x86_64: Fix svml_s_tanf8_core_avx2.S code formatting

This commit contains following formatting changes

1. Instructions proceeded by a tab.
2. Instruction less than 8 characters in length have a tab
   between it and the first operand.
3. Instruction greater than 7 characters in length have a
   space between it and the first operand.
4. Tabs after `#define`d names and their value.
5. 8 space at the beginning of line replaced by tab.
6. Indent comments with code.
7. Remove redundent .text section.
8. 1 space between line content and line comment.
9. Space after all commas.

Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>

1 files changed, 2467 insertions, 2470 deletions

diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_tanf8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_tanf8_core_avx2.S
index 7c67acf8c8..cd33fac643 100644
--- a/sysdeps/x86_64/fpu/multiarch/svml_s_tanf8_core_avx2.S
+++ b/sysdeps/x86_64/fpu/multiarch/svml_s_tanf8_core_avx2.S
@@ -47,2549 +47,2546 @@
 
 /* Offsets for data table __svml_stan_data_internal
  */
-#define _sInvPI_uisa                  	0
-#define _sPI1_uisa                    	32
-#define _sPI2_uisa                    	64
-#define _sPI3_uisa                    	96
-#define _sPI2_ha_uisa                 	128
-#define _sPI3_ha_uisa                 	160
-#define Th_tbl_uisa                   	192
-#define Tl_tbl_uisa                   	320
-#define _sPC3_uisa                    	448
-#define _sPC5_uisa                    	480
-#define _sRangeReductionVal_uisa      	512
-#define _sInvPi                       	544
-#define _sSignMask                    	576
-#define _sAbsMask                     	608
-#define _sRangeVal                    	640
-#define _sRShifter                    	672
-#define _sOne                         	704
-#define _sRangeReductionVal           	736
-#define _sPI1                         	768
-#define _sPI2                         	800
-#define _sPI3                         	832
-#define _sPI4                         	864
-#define _sPI1_FMA                     	896
-#define _sPI2_FMA                     	928
-#define _sPI3_FMA                     	960
-#define _sP0                          	992
-#define _sP1                          	1024
-#define _sQ0                          	1056
-#define _sQ1                          	1088
-#define _sQ2                          	1120
-#define _sTwo                         	1152
-#define _sCoeffs                      	1184
+#define _sInvPI_uisa			0
+#define _sPI1_uisa			32
+#define _sPI2_uisa			64
+#define _sPI3_uisa			96
+#define _sPI2_ha_uisa			128
+#define _sPI3_ha_uisa			160
+#define Th_tbl_uisa			192
+#define Tl_tbl_uisa			320
+#define _sPC3_uisa			448
+#define _sPC5_uisa			480
+#define _sRangeReductionVal_uisa	512
+#define _sInvPi				544
+#define _sSignMask			576
+#define _sAbsMask			608
+#define _sRangeVal			640
+#define _sRShifter			672
+#define _sOne				704
+#define _sRangeReductionVal		736
+#define _sPI1				768
+#define _sPI2				800
+#define _sPI3				832
+#define _sPI4				864
+#define _sPI1_FMA			896
+#define _sPI2_FMA			928
+#define _sPI3_FMA			960
+#define _sP0				992
+#define _sP1				1024
+#define _sQ0				1056
+#define _sQ1				1088
+#define _sQ2				1120
+#define _sTwo				1152
+#define _sCoeffs			1184
 
