb      $16, %r12b
+        jb        .LBL_1_6
+
+        movq      %rbx, %rdi
+        kmovw     1048(%rsp), %k4
+        movq      1056(%rsp), %rsi
+        kmovw     1040(%rsp), %k5
+        movq      1096(%rsp), %r12
+        cfi_restore (%r12)
+        kmovw     1032(%rsp), %k6
+        movq      1088(%rsp), %r13
+        cfi_restore (%r13)
+        kmovw     1024(%rsp), %k7
+        vmovups   960(%rsp), %zmm16
+        vmovups   896(%rsp), %zmm17
+        vmovups   832(%rsp), %zmm18
+        vmovups   768(%rsp), %zmm19
+        vmovups   704(%rsp), %zmm20
+        vmovups   640(%rsp), %zmm21
+        vmovups   576(%rsp), %zmm22
+        vmovups   512(%rsp), %zmm23
+        vmovups   448(%rsp), %zmm24
+        vmovups   384(%rsp), %zmm25
+        vmovups   320(%rsp), %zmm26
+        vmovups   256(%rsp), %zmm27
+        vmovups   192(%rsp), %zmm28
+        vmovups   128(%rsp), %zmm29
+        vmovups   64(%rsp), %zmm30
+        vmovups   (%rsp), %zmm31
+        movq      1080(%rsp), %r14
+        cfi_restore (%r14)
+        movq      1072(%rsp), %r15
+        cfi_restore (%r15)
+        movq      1064(%rsp), %rbx
+        vmovups   1216(%rsp), %zmm0
+        vmovups   1280(%rsp), %zmm3
+        jmp       .LBL_1_2
+
+.LBL_1_10:
+        cfi_restore_state
+        movzbl    %r12b, %r15d
+        vmovss    1156(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 1220(%rsp,%r15,8)
+        vmovss    1156(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 1284(%rsp,%r15,8)
+        jmp       .LBL_1_8
+
+.LBL_1_13:
+        movzbl    %r12b, %r15d
+        vmovss    1152(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 1216(%rsp,%r15,8)
+        vmovss    1152(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 1280(%rsp,%r15,8)
+        jmp       .LBL_1_7
+#endif
+END (_ZGVeN16vvv_sincosf_knl)
+
+ENTRY (_ZGVeN16vvv_sincosf_skx)
+#ifndef HAVE_AVX512_ASM_SUPPORT
+WRAPPER_IMPL_AVX512_fFF _ZGVdN8vvv_sincosf
+#else
+        pushq     %rbp
+        cfi_adjust_cfa_offset (8)
+        cfi_rel_offset (%rbp, 0)
+        movq      %rsp, %rbp
+        cfi_def_cfa_register (%rbp)
+        andq      $-64, %rsp
+        subq      $1344, %rsp
+        movq      __svml_ssincos_data@GOTPCREL(%rip), %rax
+        vmovaps   %zmm0, %zmm4
+        vmovups __sAbsMask(%rax), %zmm3
+        vmovups __sInvPI(%rax), %zmm5
+        vmovups __sRShifter(%rax), %zmm6
+        vmovups __sPI1_FMA(%rax), %zmm9
+        vmovups __sPI2_FMA(%rax), %zmm10
+        vmovups __sSignMask(%rax), %zmm14
+        vmovups __sOneHalf(%rax), %zmm7
+        vmovups __sPI3_FMA(%rax), %zmm12
+
+/* Absolute argument computation */
+        vandps    %zmm3, %zmm4, %zmm2
+
+/* c) Getting octant Y by 2/Pi multiplication
+   d) Add "Right Shifter" value */
+        vfmadd213ps %zmm6, %zmm2, %zmm5
+        vcmpps    $18, __sRangeReductionVal(%rax), %zmm2, %k1
+
+/* e) Treat obtained value as integer S for destination sign setting */
+        vpslld    $31, %zmm5, %zmm0
+
+/* g) Subtract "Right Shifter" (0x4B000000) value */
+        vsubps    %zmm6, %zmm5, %zmm5
+        vmovups __sA3(%rax), %zmm6
+
+/* h) Subtract Y*(PI/2) from X argument, where PI/2 divided to 3 parts:
+      X = X - Y*PI1 - Y*PI2 - Y*PI3 */
+        vmovaps   %zmm2, %zmm11
+        vfnmadd231ps %zmm5, %zmm9, %zmm11
+        vfnmadd231ps %zmm5, %zmm10, %zmm11
+        vandps    %zmm11, %zmm14, %zmm1
+        vxorps    %zmm1, %zmm7, %zmm8
