/*
* ALGORITHM DESCRIPTION:
*
- * 1+x = 2^k*(xh + xl) is computed in high-low parts; xh in [1,2)
+ * 1+x = 2^k*(xh + xl) is computed in high-low parts; xh in [1, 2)
* Get short reciprocal approximation Rcp ~ 1/xh
* R = (Rcp*xh - 1.0) + Rcp*xl
* log1p(x) = k*log(2.0) - log(Rcp) + poly(R)
/* Offsets for data table __svml_slog1p_data_internal
*/
-#define SgnMask 0
-#define sOne 16
-#define sPoly 32
-#define iHiDelta 160
-#define iLoRange 176
-#define iBrkValue 192
-#define iOffExpoMask 208
-#define sLn2 224
+#define SgnMask 0
+#define sOne 16
+#define sPoly 32
+#define iHiDelta 160
+#define iLoRange 176
+#define iBrkValue 192
+#define iOffExpoMask 208
+#define sLn2 224
#include <sysdep.h>
- .text
- .section .text.sse4,"ax",@progbits
+ .section .text.sse4, "ax", @progbits
ENTRY(_ZGVbN4v_log1pf_sse4)
- subq $72, %rsp
- cfi_def_cfa_offset(80)
- movups sOne+__svml_slog1p_data_internal(%rip), %xmm7
-
-/* compute 1+x as high, low parts */
- movaps %xmm7, %xmm1
- movaps %xmm7, %xmm5
- maxps %xmm0, %xmm1
- minps %xmm0, %xmm5
- movaps %xmm1, %xmm4
-
-/* check argument value ranges */
- movdqu iHiDelta+__svml_slog1p_data_internal(%rip), %xmm2
- addps %xmm5, %xmm4
-
-/* reduction: compute r,n */
- movdqu iBrkValue+__svml_slog1p_data_internal(%rip), %xmm3
- paddd %xmm4, %xmm2
- movdqu iOffExpoMask+__svml_slog1p_data_internal(%rip), %xmm8
- subps %xmm4, %xmm1
- psubd %xmm3, %xmm4
- addps %xmm1, %xmm5
- pand %xmm4, %xmm8
- psrad $23, %xmm4
- cvtdq2ps %xmm4, %xmm10
- pslld $23, %xmm4
- movaps %xmm7, %xmm1
- paddd %xmm3, %xmm8
- psubd %xmm4, %xmm1
- mulps %xmm5, %xmm1
-
-/* polynomial evaluation */
- subps %xmm7, %xmm8
-
-/* final reconstruction */
- mulps sLn2+__svml_slog1p_data_internal(%rip), %xmm10
- addps %xmm8, %xmm1
- movups sPoly+112+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- movdqu iLoRange+__svml_slog1p_data_internal(%rip), %xmm6
- pcmpgtd %xmm2, %xmm6
- addps sPoly+96+__svml_slog1p_data_internal(%rip), %xmm9
-
-/* combine and get argument value range mask */
- movmskps %xmm6, %edx
- movups SgnMask+__svml_slog1p_data_internal(%rip), %xmm11
- mulps %xmm1, %xmm9
- andnps %xmm0, %xmm11
- addps sPoly+80+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- addps sPoly+64+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- addps sPoly+48+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- addps sPoly+32+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- addps sPoly+16+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- addps sPoly+__svml_slog1p_data_internal(%rip), %xmm9
- mulps %xmm1, %xmm9
- mulps %xmm1, %xmm9
- addps %xmm9, %xmm1
- addps %xmm10, %xmm1
- orps %xmm11, %xmm1
- testl %edx, %edx
-
-/* Go to special inputs processing branch */
- jne L(SPECIAL_VALUES_BRANCH)
- # LOE rbx rbp r12 r13 r14 r15 edx xmm0 xmm1
-
-/* Restore registers
- * and exit the function
- */
+ subq $72, %rsp
+ cfi_def_cfa_offset(80)
+ movups sOne+__svml_slog1p_data_internal(%rip), %xmm7
+
+ /* compute 1+x as high, low parts */
+ movaps %xmm7, %xmm1
+ movaps %xmm7, %xmm5
+ maxps %xmm0, %xmm1
+ minps %xmm0, %xmm5
+ movaps %xmm1, %xmm4
+
+ /* check argument value ranges */
+ movdqu iHiDelta+__svml_slog1p_data_internal(%rip), %xmm2
+ addps %xmm5, %xmm4
+
+ /* reduction: compute r, n */
+ movdqu iBrkValue+__svml_slog1p_data_internal(%rip), %xmm3
+ paddd %xmm4, %xmm2
+ movdqu iOffExpoMask+__svml_slog1p_data_internal(%rip), %xmm8
+ subps %xmm4, %xmm1
+ psubd %xmm3, %xmm4
+ addps %xmm1, %xmm5
+ pand %xmm4, %xmm8
+ psrad $23, %xmm4
+ cvtdq2ps %xmm4, %xmm10
+ pslld $23, %xmm4
+ movaps %xmm7, %xmm1
+ paddd %xmm3, %xmm8
+ psubd %xmm4, %xmm1
+ mulps %xmm5, %xmm1
+
+ /* polynomial evaluation */
+ subps %xmm7, %xmm8
+
+ /* final reconstruction */
+ mulps sLn2+__svml_slog1p_data_internal(%rip), %xmm10
+ addps %xmm8, %xmm1
+ movups sPoly+112+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ movdqu iLoRange+__svml_slog1p_data_internal(%rip), %xmm6
+ pcmpgtd %xmm2, %xmm6
+ addps sPoly+96+__svml_slog1p_data_internal(%rip), %xmm9
+
+ /* combine and get argument value range mask */
+ movmskps %xmm6, %edx
+ movups SgnMask+__svml_slog1p_data_internal(%rip), %xmm11
+ mulps %xmm1, %xmm9
+ andnps %xmm0, %xmm11