 #include <sysdep.h>
 
-        .text
-	.section .text.avx2,"ax",@progbits
+	.section .text.avx2, "ax", @progbits
 ENTRY(_ZGVdN8v_tanf_avx2)
-        pushq     %rbp
-        cfi_def_cfa_offset(16)
-        movq      %rsp, %rbp
-        cfi_def_cfa(6, 16)
-        cfi_offset(6, -16)
-        andq      $-32, %rsp
-        pushq     %rbx
-        subq      $184, %rsp
-
-/*
- * Legacy Code
- * Here HW FMA can be unavailable
- */
-        xorl      %eax, %eax
-        vmovups   _sAbsMask+__svml_stan_data_internal(%rip), %ymm10
-
-/*
- *
- * Main path (_LA_ and _EP_)
- *
- * Octant calculation
- */
-        vmovups   _sInvPi+__svml_stan_data_internal(%rip), %ymm5
-        vmovups   _sRShifter+__svml_stan_data_internal(%rip), %ymm2
-
-/* Range reduction */
-        vmovups   _sPI1_FMA+__svml_stan_data_internal(%rip), %ymm3
-
-/* Rational approximation */
-        vmovups   _sP1+__svml_stan_data_internal(%rip), %ymm9
-        vmovaps   %ymm0, %ymm12
-        vandps    %ymm10, %ymm12, %ymm1
-        vfmadd213ps %ymm2, %ymm1, %ymm5
-        vsubps    %ymm2, %ymm5, %ymm8
-        vpslld    $30, %ymm5, %ymm6
-
-/* Inversion mask and sign calculation */
-        vpslld    $31, %ymm5, %ymm4
-        vfnmadd213ps %ymm1, %ymm8, %ymm3
-        vfnmadd231ps _sPI2_FMA+__svml_stan_data_internal(%rip), %ymm8, %ymm3
-        vfnmadd132ps _sPI3_FMA+__svml_stan_data_internal(%rip), %ymm3, %ymm8
-        vmovups   _sQ2+__svml_stan_data_internal(%rip), %ymm3
-        vmulps    %ymm8, %ymm8, %ymm13
-        vfmadd213ps _sQ1+__svml_stan_data_internal(%rip), %ymm13, %ymm3
-        vfmadd213ps _sP0+__svml_stan_data_internal(%rip), %ymm13, %ymm9
-        vfmadd213ps _sQ0+__svml_stan_data_internal(%rip), %ymm13, %ymm3
-        vmulps    %ymm9, %ymm8, %ymm8
-        vxorps    %ymm7, %ymm7, %ymm7
-        vcmpneqps %ymm7, %ymm6, %ymm2
-        vandnps   %ymm12, %ymm10, %ymm11
-        vxorps    %ymm11, %ymm4, %ymm0
-
-/* Exchanged numerator and denominator if necessary */
-        vandnps   %ymm8, %ymm2, %ymm14
-        vandps    %ymm3, %ymm2, %ymm15
-        vandps    %ymm8, %ymm2, %ymm4
-        vandnps   %ymm3, %ymm2, %ymm5
-        vorps     %ymm15, %ymm14, %ymm6
-        vorps     %ymm5, %ymm4, %ymm7
-
-/* Division */
-        vdivps    %ymm7, %ymm6, %ymm9
-
-/* Large values check */
-        vcmpnle_uqps _sRangeReductionVal+__svml_stan_data_internal(%rip), %ymm1, %ymm10
-        vmovmskps %ymm10, %edx
-
-/* Sign setting */
-        vxorps    %ymm0, %ymm9, %ymm0
-
-/*
- *
- * End of main path (_LA_ and _EP_)
- */
-
-        testl     %edx, %edx
-
-/* Go to auxilary branch */
-        jne       L(AUX_BRANCH)
-        /*  DW_CFA_expression: r3 (rbx) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -8; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x03, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xf8, 0xff, 0xff, 0xff, 0x22
-                                # LOE r12 r13 r14 r15 eax ymm0 ymm1 ymm10 ymm11 ymm12
-
-/* Return from auxilary branch
- * for out of main path inputs
- */
+	pushq	%rbp
+	cfi_def_cfa_offset(16)
+	movq	%rsp, %rbp
+	cfi_def_cfa(6, 16)
+	cfi_offset(6, -16)
+	andq	$-32, %rsp
+	pushq	%rbx
+	subq	$184, %rsp
+
+	/*
+	 * Legacy Code
+	 * Here HW FMA can be unavailable
+	 */
+	xorl	%eax, %eax
+	vmovups	_sAbsMask+__svml_stan_data_internal(%rip), %ymm10
+
+	/*
+	 *
+	 * Main path (_LA_ and _EP_)
+	 *
+	 * Octant calculation
+	 */
+	vmovups	_sInvPi+__svml_stan_data_internal(%rip), %ymm5
+	vmovups	_sRShifter+__svml_stan_data_internal(%rip), %ymm2
+
+	/* Range reduction */
+	vmovups	_sPI1_FMA+__svml_stan_data_internal(%rip), %ymm3
+
+	/* Rational approximation */
+	vmovups	_sP1+__svml_stan_data_internal(%rip), %ymm9
+	vmovaps	%ymm0, %ymm12
+	vandps	%ymm10, %ymm12, %ymm1
+	vfmadd213ps %ymm2, %ymm1, %ymm5
+	vsubps	%ymm2, %ymm5, %ymm8
+	vpslld	$30, %ymm5, %ymm6
+
+	/* Inversion mask and sign calculation */
+	vpslld	$31, %ymm5, %ymm4
+	vfnmadd213ps %ymm1, %ymm8, %ymm3
+	vfnmadd231ps _sPI2_FMA+__svml_stan_data_internal(%rip), %ymm8, %ymm3
+	vfnmadd132ps _sPI3_FMA+__svml_stan_data_internal(%rip), %ymm3, %ymm8
+	vmovups	_sQ2+__svml_stan_data_internal(%rip), %ymm3
+	vmulps	%ymm8, %ymm8, %ymm13
+	vfmadd213ps _sQ1+__svml_stan_data_internal(%rip), %ymm13, %ymm3
+	vfmadd213ps _sP0+__svml_stan_data_internal(%rip), %ymm13, %ymm9
+	vfmadd213ps _sQ0+__svml_stan_data_internal(%rip), %ymm13, %ymm3
+	vmulps	%ymm9, %ymm8, %ymm8
+	vxorps	%ymm7, %ymm7, %ymm7
+	vcmpneqps %ymm7, %ymm6, %ymm2
+	vandnps	%ymm12, %ymm10, %ymm11
+	vxorps	%ymm11, %ymm4, %ymm0
+
+	/* Exchanged numerator and denominator if necessary */
+	vandnps	%ymm8, %ymm2, %ymm14
+	vandps	%ymm3, %ymm2, %ymm15
+	vandps	%ymm8, %ymm2, %ymm4
+	vandnps	%ymm3, %ymm2, %ymm5
+	vorps	%ymm15, %ymm14, %ymm6
+	vorps	%ymm5, %ymm4, %ymm7
+
+	/* Division */
+	vdivps	%ymm7, %ymm6, %ymm9
+
+	/* Large values check */
+	vcmpnle_uqps _sRangeReductionVal+__svml_stan_data_internal(%rip), %ymm1, %ymm10
+	vmovmskps %ymm10, %edx
+
+	/* Sign setting */
+	vxorps	%ymm0, %ymm9, %ymm0
+
+	/*
+	 *
+	 * End of main path (_LA_ and _EP_)
+	 */
+
+	testl	%edx, %edx
+
+	/* Go to auxilary branch */
+	jne	L(AUX_BRANCH)
+	/*  DW_CFA_expression: r3 (rbx) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -8; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x03, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xf8, 0xff, 0xff, 0xff, 0x22
+	# LOE r12 r13 r14 r15 eax ymm0 ymm1 ymm10 ymm11 ymm12
+
+	/* Return from auxilary branch
+	 * for out of main path inputs
+	 */
 
