@@ -33,357 +33,356 @@
/* Offsets for data table __svml_sacosh_data_internal
*/
-#define sOne 0
-#define sPoly 16
-#define iBrkValue 144
-#define iOffExpoMask 160
-#define sBigThreshold 176
-#define sC2 192
-#define sC3 208
-#define sHalf 224
-#define sLargestFinite 240
-#define sThirtyOne 256
-#define sTopMask8 272
-#define XScale 288
-#define sLn2 304
+#define sOne 0
+#define sPoly 16
+#define iBrkValue 144
+#define iOffExpoMask 160
+#define sBigThreshold 176
+#define sC2 192
+#define sC3 208
+#define sHalf 224
+#define sLargestFinite 240
+#define sThirtyOne 256
+#define sTopMask8 272
+#define XScale 288
+#define sLn2 304
#include <sysdep.h>
- .text
- .section .text.sse4,"ax",@progbits
+ .section .text.sse4, "ax", @progbits
ENTRY(_ZGVbN4v_acoshf_sse4)
- subq $72, %rsp
- cfi_def_cfa_offset(80)
-
-/* Compute U = X - 1 and V = X + 1, naively first. */
- movaps %xmm0, %xmm12
-
-/* Load constants, always including One = 1 */
- movups sOne+__svml_sacosh_data_internal(%rip), %xmm2
-
-/*
- * Check that 1 < X < +inf; otherwise go to the callout function.
- * We need the callout for X = 1 to avoid division by zero below.
- * This test ensures that callout handles NaN and either infinity.
- */
- movaps %xmm0, %xmm4
- movaps %xmm2, %xmm9
-
-/*
- * Compute e = -(2 * d + d^2)
- * The first FMR is exact, and the rounding error in the other is acceptable
- * since d and e are ~ 2^-8
- */
- movaps %xmm2, %xmm10
-
-/* Finally, express Y + W = U * V accurately where Y has <= 8 bits */
- movups sTopMask8+__svml_sacosh_data_internal(%rip), %xmm5
-
-/*
- * Now we feed into the log1p code, using H in place of _VARG1 and
- * also adding L into Xl.
- * compute 1+x as high, low parts
- */
- movaps %xmm2, %xmm13
- movaps %xmm5, %xmm11
- movaps %xmm2, %xmm3
-
-/*
- * Now 1 / (1 + d)
- * = 1 / (1 + (sqrt(1 - e) - 1))
- * = 1 / sqrt(1 - e)
- * = 1 + 1/2 * e + 3/8 * e^2 + 5/16 * e^3 + 35/128 * e^4 + ...
- * So compute the first three nonconstant terms of that, so that
- * we have a relative correction (1 + Corr) to apply to S etc.
- * C1 = 1/2
- * C2 = 3/8
- * C3 = 5/16
- */
- movups sC3+__svml_sacosh_data_internal(%rip), %xmm8
-
-/*
- * The following computation can go wrong for very large X, e.g.
- * the X^2 - 1 = U * V can overflow. But for large X we have
- * acosh(X) / log(2 X) - 1 =~= 1/(4 * X^2), so for X >= 2^30
- * we can just later stick X back into the log and tweak up the exponent.
- * Actually we scale X by 2^-30 and tweak the exponent up by 31,
- * to stay in the safe range for the later log computation.
- * Compute a flag now telling us when to do this.
- */
- movaps %xmm0, %xmm1
- cmpnleps sLargestFinite+__svml_sacosh_data_internal(%rip), %xmm4
- cmpltps sBigThreshold+__svml_sacosh_data_internal(%rip), %xmm1
- cmpnltps %xmm0, %xmm3
- subps %xmm2, %xmm12
- addps %xmm0, %xmm9
-
-/* For low-accuracy versions, naivety is harmless */
- mulps %xmm12, %xmm9
- orps %xmm3, %xmm4
- movmskps %xmm4, %edx
- andps %xmm9, %xmm11
- movaps %xmm1, %xmm3
-
-/*
- * Compute R = 1/sqrt(Y + W) * (1 + d)
- * Force R to <= 8 significant bits.
- * This means that R * Y and R^2 * Y are exactly representable.
- */
- rsqrtps %xmm11, %xmm7
- subps %xmm11, %xmm9
- andps %xmm5, %xmm7
- movaps %xmm2, %xmm4
-
-/*
- * Compute S = (Y/sqrt(Y + W)) * (1 + d)
- * and T = (W/sqrt(Y + W)) * (1 + d)
- * so that S + T = sqrt(Y + W) * (1 + d)
- * S is exact, and the rounding error in T is OK.
