@@ -30,446 +30,445 @@
/* Offsets for data table __svml_datan2_data_internal
*/
-#define dPI 0
-#define dPIO2 64
-#define dA19 128
-#define dA18 192
-#define dA17 256
-#define dA16 320
-#define dA15 384
-#define dA14 448
-#define dA13 512
-#define dA12 576
-#define dA11 640
-#define dA10 704
-#define dA09 768
-#define dA08 832
-#define dA07 896
-#define dA06 960
-#define dA05 1024
-#define dA04 1088
-#define dA03 1152
-#define dA02 1216
-#define dA01 1280
-#define dA00 1344
-#define dSIGN_MASK 1408
-#define iCHK_WORK_SUB 1472
-#define iCHK_WORK_CMP 1536
-#define dABS_MASK 1600
-#define dZERO 1664
+#define dPI 0
+#define dPIO2 64
+#define dA19 128
+#define dA18 192
+#define dA17 256
+#define dA16 320
+#define dA15 384
+#define dA14 448
+#define dA13 512
+#define dA12 576
+#define dA11 640
+#define dA10 704
+#define dA09 768
+#define dA08 832
+#define dA07 896
+#define dA06 960
+#define dA05 1024
+#define dA04 1088
+#define dA03 1152
+#define dA02 1216
+#define dA01 1280
+#define dA00 1344
+#define dSIGN_MASK 1408
+#define iCHK_WORK_SUB 1472
+#define iCHK_WORK_CMP 1536
+#define dABS_MASK 1600
+#define dZERO 1664
#include <sysdep.h>
- .text
- .section .text.evex512,"ax",@progbits
+ .section .text.evex512, "ax", @progbits
ENTRY(_ZGVeN8vv_atan2_skx)
- pushq %rbp
- cfi_def_cfa_offset(16)
- movq %rsp, %rbp
- cfi_def_cfa(6, 16)
- cfi_offset(6, -16)
- andq $-64, %rsp
- subq $256, %rsp
- xorl %edx, %edx
-
-/*
- * #define NO_VECTOR_ZERO_ATAN2_ARGS
- * Declarations
- * Variables
- * Constants
- * The end of declarations
- * Implementation
- * Get r0~=1/B
- * Cannot be replaced by VQRCP(D, dR0, dB);
- * Argument Absolute values
- */
- vmovups dABS_MASK+__svml_datan2_data_internal(%rip), %zmm4
-
-/* Argument signs */
- vmovups dSIGN_MASK+__svml_datan2_data_internal(%rip), %zmm6
-
-/*
- * 1) If y<x then a= y, b=x, PIO2=0
- * 2) If y>x then a=-x, b=y, PIO2=Pi/2
- */
- vmovups dPIO2+__svml_datan2_data_internal(%rip), %zmm3
- vandpd %zmm4, %zmm0, %zmm11
- vmovaps %zmm1, %zmm7
- vandpd %zmm4, %zmm7, %zmm2
- vandpd %zmm6, %zmm7, %zmm5
- vandpd %zmm6, %zmm0, %zmm4
- vorpd %zmm6, %zmm2, %zmm12
- vcmppd $17, {sae}, %zmm2, %zmm11, %k1
- vmovdqu iCHK_WORK_CMP+__svml_datan2_data_internal(%rip), %ymm6
- vmovups %zmm11, 64(%rsp)
-
-/* Check if y and x are on main path. */
- vpsrlq $32, %zmm2, %zmm9
- vblendmpd %zmm11, %zmm12, %zmm13{%k1}
- vblendmpd %zmm2, %zmm11, %zmm15{%k1}
- vpsrlq $32, %zmm11, %zmm8
- vmovdqu iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %ymm12
- vdivpd {rn-sae}, %zmm15, %zmm13, %zmm1
- vmovups %zmm15, (%rsp)
- vpmovqd %zmm9, %ymm14
- vpmovqd %zmm8, %ymm10
- vxorpd %zmm3, %zmm3, %zmm3{%k1}
- vpsubd %ymm12, %ymm14, %ymm13
- vpsubd %ymm12, %ymm10, %ymm9
-
-/* Polynomial. */
- vmulpd {rn-sae}, %zmm1, %zmm1, %zmm12
- vpcmpgtd %ymm6, %ymm13, %ymm15
- vpcmpeqd %ymm6, %ymm13, %ymm11
- vmulpd {rn-sae}, %zmm12, %zmm12, %zmm13
- vpor %ymm11, %ymm15, %ymm8
- vmovups dA19+__svml_datan2_data_internal(%rip), %zmm11
- vmovups dA15+__svml_datan2_data_internal(%rip), %zmm15
