diff options
Diffstat (limited to 'sysdeps/x86_64/fpu/multiarch/svml_s_asinf8_core_avx2.S')
| -rw-r--r-- | sysdeps/x86_64/fpu/multiarch/svml_s_asinf8_core_avx2.S | 249 |
1 files changed, 249 insertions, 0 deletions
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_asinf8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_asinf8_core_avx2.S new file mode 100644 index 0000000000..89c156dbbb --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_asinf8_core_avx2.S @@ -0,0 +1,249 @@ +/* Function asinf vectorized with AVX2. + Copyright (C) 2021 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 + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * + * SelMask = (|x| >= 0.5) ? 1 : 0; + * R = SelMask ? sqrt(0.5 - 0.5*|x|) : |x| + * asin(x) = (SelMask ? (Pi/2 - 2*Poly(R)) : Poly(R))*(-1)^sign(x) + * + * + */ + +/* Offsets for data table __svml_sasin_data_internal + */ +#define AbsMask 0 +#define OneHalf 32 +#define SmallNorm 64 +#define One 96 +#define Two 128 +#define sqrt_coeff 160 +#define poly_coeff 224 +#define Pi2H 384 + +#include <sysdep.h> + + .text + .section .text.avx2,"ax",@progbits +ENTRY(_ZGVdN8v_asinf_avx2) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-32, %rsp + subq $96, %rsp + vmovups __svml_sasin_data_internal(%rip), %ymm5 + vmovups OneHalf+__svml_sasin_data_internal(%rip), %ymm9 + vmovups One+__svml_sasin_data_internal(%rip), %ymm6 + vmovaps %ymm0, %ymm4 + +/* x = |arg| */ + vandps %ymm4, %ymm5, %ymm3 + +/* Y = 0.5 - 0.5*x */ + vmovaps %ymm9, %ymm12 + vfnmadd231ps %ymm3, %ymm9, %ymm12 + +/* x^2 */ + vmulps %ymm3, %ymm3, %ymm7 + vcmplt_oqps %ymm3, %ymm6, %ymm8 + +/* SQ ~ -2*sqrt(Y) */ + vcmplt_oqps SmallNorm+__svml_sasin_data_internal(%rip), %ymm12, %ymm10 + vminps %ymm12, %ymm7, %ymm1 + vaddps %ymm12, %ymm12, %ymm15 + vcmpnlt_uqps %ymm9, %ymm3, %ymm0 + vrsqrtps %ymm12, %ymm11 + vmovups poly_coeff+64+__svml_sasin_data_internal(%rip), %ymm7 + vmulps %ymm1, %ymm1, %ymm6 + vmovups sqrt_coeff+__svml_sasin_data_internal(%rip), %ymm9 + vfmadd213ps poly_coeff+96+__svml_sasin_data_internal(%rip), %ymm1, %ymm7 + vmovmskps %ymm8, %edx + +/* polynomial */ + vmovups poly_coeff+__svml_sasin_data_internal(%rip), %ymm8 + vandnps %ymm11, %ymm10, %ymm13 + vmulps %ymm13, %ymm13, %ymm14 + vfmadd213ps poly_coeff+32+__svml_sasin_data_internal(%rip), %ymm1, %ymm8 + vandnps %ymm4, %ymm5, %ymm2 + vmulps %ymm15, %ymm13, %ymm5 + vfmsub213ps Two+__svml_sasin_data_internal(%rip), %ymm14, %ymm15 + vfmadd213ps %ymm7, %ymm6, %ymm8 + vfmadd213ps sqrt_coeff+32+__svml_sasin_data_internal(%rip), %ymm15, %ymm9 + vmulps %ymm15, %ymm5, %ymm15 + vfmadd213ps poly_coeff+128+__svml_sasin_data_internal(%rip), %ymm1, %ymm8 + vfmsub213ps %ymm5, %ymm15, %ymm9 + vmulps %ymm8, %ymm1, %ymm1 + vblendvps %ymm0, %ymm9, %ymm3, %ymm3 + vfmadd213ps %ymm3, %ymm3, %ymm1 + vandps Pi2H+__svml_sasin_data_internal(%rip), %ymm0, %ymm0 + vaddps %ymm1, %ymm0, %ymm10 + vxorps %ymm2, %ymm10, %ymm0 + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm4 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + 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 %ymm4, 32(%rsp) + vmovups %ymm0, 64(%rsp) + # LOE rbx r12 r13 r14 r15 edx ymm0 + + 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: -32; DW_OP_and; DW_OP_const4s: -80; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 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: -32; DW_OP_and; DW_OP_const4s: -88; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d + +/* 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 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: -80; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xb0, 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: -88; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa8, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -96; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0xa0, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 ymm0 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 32(%rsp,%r14,4), %xmm0 + call asinf@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movss %xmm0, 64(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx r15 r12d r13d +END(_ZGVdN8v_asinf_avx2) + + .section .rodata, "a" + .align 32 + +#ifdef __svml_sasin_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(32)) VUINT32 AbsMask[8][1]; + __declspec(align(32)) VUINT32 OneHalf[8][1]; + __declspec(align(32)) VUINT32 SmallNorm[8][1]; + __declspec(align(32)) VUINT32 One[8][1]; + __declspec(align(32)) VUINT32 Two[8][1]; + __declspec(align(32)) VUINT32 sqrt_coeff[2][8][1]; + __declspec(align(32)) VUINT32 poly_coeff[5][8][1]; + __declspec(align(32)) VUINT32 Pi2H[8][1]; +} __svml_sasin_data_internal; +#endif +__svml_sasin_data_internal: + /*== AbsMask ==*/ + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff + /*== OneHalf ==*/ + .align 32 + .long 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000, 0x3f000000 + /*== SmallNorm ==*/ + .align 32 + .long 0x2f800000, 0x2f800000, 0x2f800000, 0x2f800000, 0x2f800000, 0x2f800000, 0x2f800000, 0x2f800000 + /*== One ==*/ + .align 32 + .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000 + /*== Two ==*/ + .align 32 + .long 0x40000000, 0x40000000, 0x40000000, 0x40000000, 0x40000000, 0x40000000, 0x40000000, 0x40000000 + /*== sqrt_coeff[2] ==*/ + .align 32 + .long 0xbdC00004, 0xbdC00004, 0xbdC00004, 0xbdC00004, 0xbdC00004, 0xbdC00004, 0xbdC00004, 0xbdC00004 /* sqrt_coeff2 */ + .long 0x3e800001, 0x3e800001, 0x3e800001, 0x3e800001, 0x3e800001, 0x3e800001, 0x3e800001, 0x3e800001 /* sqrt_coeff1 */ + /*== poly_coeff[5] ==*/ + .align 32 + .long 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07, 0x3d2EDC07 /* poly_coeff5 */ + .long 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B, 0x3CC32A6B /* poly_coeff4 */ + .long 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4, 0x3d3A9AB4 /* poly_coeff3 */ + .long 0x3d997C12, 0x3d997C12, 0x3d997C12, 0x3d997C12, 0x3d997C12, 0x3d997C12, 0x3d997C12, 0x3d997C12 /* poly_coeff2 */ + .long 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF, 0x3e2AAAFF /* poly_coeff1 */ + /*== Pi2H ==*/ + .align 32 + .long 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB, 0x3fc90FDB + .align 32 + .type __svml_sasin_data_internal,@object + .size __svml_sasin_data_internal,.-__svml_sasin_data_internal |