+
+/* Result sign calculations */
+        vpternlogd $150, %zmm0, %zmm14, %zmm1
+        vmovups   .L_2il0floatpacket.13(%rip), %zmm14
+
+/* Add correction term 0.5 for cos() part */
+        vaddps    %zmm8, %zmm5, %zmm15
+        vfnmadd213ps %zmm11, %zmm12, %zmm5
+        vandnps   %zmm4, %zmm3, %zmm11
+        vmovups __sA7_FMA(%rax), %zmm3
+        vmovaps   %zmm2, %zmm13
+        vfnmadd231ps %zmm15, %zmm9, %zmm13
+        vxorps    %zmm0, %zmm5, %zmm9
+        vmovups __sA5_FMA(%rax), %zmm0
+        vfnmadd231ps %zmm15, %zmm10, %zmm13
+        vmulps    %zmm9, %zmm9, %zmm8
+        vfnmadd213ps %zmm13, %zmm12, %zmm15
+        vmovups __sA9_FMA(%rax), %zmm12
+        vxorps    %zmm1, %zmm15, %zmm1
+        vmulps    %zmm1, %zmm1, %zmm13
+
+/* 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)))) */
+        vmovaps   %zmm12, %zmm7
+        vfmadd213ps %zmm3, %zmm8, %zmm7
+        vfmadd213ps %zmm3, %zmm13, %zmm12
+        vfmadd213ps %zmm0, %zmm8, %zmm7
+        vfmadd213ps %zmm0, %zmm13, %zmm12
+        vfmadd213ps %zmm6, %zmm8, %zmm7
+        vfmadd213ps %zmm6, %zmm13, %zmm12
+        vmulps    %zmm8, %zmm7, %zmm10
+        vmulps    %zmm13, %zmm12, %zmm3
+        vfmadd213ps %zmm9, %zmm9, %zmm10
+        vfmadd213ps %zmm1, %zmm1, %zmm3
+        vxorps    %zmm11, %zmm10, %zmm0
+        vpandnd   %zmm2, %zmm2, %zmm14{%k1}
+        vptestmd  %zmm14, %zmm14, %k0
+        kmovw     %k0, %ecx
+        testl     %ecx, %ecx
+        jne       .LBL_2_3
+
+.LBL_2_2:
+        cfi_remember_state
+        vmovups   %zmm0, (%rdi)
+        vmovups   %zmm3, (%rsi)
+        movq      %rbp, %rsp
+        cfi_def_cfa_register (%rsp)
+        popq      %rbp
+        cfi_adjust_cfa_offset (-8)
+        cfi_restore (%rbp)
+        ret
+
+.LBL_2_3:
+        cfi_restore_state
+        vmovups   %zmm4, 1152(%rsp)
+        vmovups   %zmm0, 1216(%rsp)
+        vmovups   %zmm3, 1280(%rsp)
+        je        .LBL_2_2
+
+        xorb      %dl, %dl
+        xorl      %eax, %eax
+        kmovw     %k4, 1048(%rsp)
+        kmovw     %k5, 1040(%rsp)
+        kmovw     %k6, 1032(%rsp)
+        kmovw     %k7, 1024(%rsp)
+        vmovups   %zmm16, 960(%rsp)
+        vmovups   %zmm17, 896(%rsp)
+        vmovups   %zmm18, 832(%rsp)
+        vmovups   %zmm19, 768(%rsp)
+        vmovups   %zmm20, 704(%rsp)
+        vmovups   %zmm21, 640(%rsp)
+        vmovups   %zmm22, 576(%rsp)
+        vmovups   %zmm23, 512(%rsp)
+        vmovups   %zmm24, 448(%rsp)
+        vmovups   %zmm25, 384(%rsp)
+        vmovups   %zmm26, 320(%rsp)
+        vmovups   %zmm27, 256(%rsp)
+        vmovups   %zmm28, 192(%rsp)
+        vmovups   %zmm29, 128(%rsp)
+        vmovups   %zmm30, 64(%rsp)
+        vmovups   %zmm31, (%rsp)
+        movq      %rsi, 1056(%rsp)
+        movq      %r12, 1096(%rsp)
+        cfi_offset_rel_rsp (12, 1096)
+        movb      %dl, %r12b
+        movq      %r13, 1088(%rsp)
+        cfi_offset_rel_rsp (13, 1088)
+        movl      %eax, %r13d
+        movq      %r14, 1080(%rsp)
+        cfi_offset_rel_rsp (14, 1080)
+        movl      %ecx, %r14d
+        movq      %r15, 1072(%rsp)
+        cfi_offset_rel_rsp (15, 1072)
+        movq      %rbx, 1064(%rsp)
+        movq      %rdi, %rbx
+        cfi_remember_state
+
+.LBL_2_6:
+        btl       %r13d, %r14d
+        jc        .LBL_2_13
+
+.LBL_2_7:
+        lea       1(%r13), %esi
+        btl       %esi, %r14d
+        jc        .LBL_2_10