- vpcmpgtd %ymm6, %ymm9, %ymm14
- vpcmpeqd %ymm6, %ymm9, %ymm6
- vpor %ymm6, %ymm14, %ymm10
- vmulpd {rn-sae}, %zmm13, %zmm13, %zmm14
- vmovups dA18+__svml_datan2_data_internal(%rip), %zmm9
- vpor %ymm10, %ymm8, %ymm6
- vmovups dA17+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd231pd {rn-sae}, %zmm14, %zmm11, %zmm15
- vmovups dA14+__svml_datan2_data_internal(%rip), %zmm11
- vmovups dA12+__svml_datan2_data_internal(%rip), %zmm8
- vfmadd231pd {rn-sae}, %zmm14, %zmm9, %zmm11
- vmovups dA13+__svml_datan2_data_internal(%rip), %zmm9
- vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm9
- vmovups dA16+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm8
- vmovups dA11+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
- vmovups dA10+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
- vmovups dA09+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
- vmovups dA08+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8
- vmovups dA07+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
- vmovups dA06+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
- vmovups dA05+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
- vmovups dA04+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8
- vmovups dA03+__svml_datan2_data_internal(%rip), %zmm10
-
-/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */
- vmulpd {rn-sae}, %zmm14, %zmm8, %zmm8
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
- vmovups dA02+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
- vmovups dA01+__svml_datan2_data_internal(%rip), %zmm10
- vfmadd213pd {rn-sae}, %zmm11, %zmm12, %zmm15
- vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
- vfmadd213pd {rn-sae}, %zmm8, %zmm12, %zmm9
- vmovups __svml_datan2_data_internal(%rip), %zmm8
- vfmadd213pd {rn-sae}, %zmm9, %zmm13, %zmm15
-
-/*
- * Reconstruction.
- * dP=(R+R*dP) + dPIO2
- */
- vfmadd213pd {rn-sae}, %zmm1, %zmm1, %zmm15
- vaddpd {rn-sae}, %zmm3, %zmm15, %zmm1
- vorpd %zmm5, %zmm1, %zmm9
-
-/* if x<0, dPI = Pi, else dPI =0 */
- vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm1
- vcmppd $18, {sae}, %zmm1, %zmm7, %k2
- vaddpd {rn-sae}, %zmm8, %zmm9, %zmm9{%k2}
- vmovmskps %ymm6, %eax
- vorpd %zmm4, %zmm9, %zmm11
-
-/* Special branch for fast (vector) processing of zero arguments */
- vmovups 64(%rsp), %zmm9
- testl %eax, %eax
-
-/* Go to auxilary branch */
- jne L(AUX_BRANCH)
- # LOE rbx r12 r13 r14 r15 edx ymm6 zmm0 zmm2 zmm3 zmm4 zmm5 zmm7 zmm9 zmm11
-
-/* 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 $-64, %rsp
+ subq $256, %rsp
+ xorl %edx, %edx
+
+ /*
+ * #define NO_VECTOR_ZERO_ATAN2_ARGS
+ * Declarations
+ * Variables
+ * Constants
+ * The end of declarations
+ * Implementation
+ * Get r0~=1/B
+ * Cannot be replaced by VQRCP(D, dR0, dB);
+ * Argument Absolute values
+ */
+ vmovups dABS_MASK+__svml_datan2_data_internal(%rip), %zmm4
+
+ /* Argument signs */
+ vmovups dSIGN_MASK+__svml_datan2_data_internal(%rip), %zmm6
+
+ /*
+ * 1) If y<x then a= y, b=x, PIO2=0
