diff options
| author | Sunil K Pandey <skpgkp2@gmail.com> | 2021-12-29 08:41:18 -0800 |
|---|---|---|
| committer | Sunil K Pandey <skpgkp2@gmail.com> | 2021-12-29 11:37:29 -0800 |
| commit | 3fc9ccc20b6d0d5e4517d2e766f14ce780a228a5 (patch) | |
| tree | 21dfad9c847f54ca417c735e09b95f7f6023d690 | |
| parent | 37475ba88303929e85704693455c7294e50aba77 (diff) | |
| download | glibc-3fc9ccc20b6d0d5e4517d2e766f14ce780a228a5.tar.gz glibc-3fc9ccc20b6d0d5e4517d2e766f14ce780a228a5.tar.xz glibc-3fc9ccc20b6d0d5e4517d2e766f14ce780a228a5.zip | |
x86-64: Add vector exp2/exp2f implementation to libmvec
Implement vectorized exp2/exp2f containing SSE, AVX, AVX2 and AVX512 versions for libmvec as per vector ABI. It also contains accuracy and ABI tests for vector exp2/exp2f with regenerated ulps. Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
50 files changed, 2293 insertions, 1 deletions
diff --git a/bits/libm-simd-decl-stubs.h b/bits/libm-simd-decl-stubs.h index adf65f6bc2..36d6643eb9 100644 --- a/bits/libm-simd-decl-stubs.h +++ b/bits/libm-simd-decl-stubs.h @@ -142,4 +142,15 @@ #define __DECL_SIMD_hypotf32x #define __DECL_SIMD_hypotf64x #define __DECL_SIMD_hypotf128x + +#define __DECL_SIMD_exp2 +#define __DECL_SIMD_exp2f +#define __DECL_SIMD_exp2l +#define __DECL_SIMD_exp2f16 +#define __DECL_SIMD_exp2f32 +#define __DECL_SIMD_exp2f64 +#define __DECL_SIMD_exp2f128 +#define __DECL_SIMD_exp2f32x +#define __DECL_SIMD_exp2f64x +#define __DECL_SIMD_exp2f128x #endif diff --git a/math/bits/mathcalls.h b/math/bits/mathcalls.h index 2ed820a0dc..645088cbf3 100644 --- a/math/bits/mathcalls.h +++ b/math/bits/mathcalls.h @@ -127,7 +127,7 @@ __MATHCALL (logb,, (_Mdouble_ __x)); #ifdef __USE_ISOC99 /* Compute base-2 exponential of X. */ -__MATHCALL (exp2,, (_Mdouble_ __x)); +__MATHCALL_VEC (exp2,, (_Mdouble_ __x)); /* Compute base-2 logarithm of X. */ __MATHCALL (log2,, (_Mdouble_ __x)); diff --git a/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist b/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist index 12bb03245b..1717f2dee9 100644 --- a/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist +++ b/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist @@ -49,32 +49,40 @@ GLIBC_2.22 _ZGVeN8vvv_sincos F GLIBC_2.35 _ZGVbN2v_acos F GLIBC_2.35 _ZGVbN2v_asin F GLIBC_2.35 _ZGVbN2v_atan F +GLIBC_2.35 _ZGVbN2v_exp2 F GLIBC_2.35 _ZGVbN2vv_hypot F GLIBC_2.35 _ZGVbN4v_acosf F GLIBC_2.35 _ZGVbN4v_asinf F GLIBC_2.35 _ZGVbN4v_atanf F +GLIBC_2.35 _ZGVbN4v_exp2f F GLIBC_2.35 _ZGVbN4vv_hypotf F GLIBC_2.35 _ZGVcN4v_acos F GLIBC_2.35 _ZGVcN4v_asin F GLIBC_2.35 _ZGVcN4v_atan F +GLIBC_2.35 _ZGVcN4v_exp2 F GLIBC_2.35 _ZGVcN4vv_hypot F GLIBC_2.35 _ZGVcN8v_acosf F GLIBC_2.35 _ZGVcN8v_asinf F GLIBC_2.35 _ZGVcN8v_atanf F +GLIBC_2.35 _ZGVcN8v_exp2f F GLIBC_2.35 _ZGVcN8vv_hypotf F GLIBC_2.35 _ZGVdN4v_acos F GLIBC_2.35 _ZGVdN4v_asin F GLIBC_2.35 _ZGVdN4v_atan F +GLIBC_2.35 _ZGVdN4v_exp2 F GLIBC_2.35 _ZGVdN4vv_hypot F GLIBC_2.35 _ZGVdN8v_acosf F GLIBC_2.35 _ZGVdN8v_asinf F GLIBC_2.35 _ZGVdN8v_atanf F +GLIBC_2.35 _ZGVdN8v_exp2f F GLIBC_2.35 _ZGVdN8vv_hypotf F GLIBC_2.35 _ZGVeN16v_acosf F GLIBC_2.35 _ZGVeN16v_asinf F GLIBC_2.35 _ZGVeN16v_atanf F +GLIBC_2.35 _ZGVeN16v_exp2f F GLIBC_2.35 _ZGVeN16vv_hypotf F GLIBC_2.35 _ZGVeN8v_acos F GLIBC_2.35 _ZGVeN8v_asin F GLIBC_2.35 _ZGVeN8v_atan F +GLIBC_2.35 _ZGVeN8v_exp2 F GLIBC_2.35 _ZGVeN8vv_hypot F diff --git