+
+.LBL_2_8:
+        incb      %r12b
+        addl      $2, %r13d
+        cmpb      $16, %r12b
+        jb        .LBL_2_6
+
+        kmovw     1048(%rsp), %k4
+        movq      %rbx, %rdi
+        kmovw     1040(%rsp), %k5
+        kmovw     1032(%rsp), %k6
+        kmovw     1024(%rsp), %k7
+        vmovups   960(%rsp), %zmm16
+        vmovups   896(%rsp), %zmm17
+        vmovups   832(%rsp), %zmm18
+        vmovups   768(%rsp), %zmm19
+        vmovups   704(%rsp), %zmm20
+        vmovups   640(%rsp), %zmm21
+        vmovups   576(%rsp), %zmm22
+        vmovups   512(%rsp), %zmm23
+        vmovups   448(%rsp), %zmm24
+        vmovups   384(%rsp), %zmm25
+        vmovups   320(%rsp), %zmm26
+        vmovups   256(%rsp), %zmm27
+        vmovups   192(%rsp), %zmm28
+        vmovups   128(%rsp), %zmm29
+        vmovups   64(%rsp), %zmm30
+        vmovups   (%rsp), %zmm31
+        vmovups   1216(%rsp), %zmm0
+        vmovups   1280(%rsp), %zmm3
+        movq      1056(%rsp), %rsi
+        movq      1096(%rsp), %r12
+        cfi_restore (%r12)
+        movq      1088(%rsp), %r13
+        cfi_restore (%r13)
+        movq      1080(%rsp), %r14
+        cfi_restore (%r14)
+        movq      1072(%rsp), %r15
+        cfi_restore (%r15)
+        movq      1064(%rsp), %rbx
+        jmp       .LBL_2_2
+
+.LBL_2_10:
+        cfi_restore_state
+        movzbl    %r12b, %r15d
+        vmovss    1156(%rsp,%r15,8), %xmm0
+        vzeroupper
+        vmovss    1156(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 1220(%rsp,%r15,8)
+        vmovss    1156(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 1284(%rsp,%r15,8)
+        jmp       .LBL_2_8
+
+.LBL_2_13:
+        movzbl    %r12b, %r15d
+        vmovss    1152(%rsp,%r15,8), %xmm0
+        vzeroupper
+        vmovss    1152(%rsp,%r15,8), %xmm0
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 1216(%rsp,%r15,8)
+        vmovss    1152(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 1280(%rsp,%r15,8)
+        jmp       .LBL_2_7
+#endif
+END (_ZGVeN16vvv_sincosf_skx)
+
+	.section .rodata, "a"
+.L_2il0floatpacket.13:
+	.long	0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff,0xffffffff
+	.type	.L_2il0floatpacket.13,@object
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core.S
new file mode 100644
index 0000000000..610046b587
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf4_core.S
@@ -0,0 +1,38 @@
+/* Multiple versions of vectorized sincosf.
+   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 <init-arch.h>
+
+	.text
+ENTRY (_ZGVbN4vvv_sincosf)
+        .type   _ZGVbN4vvv_sincosf, @gnu_indirect_function
+        cmpl    $0, KIND_OFFSET+__cpu_features(%rip)
+        jne     1f
+        call    __init_cpu_features
+1:      leaq    _ZGVbN4vvv_sincosf_sse4(%rip), %rax
+        testl   $bit_SSE4_1, __cpu_features+CPUID_OFFSET+index_SSE4_1(%rip)
+        jz      2f
+        ret
+2:      leaq    _ZGVbN4vvv_sincosf_sse2(%rip), %rax
+        ret
+END (_ZGVbN4vvv_sincosf)
+libmvec_hidden_def (_ZGVbN4vvv_sincosf)
+
+#define _ZGVbN4vvv_sincosf _ZGVbN4vvv_sincosf_sse2
+#include "../svml_s_sincosf4_core.S"
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)
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core.S
new file mode 100644
index 0000000000..9e5be67fc9
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core.S
@@ -0,0 +1,38 @@
+/* Multiple versions of vectorized sincosf.