+ * 2) If y>x then a=-x, b=y, PIO2=Pi/2
+ */
+ vmovups dPIO2+__svml_datan2_data_internal(%rip), %zmm3
+ vandpd %zmm4, %zmm0, %zmm11
+ vmovaps %zmm1, %zmm7
+ vandpd %zmm4, %zmm7, %zmm2
+ vandpd %zmm6, %zmm7, %zmm5
+ vandpd %zmm6, %zmm0, %zmm4
+ vorpd %zmm6, %zmm2, %zmm12
+ vcmppd $17, {sae}, %zmm2, %zmm11, %k1
+ vmovdqu iCHK_WORK_CMP+__svml_datan2_data_internal(%rip), %ymm6
+ vmovups %zmm11, 64(%rsp)
+
+ /* Check if y and x are on main path. */
+ vpsrlq $32, %zmm2, %zmm9
+ vblendmpd %zmm11, %zmm12, %zmm13{%k1}
+ vblendmpd %zmm2, %zmm11, %zmm15{%k1}
+ vpsrlq $32, %zmm11, %zmm8
+ vmovdqu iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %ymm12
+ vdivpd {rn-sae}, %zmm15, %zmm13, %zmm1
+ vmovups %zmm15, (%rsp)
+ vpmovqd %zmm9, %ymm14
+ vpmovqd %zmm8, %ymm10
+ vxorpd %zmm3, %zmm3, %zmm3{%k1}
+ vpsubd %ymm12, %ymm14, %ymm13
+ vpsubd %ymm12, %ymm10, %ymm9
+
+ /* Polynomial. */
+ vmulpd {rn-sae}, %zmm1, %zmm1, %zmm12
+ vpcmpgtd %ymm6, %ymm13, %ymm15
+ vpcmpeqd %ymm6, %ymm13, %ymm11
+ vmulpd {rn-sae}, %zmm12, %zmm12, %zmm13
+ vpor %ymm11, %ymm15, %ymm8
+ vmovups dA19+__svml_datan2_data_internal(%rip), %zmm11
+ vmovups dA15+__svml_datan2_data_internal(%rip), %zmm15
+ vpcmpgtd %ymm6, %ymm9, %ymm14
+ vpcmpeqd %ymm6, %ymm9, %ymm6
+ vpor %ymm6, %ymm14, %ymm10
+ vmulpd {rn-sae}, %zmm13, %zmm13, %zmm14
+ vmovups dA18+__svml_datan2_data_internal(%rip), %zmm9
+ vpor %ymm10, %ymm8, %ymm6
+ vmovups dA17+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd231pd {rn-sae}, %zmm14, %zmm11, %zmm15
+ vmovups dA14+__svml_datan2_data_internal(%rip), %zmm11
+ vmovups dA12+__svml_datan2_data_internal(%rip), %zmm8
+ vfmadd231pd {rn-sae}, %zmm14, %zmm9, %zmm11
+ vmovups dA13+__svml_datan2_data_internal(%rip), %zmm9
+ vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm9
+ vmovups dA16+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm8
+ vmovups dA11+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
+ vmovups dA10+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
+ vmovups dA09+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
+ vmovups dA08+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8
+ vmovups dA07+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
+ vmovups dA06+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
+ vmovups dA05+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
+ vmovups dA04+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8
+ vmovups dA03+__svml_datan2_data_internal(%rip), %zmm10
+
+ /* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */
+ vmulpd {rn-sae}, %zmm14, %zmm8, %zmm8
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15
+ vmovups dA02+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11
+ vmovups dA01+__svml_datan2_data_internal(%rip), %zmm10
+ vfmadd213pd {rn-sae}, %zmm11, %zmm12, %zmm15
+ vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9
+ vfmadd213pd {rn-sae}, %zmm8, %zmm12, %zmm9
+ vmovups __svml_datan2_data_internal(%rip), %zmm8
+ vfmadd213pd {rn-sae}, %zmm9, %zmm13, %zmm15
+
+ /*
+ * Reconstruction.