+ addps sPoly+80+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ addps sPoly+64+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ addps sPoly+48+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ addps sPoly+32+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ addps sPoly+16+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ addps sPoly+__svml_slog1p_data_internal(%rip), %xmm9
+ mulps %xmm1, %xmm9
+ mulps %xmm1, %xmm9
+ addps %xmm9, %xmm1
+ addps %xmm10, %xmm1
+ orps %xmm11, %xmm1
+ testl %edx, %edx
+
+ /* Go to special inputs processing branch */
+ jne L(SPECIAL_VALUES_BRANCH)
+ # LOE rbx rbp r12 r13 r14 r15 edx xmm0 xmm1
+
+ /* Restore registers
+ * and exit the function
+ */
L(EXIT):
- movaps %xmm1, %xmm0
- addq $72, %rsp
- cfi_def_cfa_offset(8)
- ret
- cfi_def_cfa_offset(80)
-
-/* Branch to process
- * special inputs
- */
+ movaps %xmm1, %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 %xmm1, 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 %xmm1, 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), %xmm1
-
-/* 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 xmm1
-
-/* 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), %xmm1
+
+ /* 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 xmm1
+
+ /* Scalar math fucntion call
+ * to process special input
+ */
L(SCALAR_MATH_CALL):
- movl %r12d, %r14d
- movss 32(%rsp,%r14,4), %xmm0
- call log1pf@PLT
- # LOE rbx rbp r14 r15 r12d r13d xmm0
+ movl %r12d, %r14d
+ movss 32(%rsp, %r14, 4), %xmm0
+ call log1pf@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_log1pf_sse4)
- .section .rodata, "a"
- .align 16
+ .section .rodata, "a"
+ .align 16
#ifdef __svml_slog1p_data_internal_typedef
typedef unsigned int VUINT32;
typedef struct {
- __declspec(align(16)) VUINT32 SgnMask[4][1];
- __declspec(align(16)) VUINT32 sOne[4][1];
- __declspec(align(16)) VUINT32 sPoly[8][4][1];
- __declspec(align(16)) VUINT32 iHiDelta[4][1];
- __declspec(align(16)) VUINT32 iLoRange[4][1];
- __declspec(align(16)) VUINT32 iBrkValue[4][1];
- __declspec(align(16)) VUINT32 iOffExpoMask[4][1];
- __declspec(align(16)) VUINT32 sLn2[4][1];
+ __declspec(align(16)) VUINT32 SgnMask[4][1];
+ __declspec(align(16)) VUINT32 sOne[4][1];
+ __declspec(align(16)) VUINT32 sPoly[8][4][1];
+ __declspec(align(16)) VUINT32 iHiDelta[4][1];
+ __declspec(align(16)) VUINT32 iLoRange[4][1];
+ __declspec(align(16)) VUINT32 iBrkValue[4][1];
+ __declspec(align(16)) VUINT32 iOffExpoMask[4][1];
+ __declspec(align(16)) VUINT32 sLn2[4][1];
} __svml_slog1p_data_internal;
#endif
__svml_slog1p_data_internal:
- /*== SgnMask ==*/
- .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff
- /*== sOne = SP 1.0 ==*/
- .align 16
- .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 */
- /*== iHiDelta = SP 80000000-7f000000 ==*/
- .align 16
- .long 0x01000000, 0x01000000, 0x01000000, 0x01000000
- /*== iLoRange = SP 00800000+iHiDelta ==*/
- .align 16
- .long 0x01800000, 0x01800000, 0x01800000, 0x01800000
- /*== iBrkValue = SP 2/3 ==*/
- .align 16
- .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab
- /*== iOffExpoMask = SP significand mask ==*/
- .align 16
- .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff
- /*== sLn2 = SP ln(2) ==*/
- .align 16
- .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218
- .align 16
- .type __svml_slog1p_data_internal,@object
- .size __svml_slog1p_data_internal,.-__svml_slog1p_data_internal
+ /* SgnMask */
+ .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff
+ /* sOne = SP 1.0 */
+ .align 16
+ .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 */
+ /* iHiDelta = SP 80000000-7f000000 */
+ .align 16
+ .long 0x01000000, 0x01000000, 0x01000000, 0x01000000
+ /* iLoRange = SP 00800000+iHiDelta */
+ .align 16
+ .long 0x01800000, 0x01800000, 0x01800000, 0x01800000
+ /* iBrkValue = SP 2/3 */
+ .align 16
+ .long 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab, 0x3f2aaaab
+ /* iOffExpoMask = SP significand mask */
+ .align 16
+ .long 0x007fffff, 0x007fffff, 0x007fffff, 0x007fffff
+ /* sLn2 = SP ln(2) */
+ .align 16
+ .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218
+ .align 16
+ .type __svml_slog1p_data_internal, @object
+ .size __svml_slog1p_data_internal, .-__svml_slog1p_data_internal
projects / thirdparty / glibc.git / commitdiff