 L(AUX_BRANCH_RETURN):
-        testl     %eax, %eax
+	testl	%eax, %eax
 
-/* Go to special inputs processing branch */
-        jne       L(SPECIAL_VALUES_BRANCH)
-                                # LOE r12 r13 r14 r15 eax ymm0 ymm12
+	/* Go to special inputs processing branch */
+	jne	L(SPECIAL_VALUES_BRANCH)
+	# LOE r12 r13 r14 r15 eax ymm0 ymm12
 
-/* Restore registers
- * and exit the function
- */
+	/* Restore registers
+	 * and exit the function
+	 */
 
 L(EXIT):
-        addq      $184, %rsp
-        cfi_restore(3)
-        popq      %rbx
-        movq      %rbp, %rsp
-        popq      %rbp
-        cfi_def_cfa(7, 8)
-        cfi_restore(6)
-        ret
-        cfi_def_cfa(6, 16)
-        /*  DW_CFA_expression: r3 (rbx) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -8; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x03, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xf8, 0xff, 0xff, 0xff, 0x22
-        cfi_offset(6, -16)
-
-/* Branch to process
- * special inputs
- */
+	addq	$184, %rsp
+	cfi_restore(3)
+	popq	%rbx
+	movq	%rbp, %rsp
+	popq	%rbp
+	cfi_def_cfa(7, 8)
+	cfi_restore(6)
+	ret
+	cfi_def_cfa(6, 16)
+	/*  DW_CFA_expression: r3 (rbx) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -8; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x03, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xf8, 0xff, 0xff, 0xff, 0x22
+	cfi_offset(6, -16)
+
+	/* Branch to process
+	 * special inputs
+	 */
 