+   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 <init-arch.h>
+
+	.text
+ENTRY (_ZGVdN8vvv_sincosf)
+        .type   _ZGVdN8vvv_sincosf, @gnu_indirect_function
+        cmpl    $0, KIND_OFFSET+__cpu_features(%rip)
+        jne     1f
+        call    __init_cpu_features
+1:      leaq    _ZGVdN8vvv_sincosf_avx2(%rip), %rax
+        testl   $bit_AVX2_Usable, __cpu_features+FEATURE_OFFSET+index_AVX2_Usable(%rip)
+        jz      2f
+        ret
+2:      leaq    _ZGVdN8vvv_sincosf_sse_wrapper(%rip), %rax
+        ret
+END (_ZGVdN8vvv_sincosf)
+libmvec_hidden_def (_ZGVdN8vvv_sincosf)
+
+#define _ZGVdN8vvv_sincosf _ZGVdN8vvv_sincosf_sse_wrapper
+#include "../svml_s_sincosf8_core.S"
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core_avx2.S
new file mode 100644
index 0000000000..153c315799
--- /dev/null
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf8_core_avx2.S
@@ -0,0 +1,241 @@
+/* Function sincosf vectorized with AVX2.
+   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(_ZGVdN8vvv_sincosf_avx2)
+/*
+   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      $448, %rsp
+        movq      __svml_ssincos_data@GOTPCREL(%rip), %rax
+        vmovdqa   %ymm0, %ymm5
+        vmovups   %ymm13, 352(%rsp)
+        vmovups __sAbsMask(%rax), %ymm2
+        vmovups __sInvPI(%rax), %ymm1
+        vmovups __sPI1_FMA(%rax), %ymm13
+        vmovups   %ymm15, 288(%rsp)
+
+/* Absolute argument computation */
+        vandps    %ymm2, %ymm5, %ymm4
+
+/* c) Getting octant Y by 2/Pi multiplication
+   d) Add "Right Shifter" value */
+        vfmadd213ps __sRShifter(%rax), %ymm4, %ymm1
+
+/* e) Treat obtained value as integer S for destination sign setting */
+        vpslld    $31, %ymm1, %ymm0
+
+/* g) Subtract "Right Shifter" (0x4B000000) value */
+        vsubps __sRShifter(%rax), %ymm1, %ymm1
+
+/* h) Subtract Y*(PI/2) from X argument, where PI/2 divided to 3 parts:
+      X = X - Y*PI1 - Y*PI2 - Y*PI3 */
+        vmovdqa   %ymm4, %ymm7
+        vfnmadd231ps %ymm1, %ymm13, %ymm7
+        vfnmadd231ps __sPI2_FMA(%rax), %ymm1, %ymm7
+        vandps __sSignMask(%rax), %ymm7, %ymm15
+        vxorps __sOneHalf(%rax), %ymm15, %ymm6
+
+/* Add correction term 0.5 for cos() part */
+        vaddps    %ymm6, %ymm1, %ymm6
+        vmovdqa   %ymm4, %ymm3
+        vfnmadd231ps %ymm6, %ymm13, %ymm3
+        vmovups __sPI3_FMA(%rax), %ymm13
+        vcmpnle_uqps __sRangeReductionVal(%rax), %ymm4, %ymm4
+        vfnmadd231ps __sPI2_FMA(%rax), %ymm6, %ymm3
+        vfnmadd213ps %ymm7, %ymm13, %ymm1
+        vfnmadd213ps %ymm3, %ymm13, %ymm6
+
+/* Result sign calculations */
+        vxorps __sSignMask(%rax), %ymm15, %ymm3
+        vxorps    %ymm0, %ymm3, %ymm7
+        vxorps    %ymm7, %ymm6, %ymm3
+        vxorps    %ymm0, %ymm1, %ymm15
+        vandnps   %ymm5, %ymm2, %ymm6
+        vmovups __sA7_FMA(%rax), %ymm2
+        vmulps    %ymm15, %ymm15, %ymm13
+        vmovups __sA9_FMA(%rax), %ymm7
+        vmulps    %ymm3, %ymm3, %ymm1
+
+/* 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)))) */
+        vmovdqa   %ymm2, %ymm0