- */
- mulps %xmm7, %xmm11
- movaps %xmm7, %xmm6
- mulps %xmm7, %xmm9
- mulps %xmm11, %xmm6
- mulps %xmm9, %xmm7
-
-/*
- * For low-accuracy versions, the computation can be done
- * just as U + ((S + T) + (S + T) * Corr)
- */
- addps %xmm9, %xmm11
- subps %xmm6, %xmm10
- movaps %xmm2, %xmm9
- subps %xmm7, %xmm10
- mulps %xmm10, %xmm8
-
-/* Now multiplex to the case X = 2^-30 * input, Xl = 0 in the "big" case. */
- movups XScale+__svml_sacosh_data_internal(%rip), %xmm14
- mulps %xmm0, %xmm14
- addps sC2+__svml_sacosh_data_internal(%rip), %xmm8
- mulps %xmm10, %xmm8
- andnps %xmm14, %xmm3
-
-/*
- * Now resume the main code.
- * reduction: compute r,n
- */
- movdqu iBrkValue+__svml_sacosh_data_internal(%rip), %xmm14
- movdqu iOffExpoMask+__svml_sacosh_data_internal(%rip), %xmm5
-
-/* Add 31 to the exponent in the "large" case to get log(2 * input) */
- movups sThirtyOne+__svml_sacosh_data_internal(%rip), %xmm6
- addps sHalf+__svml_sacosh_data_internal(%rip), %xmm8
- mulps %xmm8, %xmm10
- movaps %xmm1, %xmm8
- mulps %xmm11, %xmm10
- addps %xmm10, %xmm11
- addps %xmm11, %xmm12
- maxps %xmm12, %xmm13
- minps %xmm12, %xmm9
- movaps %xmm13, %xmm15
- addps %xmm9, %xmm15
- subps %xmm15, %xmm13
- andps %xmm1, %xmm15
- orps %xmm15, %xmm3
- addps %xmm13, %xmm9
- psubd %xmm14, %xmm3
- andps %xmm1, %xmm9
- pand %xmm3, %xmm5
- psrad $23, %xmm3
- cvtdq2ps %xmm3, %xmm7
- pslld $23, %xmm3
- paddd %xmm14, %xmm5
- psubd %xmm3, %xmm4
-
-/* polynomial evaluation */
- subps %xmm2, %xmm5
- mulps %xmm4, %xmm9
- addps %xmm7, %xmm6
- movups sPoly+112+__svml_sacosh_data_internal(%rip), %xmm2
- andnps %xmm6, %xmm8
- andps %xmm1, %xmm7
- addps %xmm5, %xmm9
- mulps %xmm9, %xmm2
- orps %xmm7, %xmm8
-
-/* final reconstruction */
- mulps sLn2+__svml_sacosh_data_internal(%rip), %xmm8
- addps sPoly+96+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+80+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+64+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+48+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+32+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+16+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- addps sPoly+__svml_sacosh_data_internal(%rip), %xmm2
- mulps %xmm9, %xmm2
- mulps %xmm9, %xmm2
- addps %xmm2, %xmm9
- addps %xmm8, %xmm9
- testl %edx, %edx
-
-/* Go to special inputs processing branch */
- jne L(SPECIAL_VALUES_BRANCH)
- # LOE rbx rbp r12 r13 r14 r15 edx xmm0 xmm9
-
-/* Restore registers
- * and exit the function
- */
+ subq $72, %rsp
+ cfi_def_cfa_offset(80)
+
+ /* Compute U = X - 1 and V = X + 1, naively first. */
+ movaps %xmm0, %xmm12
+
+ /* Load constants, always including One = 1 */
+ movups sOne+__svml_sacosh_data_internal(%rip), %xmm2
+
+ /*
+ * Check that 1 < X < +inf; otherwise go to the callout function.
+ * We need the callout for X = 1 to avoid division by zero below.
+ * This test ensures that callout handles NaN and either infinity.