 L(SPECIAL_VALUES_BRANCH):
-        vmovups   %ymm12, 32(%rsp)
-        vmovups   %ymm0, 64(%rsp)
-                                # LOE r12 r13 r14 r15 eax ymm0
-
-        xorl      %ebx, %ebx
-                                # LOE r12 r13 r14 r15 eax ebx
-
-        vzeroupper
-        movq      %r12, 8(%rsp)
-        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
-        movl      %eax, %r12d
-        movq      %r13, (%rsp)
-        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
-                                # LOE r14 r15 ebx r12d
-
-/* Range mask
- * bits check
- */
+	vmovups	%ymm12, 32(%rsp)
+	vmovups	%ymm0, 64(%rsp)
+	# LOE r12 r13 r14 r15 eax ymm0
+
+	xorl	%ebx, %ebx
+	# LOE r12 r13 r14 r15 eax ebx
+
+	vzeroupper
+	movq	%r12, 8(%rsp)
+	/*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+	movl	%eax, %r12d
+	movq	%r13, (%rsp)
+	/*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+	# LOE r14 r15 ebx r12d
+
+	/* Range mask
+	 * bits check
+	 */
 
 L(RANGEMASK_CHECK):
-        btl       %ebx, %r12d
+	btl	%ebx, %r12d
 
-/* Call scalar math function */
-        jc        L(SCALAR_MATH_CALL)
-                                # LOE r14 r15 ebx r12d
+	/* Call scalar math function */
+	jc	L(SCALAR_MATH_CALL)
+	# LOE r14 r15 ebx r12d
 
-/* Special inputs
- * processing loop
- */
+	/* Special inputs
+	 * processing loop
+	 */
 
 L(SPECIAL_VALUES_LOOP):
-        incl      %ebx
-        cmpl      $8, %ebx
-
-/* Check bits in range mask */
-        jl        L(RANGEMASK_CHECK)
-                                # LOE r14 r15 ebx r12d
-
-        movq      8(%rsp), %r12
-        cfi_restore(12)
-        movq      (%rsp), %r13
-        cfi_restore(13)
-        vmovups   64(%rsp), %ymm0
-
-/* Go to exit */
-        jmp       L(EXIT)
-        /*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
-        /*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
-        .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
-                                # LOE r12 r13 r14 r15 ymm0
-
-/* Scalar math fucntion call
- * to process special input
- */
+	incl	%ebx
+	cmpl	$8, %ebx
+
+	/* Check bits in range mask */
+	jl	L(RANGEMASK_CHECK)
+	# LOE r14 r15 ebx r12d
+
+	movq	8(%rsp), %r12
+	cfi_restore(12)
+	movq	(%rsp), %r13
+	cfi_restore(13)
+	vmovups	64(%rsp), %ymm0
+
+	/* Go to exit */
+	jmp	L(EXIT)
+	/*  DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -184; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 0xff, 0xff, 0xff, 0x22
+	/*  DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -192; DW_OP_plus)  */
+	.cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22
+	# LOE r12 r13 r14 r15 ymm0
+
+	/* Scalar math fucntion call
+	 * to process special input
+	 */
 
 L(SCALAR_MATH_CALL):
-        movl      %ebx, %r13d
-        movss     32(%rsp,%r13,4), %xmm0
-        call      tanf@PLT
-                                # LOE r13 r14 r15 ebx r12d xmm0
+	movl	%ebx, %r13d
+	movss	32(%rsp, %r13, 4), %xmm0
+	call	tanf@PLT
+	# LOE r13 r14 r15 ebx r12d xmm0
 
-        movss     %xmm0, 64(%rsp,%r13,4)
+	movss	%xmm0, 64(%rsp, %r13, 4)
 
-/* Process special inputs in loop */
-        jmp       L(SPECIAL_VALUES_LOOP)
-        cfi_restore(12)
-        cfi_restore(13)
-                                # LOE r14 r15 ebx r12d
+	/* Process special inputs in loop */
+	jmp	L(SPECIAL_VALUES_LOOP)
+	cfi_restore(12)
+	cfi_restore(13)
+	# LOE r14 r15 ebx r12d
 
-/* Auxilary branch
- * for out of main path inputs
- */
+	/* Auxilary branch
+	 * for out of main path inputs
+	 */
 