+ * dP=(R+R*dP) + dPIO2
+ */
+ vfmadd213pd {rn-sae}, %zmm1, %zmm1, %zmm15
+ vaddpd {rn-sae}, %zmm3, %zmm15, %zmm1
+ vorpd %zmm5, %zmm1, %zmm9
+
+ /* if x<0, dPI = Pi, else dPI =0 */
+ vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm1
+ vcmppd $18, {sae}, %zmm1, %zmm7, %k2
+ vaddpd {rn-sae}, %zmm8, %zmm9, %zmm9{%k2}
+ vmovmskps %ymm6, %eax
+ vorpd %zmm4, %zmm9, %zmm11
+
+ /* Special branch for fast (vector) processing of zero arguments */
+ vmovups 64(%rsp), %zmm9
+ testl %eax, %eax
+
+ /* Go to auxilary branch */
+ jne L(AUX_BRANCH)
+ # LOE rbx r12 r13 r14 r15 edx ymm6 zmm0 zmm2 zmm3 zmm4 zmm5 zmm7 zmm9 zmm11
+
+ /* Return from auxilary branch
+ * for out of main path inputs
+ */
L(AUX_BRANCH_RETURN):
-/*
- * Special branch for fast (vector) processing of zero arguments
- * The end of implementation
- */
- testl %edx, %edx
+ /*
+ * Special branch for fast (vector) processing of zero arguments
+ * The end of implementation
+ */
+ testl %edx, %edx
-/* Go to special inputs processing branch */
- jne L(SPECIAL_VALUES_BRANCH)
- # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11
+ /* Go to special inputs processing branch */
+ jne L(SPECIAL_VALUES_BRANCH)
+ # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11
-/* Restore registers
- * and exit the function
- */
+ /* Restore registers
+ * and exit the function
+ */
L(EXIT):
- vmovaps %zmm11, %zmm0
- movq %rbp, %rsp
- popq %rbp
- cfi_def_cfa(7, 8)
- cfi_restore(6)
- ret
- cfi_def_cfa(6, 16)
- cfi_offset(6, -16)
-
-/* Branch to process
- * special inputs
- */
+ vmovaps %zmm11, %zmm0
+ movq %rbp, %rsp
+ popq %rbp
+ cfi_def_cfa(7, 8)
+ cfi_restore(6)
+ ret
+ cfi_def_cfa(6, 16)
+ cfi_offset(6, -16)
+
+ /* Branch to process
+ * special inputs
+ */
L(SPECIAL_VALUES_BRANCH):
- vmovups %zmm0, 64(%rsp)
- vmovups %zmm7, 128(%rsp)
- vmovups %zmm11, 192(%rsp)
- # LOE rbx r12 r13 r14 r15 edx zmm11
-
- xorl %eax, %eax
- # LOE rbx r12 r13 r14 r15 eax edx
-
- vzeroupper
- movq %r12, 16(%rsp)
- /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */
- .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22
- movl %eax, %r12d
- movq %r13, 8(%rsp)
- /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */
- .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22
- movl %edx, %r13d
- movq %r14, (%rsp)
- /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */
- .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22
- # LOE rbx r15 r12d r13d
-
-/* Range mask
- * bits check
- */
+ vmovups %zmm0, 64(%rsp)
+ vmovups %zmm7, 128(%rsp)
+ vmovups %zmm11, 192(%rsp)
+ # LOE rbx r12 r13 r14 r15 edx zmm11
+
+ xorl %eax, %eax
+ # LOE rbx r12 r13 r14 r15 eax edx
+
+ vzeroupper
+ movq %r12, 16(%rsp)
+ /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22
+ movl %eax, %r12d
+ movq %r13, 8(%rsp)
+ /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22
+ movl %edx, %r13d
+ movq %r14, (%rsp)
+ /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22
+ # LOE rbx r15 r12d r13d
+
+ /* Range mask
+ * bits check
+ */
L(RANGEMASK_CHECK):
- btl %r12d, %r13d
+ btl %r12d, %r13d
-/* Call scalar math function */
- jc L(SCALAR_MATH_CALL)
- # LOE rbx r15 r12d r13d
+ /* Call scalar math function */
+ jc L(SCALAR_MATH_CALL)
+ # LOE rbx r15 r12d r13d
-/* Special inputs
- * processing loop
- */
+ /* Special inputs
+ * processing loop
+ */
L(SPECIAL_VALUES_LOOP):
- incl %r12d
- cmpl $8, %r12d