+ */
+ movaps %xmm0, %xmm4
+ movaps %xmm2, %xmm9
+
+ /*
+ * Compute e = -(2 * d + d^2)
+ * The first FMR is exact, and the rounding error in the other is acceptable
+ * since d and e are ~ 2^-8
+ */
+ movaps %xmm2, %xmm10
+
+ /* Finally, express Y + W = U * V accurately where Y has <= 8 bits */
+ movups sTopMask8+__svml_sacosh_data_internal(%rip), %xmm5
+
+ /*
+ * Now we feed into the log1p code, using H in place of _VARG1 and
+ * also adding L into Xl.
+ * compute 1+x as high, low parts
+ */
+ movaps %xmm2, %xmm13
+ movaps %xmm5, %xmm11
+ movaps %xmm2, %xmm3
+
+ /*
+ * Now 1 / (1 + d)
+ * = 1 / (1 + (sqrt(1 - e) - 1))
+ * = 1 / sqrt(1 - e)
+ * = 1 + 1/2 * e + 3/8 * e^2 + 5/16 * e^3 + 35/128 * e^4 + ...
+ * So compute the first three nonconstant terms of that, so that
+ * we have a relative correction (1 + Corr) to apply to S etc.
+ * C1 = 1/2
+ * C2 = 3/8
+ * C3 = 5/16
+ */
+ movups sC3+__svml_sacosh_data_internal(%rip), %xmm8
+
+ /*
+ * The following computation can go wrong for very large X, e.g.
+ * the X^2 - 1 = U * V can overflow. But for large X we have
+ * acosh(X) / log(2 X) - 1 =~= 1/(4 * X^2), so for X >= 2^30
+ * we can just later stick X back into the log and tweak up the exponent.
+ * Actually we scale X by 2^-30 and tweak the exponent up by 31,
+ * to stay in the safe range for the later log computation.
+ * Compute a flag now telling us when to do this.
+ */
+ movaps %xmm0, %xmm1
+ cmpnleps sLargestFinite+__svml_sacosh_data_internal(%rip), %xmm4
+ cmpltps sBigThreshold+__svml_sacosh_data_internal(%rip), %xmm1
+ cmpnltps %xmm0, %xmm3
+ subps %xmm2, %xmm12
+ addps %xmm0, %xmm9
+
+ /* For low-accuracy versions, naivety is harmless */
+ mulps %xmm12, %xmm9
+ orps %xmm3, %xmm4
+ movmskps %xmm4, %edx
+ andps %xmm9, %xmm11
+ movaps %xmm1, %xmm3
+
+ /*
+ * Compute R = 1/sqrt(Y + W) * (1 + d)
+ * Force R to <= 8 significant bits.
+ * This means that R * Y and R^2 * Y are exactly representable.
+ */
+ rsqrtps %xmm11, %xmm7
+ subps %xmm11, %xmm9
+ andps %xmm5, %xmm7
+ movaps %xmm2, %xmm4
+
+ /*
+ * Compute S = (Y/sqrt(Y + W)) * (1 + d)
+ * and T = (W/sqrt(Y + W)) * (1 + d)
+ * so that S + T = sqrt(Y + W) * (1 + d)
+ * S is exact, and the rounding error in T is OK.
+ */
+ mulps %xmm7, %xmm11
+ movaps %xmm7, %xmm6
+ mulps %xmm7, %xmm9
+ mulps %xmm11, %xmm6
+ mulps %xmm9, %xmm7
+
+ /*
+ * For low-accuracy versions, the computation can be done
+ * just as U + ((S + T) + (S + T) * Corr)
+ */
+ addps %xmm9, %xmm11
+ subps %xmm6, %xmm10
+ movaps %xmm2, %xmm9
+ subps %xmm7, %xmm10
+ mulps %xmm10, %xmm8
+
+ /* Now multiplex to the case X = 2^-30 * input, Xl = 0 in the "big" case. */
+ movups XScale+__svml_sacosh_data_internal(%rip), %xmm14
+ mulps %xmm0, %xmm14
+ addps sC2+__svml_sacosh_data_internal(%rip), %xmm8
+ mulps %xmm10, %xmm8
+ andnps %xmm14, %xmm3
+
+ /*
+ * Now resume the main code.