 L(AUX_BRANCH):
-        vpand     .FLT_16(%rip), %ymm1, %ymm5
-
-/*
- * Get the (2^a / 2pi) mod 1 values from the table.
- * Because doesn't have I-type gather, we need a trivial cast
- */
-        lea       __svml_stan_reduction_data_internal(%rip), %rdx
-        vmovups   %ymm11, 64(%rsp)
-        vmovups   .FLT_15(%rip), %ymm7
-        vmovups   %ymm10, 96(%rsp)
-        vmovups   %ymm0, 128(%rsp)
-        vpsrld    $23, %ymm5, %ymm6
-        vpslld    $1, %ymm6, %ymm11
-        vpaddd    %ymm6, %ymm11, %ymm13
-        vpslld    $2, %ymm13, %ymm15
-        vandps    %ymm7, %ymm12, %ymm14
-        vcmpeqps  %ymm7, %ymm14, %ymm10
-        vmovmskps %ymm10, %eax
-        vextractf128 $1, %ymm15, %xmm7
-        vmovd     %xmm15, %ecx
-        vmovd     %xmm7, %r8d
-        vmovd     (%rcx,%rdx), %xmm8
-        vpextrd   $1, %xmm15, %ebx
-        vpextrd   $2, %xmm15, %esi
-        vpextrd   $3, %xmm15, %edi
-        vpextrd   $1, %xmm7, %r10d
-        vpextrd   $2, %xmm7, %r9d
-        vpextrd   $3, %xmm7, %r11d
-        vmovd     (%rbx,%rdx), %xmm3
-        vmovd     (%rsi,%rdx), %xmm2
-        vmovd     (%rdi,%rdx), %xmm14
-        vmovd     (%r8,%rdx), %xmm10
-        vmovd     (%r10,%rdx), %xmm5
-        vmovd     (%r9,%rdx), %xmm11
-        vmovd     (%r11,%rdx), %xmm6
-        vpunpckldq %xmm3, %xmm8, %xmm4
-        vpunpckldq %xmm14, %xmm2, %xmm0
-        vpunpckldq %xmm5, %xmm10, %xmm13
-        vpunpckldq %xmm6, %xmm11, %xmm15
-        vpunpcklqdq %xmm0, %xmm4, %xmm9
-        vmovd     4(%rcx,%rdx), %xmm3
-        vmovd     4(%rbx,%rdx), %xmm2
-        vmovd     4(%rsi,%rdx), %xmm14
-        vmovd     4(%rdi,%rdx), %xmm4
-        vpunpcklqdq %xmm15, %xmm13, %xmm8
-        vmovd     4(%r8,%rdx), %xmm5
-        vmovd     4(%r10,%rdx), %xmm6
-        vmovd     4(%r9,%rdx), %xmm13
-        vmovd     4(%r11,%rdx), %xmm15
-        vpunpckldq %xmm2, %xmm3, %xmm0
-        vpunpckldq %xmm4, %xmm14, %xmm7
-        vpunpckldq %xmm15, %xmm13, %xmm3
-        vpunpcklqdq %xmm7, %xmm0, %xmm10
-        vmovd     8(%rsi,%rdx), %xmm0
-        vmovd     8(%rdi,%rdx), %xmm7
-        vmovd     8(%rcx,%rdx), %xmm14
-        vmovd     8(%rbx,%rdx), %xmm4
-        vmovd     8(%r8,%rdx), %xmm15
-        vinsertf128 $1, %xmm8, %ymm9, %ymm11
-        vpunpckldq %xmm6, %xmm5, %xmm8
-        vpunpcklqdq %xmm3, %xmm8, %xmm2
-        vpunpckldq %xmm7, %xmm0, %xmm6
-
-/*
- * Also get the significand as an integer
- * NB: adding in the integer bit is wrong for denorms!
- * To make this work for denorms we should do something slightly different
- */
-        vpand     .FLT_17(%rip), %ymm1, %ymm7
-        vmovd     8(%r10,%rdx), %xmm8
-        vmovd     8(%r9,%rdx), %xmm3
-        vpunpckldq %xmm4, %xmm14, %xmm5
-        vpunpckldq %xmm8, %xmm15, %xmm14
-
-/*  Load constants (not all needed at once)  */
-        lea       _sCoeffs+36+__svml_stan_data_internal(%rip), %r9
-        vpunpcklqdq %xmm6, %xmm5, %xmm13
-        vpaddd    .FLT_18(%rip), %ymm7, %ymm5
-        vinsertf128 $1, %xmm2, %ymm10, %ymm9
-        vmovd     8(%r11,%rdx), %xmm2