-
-/* Check bits in range mask */
- jl L(RANGEMASK_CHECK)
- # LOE rbx r15 r12d r13d
-
- movq 16(%rsp), %r12
- cfi_restore(12)
- movq 8(%rsp), %r13
- cfi_restore(13)
- movq (%rsp), %r14
- cfi_restore(14)
- vmovups 192(%rsp), %zmm11
-
-/* Go to exit */
- jmp L(EXIT)
- /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */
- .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22
- /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */
- .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22
- /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */
- .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22
- # LOE rbx r12 r13 r14 r15 zmm11
-
-/* Scalar math fucntion call
- * to process special input
- */
+ incl %r12d
+ cmpl $8, %r12d
+
+ /* Check bits in range mask */
+ jl L(RANGEMASK_CHECK)
+ # LOE rbx r15 r12d r13d
+
+ movq 16(%rsp), %r12
+ cfi_restore(12)
+ movq 8(%rsp), %r13
+ cfi_restore(13)
+ movq (%rsp), %r14
+ cfi_restore(14)
+ vmovups 192(%rsp), %zmm11
+
+ /* Go to exit */
+ jmp L(EXIT)
+ /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22
+ /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22
+ /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */
+ .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22
+ # LOE rbx r12 r13 r14 r15 zmm11
+
+ /* Scalar math fucntion call
+ * to process special input
+ */
L(SCALAR_MATH_CALL):
- movl %r12d, %r14d
- movsd 64(%rsp,%r14,8), %xmm0
- movsd 128(%rsp,%r14,8), %xmm1
- call atan2@PLT
- # LOE rbx r14 r15 r12d r13d xmm0
-
- movsd %xmm0, 192(%rsp,%r14,8)
-
-/* Process special inputs in loop */
- jmp L(SPECIAL_VALUES_LOOP)
- cfi_restore(12)
- cfi_restore(13)
- cfi_restore(14)
- # LOE rbx r15 r12d r13d
-
-/* Auxilary branch
- * for out of main path inputs
- */
+ movl %r12d, %r14d
+ movsd 64(%rsp, %r14, 8), %xmm0
+ movsd 128(%rsp, %r14, 8), %xmm1
+ call atan2@PLT
+ # LOE rbx r14 r15 r12d r13d xmm0
-L(AUX_BRANCH):
-/* Check if at least on of Y or Y is zero: iAXAYZERO */
- vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm8
-
-/* Check if both X & Y are not NaNs: iXYnotNAN */
- vcmppd $3, {sae}, %zmm7, %zmm7, %k1
- vcmppd $3, {sae}, %zmm0, %zmm0, %k2
- vcmppd $4, {sae}, %zmm8, %zmm2, %k3
- vcmppd $4, {sae}, %zmm8, %zmm9, %k4
-
-/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */
- vpcmpgtq %zmm7, %zmm8, %k6
- vpternlogd $0xff, %zmm1, %zmm1, %zmm10
- vmovaps %zmm10, %zmm15
- vmovaps %zmm10, %zmm12
- vmovaps %zmm10, %zmm13
- vpandnq %zmm2, %zmm2, %zmm15{%k3}
- vmovaps %zmm10, %zmm2
- vpandnq %zmm7, %zmm7, %zmm12{%k1}
- vpandnq %zmm0, %zmm0, %zmm13{%k2}
- vpandnq %zmm9, %zmm9, %zmm2{%k4}
- vandpd %zmm13, %zmm12, %zmm14
- vorpd %zmm2, %zmm15, %zmm9
- vpsrlq $32, %zmm14, %zmm1
- vpsrlq $32, %zmm9, %zmm2
- vpmovqd %zmm1, %ymm1
- vpmovqd %zmm2, %ymm9
-
-/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */
- vpand %ymm1, %ymm9, %ymm2
+ movsd %xmm0, 192(%rsp, %r14, 8)
-/*
- * Path for zero arguments (at least one of both)
- * Check if both args are zeros (den. is zero)
- */
- vmovups (%rsp), %zmm1
-
-/* Exclude from previous callout mask zero (and not NaN) arguments */
- vpandn %ymm6, %ymm2, %ymm6
- vcmppd $4, {sae}, %zmm8, %zmm1, %k5
-
-/* Go to callout */
- vmovmskps %ymm6, %edx
- vpandnq %zmm1, %zmm1, %zmm10{%k5}
-
-/* Set sPIO2 to zero if den. is zero */
- vpandnq %zmm3, %zmm10, %zmm3
- vpandq %zmm10, %zmm8, %zmm1
- vporq %zmm1, %zmm3, %zmm3