+ * reduction: compute r, n
+ */
+ movdqu iBrkValue+__svml_sacosh_data_internal(%rip), %xmm14
+ movdqu iOffExpoMask+__svml_sacosh_data_internal(%rip), %xmm5
+
+ /* Add 31 to the exponent in the "large" case to get log(2 * input) */
+ movups sThirtyOne+__svml_sacosh_data_internal(%rip), %xmm6
+ addps sHalf+__svml_sacosh_data_internal(%rip), %xmm8
+ mulps %xmm8, %xmm10
+ movaps %xmm1, %xmm8
+ mulps %xmm11, %xmm10
+ addps %xmm10, %xmm11
+ addps %xmm11, %xmm12
+ maxps %xmm12, %xmm13
+ minps %xmm12, %xmm9
+ movaps %xmm13, %xmm15
+ addps %xmm9, %xmm15
+ subps %xmm15, %xmm13
+ andps %xmm1, %xmm15
+ orps %xmm15, %xmm3
+ addps %xmm13, %xmm9
+ psubd %xmm14, %xmm3
+ andps %xmm1, %xmm9
+ pand %xmm3, %xmm5
+ psrad $23, %xmm3
+ cvtdq2ps %xmm3, %xmm7
+ pslld $23, %xmm3
+ paddd %xmm14, %xmm5
+ psubd %xmm3, %xmm4
+
+ /* polynomial evaluation */
+ subps %xmm2, %xmm5
+ mulps %xmm4, %xmm9
+ addps %xmm7, %xmm6
+ movups sPoly+112+__svml_sacosh_data_internal(%rip), %xmm2
+ andnps %xmm6, %xmm8
+ andps %xmm1, %xmm7
+ addps %xmm5, %xmm9
+ mulps %xmm9, %xmm2
+ orps %xmm7, %xmm8
+
+ /* final reconstruction */
+ mulps sLn2+__svml_sacosh_data_internal(%rip), %xmm8
+ addps sPoly+96+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+80+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+64+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+48+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+32+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+16+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ addps sPoly+__svml_sacosh_data_internal(%rip), %xmm2
+ mulps %xmm9, %xmm2
+ mulps %xmm9, %xmm2
+ addps %xmm2, %xmm9
+ addps %xmm8, %xmm9
+ testl %edx, %edx
+
+ /* Go to special inputs processing branch */
+ jne L(SPECIAL_VALUES_BRANCH)
+ # LOE rbx rbp r12 r13 r14 r15 edx xmm0 xmm9
+
+ /* Restore registers
+ * and exit the function
+ */
L(EXIT):
- movaps %xmm9, %xmm0
- addq $72, %rsp
- cfi_def_cfa_offset(8)
- ret
- cfi_def_cfa_offset(80)
-
-/* Branch to process
- * special inputs
- */
+ movaps %xmm9, %xmm0
+ addq $72, %rsp
+ cfi_def_cfa_offset(8)
+ ret
+ cfi_def_cfa_offset(80)
+
+ /* Branch to process
+ * special inputs
+ */
L(SPECIAL_VALUES_BRANCH):
- movups %xmm0, 32(%rsp)
- movups %xmm9, 48(%rsp)
- # LOE rbx rbp r12 r13 r14 r15 edx
-
- xorl %eax, %eax
- movq %r12, 16(%rsp)
- cfi_offset(12, -64)
- movl %eax, %r12d
- movq %r13, 8(%rsp)
- cfi_offset(13, -72)
- movl %edx, %r13d
- movq %r14, (%rsp)
- cfi_offset(14, -80)
- # LOE rbx rbp r15 r12d r13d
-
-/* Range mask
- * bits check
- */
+ movups %xmm0, 32(%rsp)
+ movups %xmm9, 48(%rsp)
+ # LOE rbx rbp r12 r13 r14 r15 edx
+
+ xorl %eax, %eax
+ movq %r12, 16(%rsp)
+ cfi_offset(12, -64)
+ movl %eax, %r12d
+ movq %r13, 8(%rsp)
+ cfi_offset(13, -72)
+ movl %edx, %r13d
+ movq %r14, (%rsp)
+ cfi_offset(14, -80)
+ # LOE rbx rbp 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 rbp r15 r12d r13d
+ /* Call scalar math function */
+ jc L(SCALAR_MATH_CALL)
+ # LOE rbx rbp r15 r12d r13d
-/* Special inputs
- * processing loop
- */
+ /* Special inputs
+ * processing loop
+ */
L(SPECIAL_VALUES_LOOP):
- incl %r12d
- cmpl $4, %r12d
-
-/* Check bits in range mask */
- jl L(RANGEMASK_CHECK)
- # LOE rbx rbp r15 r12d r13d
-
- movq 16(%rsp), %r12
- cfi_restore(12)
- movq 8(%rsp), %r13
- cfi_restore(13)
- movq (%rsp), %r14
- cfi_restore(14)
- movups 48(%rsp), %xmm9
-
-/* Go to exit */
- jmp L(EXIT)
- cfi_offset(12, -64)
- cfi_offset(13, -72)
- cfi_offset(14, -80)