-        vpunpckldq %xmm2, %xmm3, %xmm4
-        vpunpcklqdq %xmm4, %xmm14, %xmm0
-
-/*
- * Break the P_xxx and m into 16-bit chunks ready for
- * the long multiplication via 16x16->32 multiplications
- */
-        vmovdqu   .FLT_19(%rip), %ymm14
-        vpsrld    $16, %ymm5, %ymm10
-        vpand     %ymm14, %ymm5, %ymm5
-        vpand     %ymm14, %ymm9, %ymm3
-        vpand     %ymm14, %ymm11, %ymm7
-        vpsrld    $16, %ymm11, %ymm11
-        vpmulld   %ymm3, %ymm5, %ymm8
-        vpmulld   %ymm3, %ymm10, %ymm3
-        vinsertf128 $1, %xmm0, %ymm13, %ymm13
-        vpsrld    $16, %ymm13, %ymm6
-        vpand     %ymm14, %ymm13, %ymm15
-        vpsrld    $16, %ymm9, %ymm0
-        vpmulld   %ymm6, %ymm10, %ymm13
-        vpmulld   %ymm6, %ymm5, %ymm6
-        vpsrld    $16, %ymm6, %ymm6
-        vpmulld   %ymm15, %ymm10, %ymm4
-        vpand     %ymm14, %ymm8, %ymm15
-        vpaddd    %ymm6, %ymm13, %ymm13
-        vpsrld    $16, %ymm4, %ymm4
-        vpsrld    $16, %ymm8, %ymm8
-        vpaddd    %ymm13, %ymm15, %ymm15
-        vpmulld   %ymm0, %ymm5, %ymm2
-        vpaddd    %ymm15, %ymm4, %ymm13
-        vpand     %ymm14, %ymm2, %ymm4
-        vpaddd    %ymm8, %ymm3, %ymm15
-        vpsrld    $16, %ymm2, %ymm2
-        vpsrld    $16, %ymm13, %ymm6
-
-/* Assemble reduced argument from the pieces */
-        vpand     %ymm14, %ymm13, %ymm13
-        vpaddd    %ymm15, %ymm4, %ymm8
-        vpmulld   %ymm7, %ymm5, %ymm9
-        vpmulld   %ymm0, %ymm10, %ymm0
-        vpaddd    %ymm8, %ymm6, %ymm4
-        vpand     %ymm14, %ymm9, %ymm6
-        vpaddd    %ymm2, %ymm0, %ymm8
-        vpsrld    $16, %ymm9, %ymm3
-        vpsrld    $16, %ymm4, %ymm15
-        vpslld    $16, %ymm4, %ymm4
-        vpaddd    %ymm8, %ymm6, %ymm6
-        vpaddd    %ymm6, %ymm15, %ymm0
-        vpmulld   %ymm11, %ymm5, %ymm6
-
-/* Now do the big multiplication and carry propagation */
-        vpmulld   %ymm7, %ymm10, %ymm8
-        vpand     %ymm14, %ymm6, %ymm2
-        vpaddd    %ymm3, %ymm8, %ymm5
-        vpsrld    $16, %ymm0, %ymm15
-        vpand     %ymm14, %ymm0, %ymm0
-
-/*
- * We want to incorporate the original sign now too.
- * Do it here for convenience in getting the right N value,
- * though we could wait right to the end if we were prepared
- * to modify the sign of N later too.
- * So get the appropriate sign mask now (or sooner).
- */
-        vpand     .FLT_20(%rip), %ymm1, %ymm3
-        vpaddd    %ymm5, %ymm2, %ymm7
-        vpaddd    %ymm13, %ymm4, %ymm8
-
-/*
- * Now round at the 2^-8 bit position for reduction mod pi/2^7
- * instead of the original 2pi (but still with the same 2pi scaling).
- * Use a shifter of 2^15 + 2^14.
- * The N we get is our final version; it has an offset of
- * 2^8 because of the implicit integer bit, and anyway for negative
- * starting value it's a 2s complement thing. But we need to mask
- * off the exponent part anyway so it's fine.
- */
-        vmovups   .FLT_22(%rip), %ymm14
-        vpaddd    %ymm7, %ymm15, %ymm15
-
-/*
- * Create floating-point high part, implicitly adding integer bit 1