+        vfmadd231ps __sA9_FMA(%rax), %ymm13, %ymm0
+        vfmadd213ps %ymm2, %ymm1, %ymm7
+        vfmadd213ps __sA5_FMA(%rax), %ymm13, %ymm0
+        vfmadd213ps __sA5_FMA(%rax), %ymm1, %ymm7
+        vfmadd213ps __sA3(%rax), %ymm13, %ymm0
+        vfmadd213ps __sA3(%rax), %ymm1, %ymm7
+        vmulps    %ymm13, %ymm0, %ymm13
+        vmulps    %ymm1, %ymm7, %ymm1
+        vfmadd213ps %ymm15, %ymm15, %ymm13
+        vfmadd213ps %ymm3, %ymm3, %ymm1
+        vmovmskps %ymm4, %ecx
+        vxorps    %ymm6, %ymm13, %ymm0
+        testl     %ecx, %ecx
+        jne       .LBL_1_3
+
+.LBL_1_2:
+        cfi_remember_state
+        vmovups   352(%rsp), %ymm13
+        vmovups   288(%rsp), %ymm15
+        vmovups   %ymm0, (%rdi)
+        vmovups   %ymm1, (%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
+        vmovups   %ymm5, 256(%rsp)
+        vmovups   %ymm0, 320(%rsp)
+        vmovups   %ymm1, 384(%rsp)
+        je        .LBL_1_2
+
+        xorb      %dl, %dl
+        xorl      %eax, %eax
+        vmovups   %ymm8, 160(%rsp)
+        vmovups   %ymm9, 128(%rsp)
+        vmovups   %ymm10, 96(%rsp)
+        vmovups   %ymm11, 64(%rsp)
+        vmovups   %ymm12, 32(%rsp)
+        vmovups   %ymm14, (%rsp)
+        movq      %rsi, 192(%rsp)
+        movq      %r12, 232(%rsp)
+        cfi_offset_rel_rsp (12, 232)
+        movb      %dl, %r12b
+        movq      %r13, 224(%rsp)
+        cfi_offset_rel_rsp (13, 224)
+        movl      %eax, %r13d
+        movq      %r14, 216(%rsp)
+        cfi_offset_rel_rsp (14, 216)
+        movl      %ecx, %r14d
+        movq      %r15, 208(%rsp)
+        cfi_offset_rel_rsp (14, 208)
+        movq      %rbx, 200(%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
+
+        vmovups   160(%rsp), %ymm8
+        movq      %rbx, %rdi
+        vmovups   128(%rsp), %ymm9
+        vmovups   96(%rsp), %ymm10
+        vmovups   64(%rsp), %ymm11
+        vmovups   32(%rsp), %ymm12
+        vmovups   (%rsp), %ymm14
+        vmovups   320(%rsp), %ymm0
+        vmovups   384(%rsp), %ymm1
+        movq      192(%rsp), %rsi
+        movq      232(%rsp), %r12
+        cfi_restore (%r12)
+        movq      224(%rsp), %r13
+        cfi_restore (%r13)
+        movq      216(%rsp), %r14
+        cfi_restore (%r14)
+        movq      208(%rsp), %r15
+        cfi_restore (%r15)
+        movq      200(%rsp), %rbx
+        jmp       .LBL_1_2
+
+.LBL_1_10:
+        cfi_restore_state
+        movzbl    %r12b, %r15d
+        vmovss    260(%rsp,%r15,8), %xmm0
+        vzeroupper
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 324(%rsp,%r15,8)
+        vmovss    260(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 388(%rsp,%r15,8)
+        jmp       .LBL_1_8
+
+.LBL_1_13:
+        movzbl    %r12b, %r15d
+        vmovss    256(%rsp,%r15,8), %xmm0
+        vzeroupper
+
+        call      sinf@PLT
+
+        vmovss    %xmm0, 320(%rsp,%r15,8)
+        vmovss    256(%rsp,%r15,8), %xmm0
+
+        call      cosf@PLT
+
+        vmovss    %xmm0, 384(%rsp,%r15,8)
+        jmp       .LBL_1_7
+
+END(_ZGVdN8vvv_sincosf_avx2)