- # LOE rbx rbp r12 r13 r14 r15 xmm9
-
-/* Scalar math fucntion call
- * to process special input
- */
+ incl %r12d
+ cmpl $4, %r12d
+
+ /* Check bits in range mask */
+ jl L(RANGEMASK_CHECK)
+ # LOE rbx rbp r15 r12d r13d
+
+ movq 16(%rsp), %r12
+ cfi_restore(12)
+ movq 8(%rsp), %r13
+ cfi_restore(13)
+ movq (%rsp), %r14
+ cfi_restore(14)
+ movups 48(%rsp), %xmm9
+
+ /* Go to exit */
+ jmp L(EXIT)
+ cfi_offset(12, -64)
+ cfi_offset(13, -72)
+ cfi_offset(14, -80)
+ # LOE rbx rbp r12 r13 r14 r15 xmm9
+
+ /* Scalar math fucntion call
+ * to process special input
+ */
L(SCALAR_MATH_CALL):
- movl %r12d, %r14d
- movss 32(%rsp,%r14,4), %xmm0
- call acoshf@PLT
- # LOE rbx rbp r14 r15 r12d r13d xmm0
+ movl %r12d, %r14d
+ movss 32(%rsp, %r14, 4), %xmm0
+ call acoshf@PLT
+ # LOE rbx rbp r14 r15 r12d r13d xmm0
- movss %xmm0, 48(%rsp,%r14,4)
+ movss %xmm0, 48(%rsp, %r14, 4)
-/* Process special inputs in loop */
- jmp L(SPECIAL_VALUES_LOOP)
- # LOE rbx rbp r15 r12d r13d
+ /* Process special inputs in loop */
+ jmp L(SPECIAL_VALUES_LOOP)
+ # LOE rbx rbp r15 r12d r13d
END(_ZGVbN4v_acoshf_sse4)
- .section .rodata, "a"
- .align 16
+ .section .rodata, "a"
+ .align 16
#ifdef __svml_sacosh_data_internal_typedef
typedef unsigned int VUINT32;
typedef struct {
- __declspec(align(16)) VUINT32 sOne[4][1];
- __declspec(align(16)) VUINT32 sPoly[8][4][1];
- __declspec(align(16)) VUINT32 iBrkValue[4][1];
- __declspec(align(16)) VUINT32 iOffExpoMask[4][1];
- __declspec(align(16)) VUINT32 sBigThreshold[4][1];
- __declspec(align(16)) VUINT32 sC2[4][1];
- __declspec(align(16)) VUINT32 sC3[4][1];
- __declspec(align(16)) VUINT32 sHalf[4][1];
- __declspec(align(16)) VUINT32 sLargestFinite[4][1];
- __declspec(align(16)) VUINT32 sThirtyOne[4][1];
- __declspec(align(16)) VUINT32 sTopMask8[4][1];
- __declspec(align(16)) VUINT32 XScale[4][1];
- __declspec(align(16)) VUINT32 sLn2[4][1];
+ __declspec(align(16)) VUINT32 sOne[4][1];
+ __declspec(align(16)) VUINT32 sPoly[8][4][1];
+ __declspec(align(16)) VUINT32 iBrkValue[4][1];
+ __declspec(align(16)) VUINT32 iOffExpoMask[4][1];
+ __declspec(align(16)) VUINT32 sBigThreshold[4][1];
+ __declspec(align(16)) VUINT32 sC2[4][1];
+ __declspec(align(16)) VUINT32 sC3[4][1];
+ __declspec(align(16)) VUINT32 sHalf[4][1];
+ __declspec(align(16)) VUINT32 sLargestFinite[4][1];
+ __declspec(align(16)) VUINT32 sThirtyOne[4][1];
+ __declspec(align(16)) VUINT32 sTopMask8[4][1];
+ __declspec(align(16)) VUINT32 XScale[4][1];
+ __declspec(align(16)) VUINT32 sLn2[4][1];
} __svml_sacosh_data_internal;
#endif
__svml_sacosh_data_internal:
- /*== sOne = SP 1.0 ==*/
- .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000
- /*== sPoly[] = SP polynomial ==*/
- .align 16
- .long 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000 /* -5.0000000000000000000000000e-01 P0 */
- .long 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94 /* 3.3333265781402587890625000e-01 P1 */
- .long 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e /* -2.5004237890243530273437500e-01 P2 */
- .long 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190 /* 2.0007920265197753906250000e-01 P3 */
- .long 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37 /* -1.6472326219081878662109375e-01 P4 */
- .long 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12 /* 1.4042308926582336425781250e-01 P5 */
- .long 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3 /* -1.5122179687023162841796875e-01 P6 */