- * Incorporate overall sign at this stage too.
- */
-        vpxor     .FLT_21(%rip), %ymm3, %ymm11
-
-/*
- * Create floating-point low and medium parts, respectively
- * lo_17, ... lo_0, 0, ..., 0
- * hi_8, ... hi_0, lo_31, ..., lo_18
- * then subtract off the implicitly added integer bits,
- * 2^-46 and 2^-23, respectively.
- * Put the original sign into all of them at this stage.
- */
-        vpxor     .FLT_23(%rip), %ymm3, %ymm7
-        vpslld    $16, %ymm15, %ymm9
-        vpaddd    %ymm0, %ymm9, %ymm2
-        vpand     .FLT_24(%rip), %ymm8, %ymm0
-        vpsrld    $18, %ymm8, %ymm8
-        vpsrld    $9, %ymm2, %ymm10
-        vpslld    $5, %ymm0, %ymm4
-        vpor      %ymm11, %ymm10, %ymm6
-        vpxor     .FLT_25(%rip), %ymm3, %ymm11
-        vpand     .FLT_26(%rip), %ymm2, %ymm3
-        vpor      %ymm7, %ymm4, %ymm5
-
-/*
- * If the magnitude of the input is <= 2^-20, then
- * just pass through the input, since no reduction will be needed and
- * the main path will only work accurately if the reduced argument is
- * about >= 2^-40 (which it is for all large pi multiples)
- */
-        vmovups   .FLT_30(%rip), %ymm4
-        vpslld    $14, %ymm3, %ymm2
-
-/*
- * Now multiply those numbers all by 2 pi, reasonably accurately.
- * (RHi + RLo) * (pi_lead + pi_trail) ~=
- * RHi * pi_lead + (RHi * pi_trail + RLo * pi_lead)
- */
-        vmovups   .FLT_27(%rip), %ymm3
-        vaddps    %ymm14, %ymm6, %ymm13
-        vpor      %ymm8, %ymm2, %ymm9
-        vsubps    %ymm14, %ymm13, %ymm15
-
-/* Grab our final N value as an integer, appropriately masked mod 2^8 */
-        vpand     .FLT_31(%rip), %ymm13, %ymm13
-        vpor      %ymm11, %ymm9, %ymm10
-        vsubps    %ymm15, %ymm6, %ymm6
-        vsubps    %ymm7, %ymm5, %ymm15
-        vsubps    %ymm11, %ymm10, %ymm14
-
-/* Now add them up into 2 reasonably aligned pieces */
-        vaddps    %ymm14, %ymm6, %ymm2
-        vsubps    %ymm2, %ymm6, %ymm6
-        vmulps    %ymm2, %ymm3, %ymm7
-        vaddps    %ymm6, %ymm14, %ymm8
-        vaddps    %ymm8, %ymm15, %ymm8
-        vmovaps   %ymm3, %ymm15
-        vfmsub213ps %ymm7, %ymm2, %ymm15
-        vandps    .FLT_29(%rip), %ymm1, %ymm0
-        vfmadd132ps .FLT_28(%rip), %ymm15, %ymm2
-        vcmpgt_oqps %ymm4, %ymm0, %ymm9
-        vcmple_oqps %ymm4, %ymm0, %ymm5
-
-/*
- * The output is _VRES_R (high) + _VRES_E (low), and the integer part is _VRES_IND
- * Set sRp2 = _VRES_R^2 and then resume the original code.
- * Argument reduction is now finished: x = n * pi/128 + r
- * where n = iIndex and r = sR (high) + sE (low).
- * But we have n modulo 256, needed for sin/cos with period 2pi
- * but we want it modulo 128 since tan has period pi.
- */
-        vpand     .FLT_32(%rip), %ymm13, %ymm0
-        vfmadd213ps %ymm2, %ymm3, %ymm8
-        vpslld    $2, %ymm0, %ymm2
-        vandps    %ymm1, %ymm5, %ymm1
-        vandps    %ymm7, %ymm9, %ymm6
-        vorps     %ymm6, %ymm1, %ymm15
-        vpaddd    %ymm0, %ymm2, %ymm1
-        vpslld    $3, %ymm1, %ymm4
-        vand