- vorpd %zmm5, %zmm3, %zmm1
- vmovups __svml_datan2_data_internal(%rip), %zmm5
- vaddpd {rn-sae}, %zmm5, %zmm1, %zmm1{%k6}
- vorpd %zmm4, %zmm1, %zmm1
-
-/* Merge results from main and spec path */
- vpmovzxdq %ymm2, %zmm4
- vpsllq $32, %zmm4, %zmm2
- vpord %zmm2, %zmm4, %zmm3
- vpandnq %zmm11, %zmm3, %zmm11
- vpandq %zmm3, %zmm1, %zmm1
- vporq %zmm1, %zmm11, %zmm11
-
-/* Return to main vector processing path */
- jmp L(AUX_BRANCH_RETURN)
- # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11
+ /* Process special inputs in loop */
+ jmp L(SPECIAL_VALUES_LOOP)
+ cfi_restore(12)
+ cfi_restore(13)
+ cfi_restore(14)
+ # LOE rbx r15 r12d r13d
+
+ /* Auxilary branch
+ * for out of main path inputs
+ */
+
+L(AUX_BRANCH):
+ /* Check if at least on of Y or Y is zero: iAXAYZERO */
+ vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm8
+
+ /* Check if both X & Y are not NaNs: iXYnotNAN */
+ vcmppd $3, {sae}, %zmm7, %zmm7, %k1
+ vcmppd $3, {sae}, %zmm0, %zmm0, %k2
+ vcmppd $4, {sae}, %zmm8, %zmm2, %k3
+ vcmppd $4, {sae}, %zmm8, %zmm9, %k4
+
+ /* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */
+ vpcmpgtq %zmm7, %zmm8, %k6
+ vpternlogd $0xff, %zmm1, %zmm1, %zmm10
+ vmovaps %zmm10, %zmm15
+ vmovaps %zmm10, %zmm12
+ vmovaps %zmm10, %zmm13
+ vpandnq %zmm2, %zmm2, %zmm15{%k3}
+ vmovaps %zmm10, %zmm2
+ vpandnq %zmm7, %zmm7, %zmm12{%k1}
+ vpandnq %zmm0, %zmm0, %zmm13{%k2}
+ vpandnq %zmm9, %zmm9, %zmm2{%k4}
+ vandpd %zmm13, %zmm12, %zmm14
+ vorpd %zmm2, %zmm15, %zmm9
+ vpsrlq $32, %zmm14, %zmm1
+ vpsrlq $32, %zmm9, %zmm2
+ vpmovqd %zmm1, %ymm1
+ vpmovqd %zmm2, %ymm9
+
+ /* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */
+ vpand %ymm1, %ymm9, %ymm2
+
+ /*
+ * Path for zero arguments (at least one of both)
+ * Check if both args are zeros (den. is zero)
+ */
+ vmovups (%rsp), %zmm1
+
+ /* Exclude from previous callout mask zero (and not NaN) arguments */
+ vpandn %ymm6, %ymm2, %ymm6
+ vcmppd $4, {sae}, %zmm8, %zmm1, %k5
+
+ /* Go to callout */
+ vmovmskps %ymm6, %edx
+ vpandnq %zmm1, %zmm1, %zmm10{%k5}
+
+ /* Set sPIO2 to zero if den. is zero */
+ vpandnq %zmm3, %zmm10, %zmm3
+ vpandq %zmm10, %zmm8, %zmm1
+ vporq %zmm1, %zmm3, %zmm3
+ vorpd %zmm5, %zmm3, %zmm1
+ vmovups __svml_datan2_data_internal(%rip), %zmm5
+ vaddpd {rn-sae}, %zmm5, %zmm1, %zmm1{%k6}
+ vorpd %zmm4, %zmm1, %zmm1
+
+ /* Merge results from main and spec path */
+ vpmovzxdq %ymm2, %zmm4
+ vpsllq $32, %zmm4, %zmm2
+ vpord %zmm2, %zmm4, %zmm3
+ vpandnq %zmm11, %zmm3, %zmm11
+ vpandq %zmm3, %zmm1, %zmm1
+ vporq %zmm1, %zmm11, %zmm11
+
+ /* Return to main vector processing path */
+ jmp L(AUX_BRANCH_RETURN)
+ # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11
END(_ZGVeN8vv_atan2_skx)
- .section .rodata, "a"
- .align 64
+ .section .rodata, "a"
+ .align 64
#ifdef __svml_datan2_data_internal_typedef
typedef unsigned int VUINT32;
typedef struct {
- __declspec(align(64)) VUINT32 dPI[8][2];
- __declspec(align(64)) VUINT32 dPIO2[8][2];
- __declspec(align(64)) VUINT32 dA19[8][2];
- __declspec(align(64)) VUINT32 dA18[8][2];
- __declspec(align(64)) VUINT32 dA17[8][2];
- __declspec(align(64)) VUINT32 dA16[8][2];
- __declspec(align(64)) VUINT32 dA15[8][2];
- __declspec(align(64)) VUINT32 dA14[8][2];
- __declspec(align(64)) VUINT32 dA13[8][2];
- __declspec(align(64)) VUINT32 dA12[8][2];
- __declspec(align(64)) VUINT32 dA11[8][2];