- .long 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed /* 1.3820238411426544189453125e-01 P7 */
- /*== iBrkValue = SP 2/3 ==*/
- .align 16
- .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab
- /*== iOffExpoMask = SP significand mask ==*/
- .align 16
- .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff
- /*== sBigThreshold ==*/
- .align 16
- .long 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000
- /*== sC2 ==*/
- .align 16
- .long 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000
- /*== sC3 ==*/
- .align 16
- .long 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000
- /*== sHalf ==*/
- .align 16
- .long 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000
- /*== sLargestFinite ==*/
- .align 16
- .long 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF
- /*== sThirtyOne ==*/
- .align 16
- .long 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000
- /*== sTopMask8 ==*/
- .align 16
- .long 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000
- /*== XScale ==*/
- .align 16
- .long 0x30800000, 0x30800000, 0x30800000, 0x30800000
- /*== sLn2 = SP ln(2) ==*/
- .align 16
- .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218
- .align 16
- .type __svml_sacosh_data_internal,@object
- .size __svml_sacosh_data_internal,.-__svml_sacosh_data_internal
+ /* sOne = SP 1.0 */
+ .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000
+ /* sPoly[] = SP polynomial */
+ .align 16
+ .long 0xbf000000, 0xbf000000, 0xbf000000, 0xbf000000 /* -5.0000000000000000000000000e-01 P0 */
+ .long 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94, 0x3eaaaa94 /* 3.3333265781402587890625000e-01 P1 */
+ .long 0xbe80058e, 0xbe80058e, 0xbe80058e, 0xbe80058e /* -2.5004237890243530273437500e-01 P2 */
+ .long 0x3e4ce190, 0x3e4ce190, 0x3e4ce190, 0x3e4ce190 /* 2.0007920265197753906250000e-01 P3 */
+ .long 0xbe28ad37, 0xbe28ad37, 0xbe28ad37, 0xbe28ad37 /* -1.6472326219081878662109375e-01 P4 */
+ .long 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12, 0x3e0fcb12 /* 1.4042308926582336425781250e-01 P5 */
+ .long 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3, 0xbe1ad9e3 /* -1.5122179687023162841796875e-01 P6 */
+ .long 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed, 0x3e0d84ed /* 1.3820238411426544189453125e-01 P7 */
+ /* iBrkValue = SP 2/3 */
+ .align 16
+ .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab
+ /* iOffExpoMask = SP significand mask */
+ .align 16
+ .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff
+ /* sBigThreshold */
+ .align 16
+ .long 0x4E800000, 0x4E800000, 0x4E800000, 0x4E800000
+ /* sC2 */
+ .align 16
+ .long 0x3EC00000, 0x3EC00000, 0x3EC00000, 0x3EC00000
+ /* sC3 */
+ .align 16
+ .long 0x3EA00000, 0x3EA00000, 0x3EA00000, 0x3EA00000
+ /* sHalf */
+ .align 16
+ .long 0x3F000000, 0x3F000000, 0x3F000000, 0x3F000000
+ /* sLargestFinite */
+ .align 16
+ .long 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF, 0x7F7FFFFF
+ /* sThirtyOne */
+ .align 16
+ .long 0x41F80000, 0x41F80000, 0x41F80000, 0x41F80000
+ /* sTopMask8 */
+ .align 16
+ .long 0xFFFF0000, 0xFFFF0000, 0xFFFF0000, 0xFFFF0000
+ /* XScale */
+ .align 16
+ .long 0x30800000, 0x30800000, 0x30800000, 0x30800000
+ /* sLn2 = SP ln(2) */
+ .align 16
+ .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218
+ .align 16
+ .type __svml_sacosh_data_internal, @object
+ .size __svml_sacosh_data_internal, .-__svml_sacosh_data_internal