- __declspec(align(64)) VUINT32 dA10[8][2];
- __declspec(align(64)) VUINT32 dA09[8][2];
- __declspec(align(64)) VUINT32 dA08[8][2];
- __declspec(align(64)) VUINT32 dA07[8][2];
- __declspec(align(64)) VUINT32 dA06[8][2];
- __declspec(align(64)) VUINT32 dA05[8][2];
- __declspec(align(64)) VUINT32 dA04[8][2];
- __declspec(align(64)) VUINT32 dA03[8][2];
- __declspec(align(64)) VUINT32 dA02[8][2];
- __declspec(align(64)) VUINT32 dA01[8][2];
- __declspec(align(64)) VUINT32 dA00[8][2];
- __declspec(align(64)) VUINT32 dSIGN_MASK[8][2];
- __declspec(align(64)) VUINT32 iCHK_WORK_SUB[16][1];
- __declspec(align(64)) VUINT32 iCHK_WORK_CMP[16][1];
- __declspec(align(64)) VUINT32 dABS_MASK[8][2];
- __declspec(align(64)) VUINT32 dZERO[8][2];
+ __declspec(align(64)) VUINT32 dPI[8][2];
+ __declspec(align(64)) VUINT32 dPIO2[8][2];
+ __declspec(align(64)) VUINT32 dA19[8][2];
+ __declspec(align(64)) VUINT32 dA18[8][2];
+ __declspec(align(64)) VUINT32 dA17[8][2];
+ __declspec(align(64)) VUINT32 dA16[8][2];
+ __declspec(align(64)) VUINT32 dA15[8][2];
+ __declspec(align(64)) VUINT32 dA14[8][2];
+ __declspec(align(64)) VUINT32 dA13[8][2];
+ __declspec(align(64)) VUINT32 dA12[8][2];
+ __declspec(align(64)) VUINT32 dA11[8][2];
+ __declspec(align(64)) VUINT32 dA10[8][2];
+ __declspec(align(64)) VUINT32 dA09[8][2];
+ __declspec(align(64)) VUINT32 dA08[8][2];
+ __declspec(align(64)) VUINT32 dA07[8][2];
+ __declspec(align(64)) VUINT32 dA06[8][2];
+ __declspec(align(64)) VUINT32 dA05[8][2];
+ __declspec(align(64)) VUINT32 dA04[8][2];
+ __declspec(align(64)) VUINT32 dA03[8][2];
+ __declspec(align(64)) VUINT32 dA02[8][2];
+ __declspec(align(64)) VUINT32 dA01[8][2];
+ __declspec(align(64)) VUINT32 dA00[8][2];
+ __declspec(align(64)) VUINT32 dSIGN_MASK[8][2];
+ __declspec(align(64)) VUINT32 iCHK_WORK_SUB[16][1];
+ __declspec(align(64)) VUINT32 iCHK_WORK_CMP[16][1];
+ __declspec(align(64)) VUINT32 dABS_MASK[8][2];
+ __declspec(align(64)) VUINT32 dZERO[8][2];
} __svml_datan2_data_internal;
#endif
__svml_datan2_data_internal:
- .quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 //dPI
- .align 64
- .quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 //dPIO2
- .align 64
- .quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19
- .align 64
- .quad 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209 // dA18
- .align 64
- .quad 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23 // dA17
- .align 64
- .quad 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3 // dA16
- .align 64
- .quad 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD // dA15
- .align 64
- .quad 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5 // dA14
- .align 64
- .quad 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C // dA13
- .align 64
- .quad 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F // dA12
- .align 64
- .quad 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC // dA11
- .align 64
- .quad 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B // dA10
- .align 64
- .quad 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954 // dA09
- .align 64
- .quad 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF // dA08
- .align 64
- .quad 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0 // dA07
- .align 64
- .quad 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651 // dA06
- .align 64
- .quad 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94 // dA05
- .align 64
- .quad 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5 // dA04
- .align 64
- .quad 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7 // dA03
- .align 64
- .quad 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34 // dA02
- .align 64
- .quad 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5 // dA01
- .align 64
- .quad 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000 // dA00
- .align 64
- .quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 //dSIGN_MASK
- .align 64
- .long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 //iCHK_WORK_SUB
- .align 64
- .long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 //iCHK_WORK_CMP
- .align 64
- .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff //dABS_MASK
- .align 64
- .quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 //dZERO
- .align 64
- .type __svml_datan2_data_internal,@object
- .size __svml_datan2_data_internal,.-__svml_datan2_data_internal
+ .quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 // dPI
+ .align 64
+ .quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 // dPIO2
+ .align 64
+ .quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19
+ .align 64
+ .quad 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209 // dA18
+ .align 64
+ .quad 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23 // dA17
+ .align 64
+ .quad 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3 // dA16
+ .align 64
+ .quad 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD // dA15
+ .align 64
+ .quad 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5 // dA14
+ .align 64
+ .quad 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C // dA13
+ .align 64
+ .quad 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F // dA12
+ .align 64
+ .quad 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC // dA11
+ .align 64
+ .quad 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B // dA10
+ .align 64
+ .quad 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954 // dA09
+ .align 64
+ .quad 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF // dA08
+ .align 64
+ .quad 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0 // dA07
+ .align 64
+ .quad 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651 // dA06
+ .align 64
+ .quad 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94 // dA05
+ .align 64
+ .quad 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5 // dA04
+ .align 64
+ .quad 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7 // dA03
+ .align 64
+ .quad 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34 // dA02
+ .align 64
+ .quad 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5 // dA01
+ .align 64
+ .quad 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000 // dA00
+ .align 64
+ .quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 // dSIGN_MASK
+ .align 64
+ .long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 // iCHK_WORK_SUB
+ .align 64
+ .long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 // iCHK_WORK_CMP
+ .align 64
+ .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff // dABS_MASK
+ .align 64
+ .quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 // dZERO
+ .align 64
+ .type __svml_datan2_data_internal, @object
+ .size __svml_datan2_data_internal, .-__svml_datan2_data_internal