a/sysdeps/x86/fpu/bits/math-vector.h b/sysdeps/x86/fpu/bits/math-vector.h index 437977c5fd..c7a972521b 100644 --- a/sysdeps/x86/fpu/bits/math-vector.h +++ b/sysdeps/x86/fpu/bits/math-vector.h @@ -74,6 +74,10 @@ # define __DECL_SIMD_hypot __DECL_SIMD_x86_64 # undef __DECL_SIMD_hypotf # define __DECL_SIMD_hypotf __DECL_SIMD_x86_64 +# undef __DECL_SIMD_exp2 +# define __DECL_SIMD_exp2 __DECL_SIMD_x86_64 +# undef __DECL_SIMD_exp2f +# define __DECL_SIMD_exp2f __DECL_SIMD_x86_64 # endif #endif diff --git a/sysdeps/x86/fpu/finclude/math-vector-fortran.h b/sysdeps/x86/fpu/finclude/math-vector-fortran.h index cda31479a6..0994e6dfac 100644 --- a/sysdeps/x86/fpu/finclude/math-vector-fortran.h +++ b/sysdeps/x86/fpu/finclude/math-vector-fortran.h @@ -36,6 +36,8 @@ !GCC$ builtin (asinf) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (hypot) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (hypotf) attributes simd (notinbranch) if('x86_64') +!GCC$ builtin (exp2) attributes simd (notinbranch) if('x86_64') +!GCC$ builtin (exp2f) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (cos) attributes simd (notinbranch) if('x32') !GCC$ builtin (cosf) attributes simd (notinbranch) if('x32') @@ -57,3 +59,5 @@ !GCC$ builtin (asinf) attributes simd (notinbranch) if('x32') !GCC$ builtin (hypot) attributes simd (notinbranch) if('x32') !GCC$ builtin (hypotf) attributes simd (notinbranch) if('x32') +!GCC$ builtin (exp2) attributes simd (notinbranch) if('x32') +!GCC$ builtin (exp2f) attributes simd (notinbranch) if('x32') diff --git a/sysdeps/x86_64/fpu/Makeconfig b/sysdeps/x86_64/fpu/Makeconfig index 7769a02731..03b2364417 100644 --- a/sysdeps/x86_64/fpu/Makeconfig +++ b/sysdeps/x86_64/fpu/Makeconfig @@ -27,6 +27,7 @@ libmvec-funcs = \ atan \ cos \ exp \ + exp2 \ hypot \ log \ pow \ diff --git a/sysdeps/x86_64/fpu/Versions b/sysdeps/x86_64/fpu/Versions index e359e5dc2c..12b7ad1830 100644 --- a/sysdeps/x86_64/fpu/Versions +++ b/sysdeps/x86_64/fpu/Versions @@ -17,10 +17,12 @@ libmvec { _ZGVbN2v_acos; _ZGVcN4v_acos; _ZGVdN4v_acos; _ZGVeN8v_acos; _ZGVbN2v_asin; _ZGVcN4v_asin; _ZGVdN4v_asin; _ZGVeN8v_asin; _ZGVbN2v_atan; _ZGVcN4v_atan; _ZGVdN4v_atan; _ZGVeN8v_atan; + _ZGVbN2v_exp2; _ZGVcN4v_exp2; _ZGVdN4v_exp2; _ZGVeN8v_exp2; _ZGVbN2vv_hypot; _ZGVcN4vv_hypot; _ZGVdN4vv_hypot; _ZGVeN8vv_hypot; _ZGVbN4v_acosf; _ZGVcN8v_acosf; _ZGVdN8v_acosf; _ZGVeN16v_acosf; _ZGVbN4v_asinf; _ZGVcN8v_asinf; _ZGVdN8v_asinf; _ZGVeN16v_asinf; _ZGVbN4v_atanf; _ZGVcN8v_atanf; _ZGVdN8v_atanf; _ZGVeN16v_atanf; + _ZGVbN4v_exp2f; _ZGVcN8v_exp2f; _ZGVdN8v_exp2f; _ZGVeN16v_exp2f; _ZGVbN4vv_hypotf; _ZGVcN8vv_hypotf; _ZGVdN8vv_hypotf; _ZGVeN16vv_hypotf; } } diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps index a7513ec94e..bc4479ad39 100644 --- a/sysdeps/x86_64/fpu/libm-test-ulps +++ b/sysdeps/x86_64/fpu/libm-test-ulps @@ -1276,6 +1276,26 @@ float: 1 float128: 2 ldouble: 1 +Function: "exp2_vlen16": +float: 1 + +Function: "exp2_vlen2": +double: 1 + +Function: "exp2_vlen4": +double: 1 +float: 1 + +Function: "exp2_vlen4_avx2": +double: 1 + +Function: "exp2_vlen8": +double: 1 +float: 1 + +Function: "exp2_vlen8_avx2": +float: 1 + Function: "exp_downward": double: 1 float: 1 diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core-sse2.S new file mode 100644 index 0000000000..330260baaa --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core-sse2.S @@ -0,0 +1,20 @@ +/* SSE2 version of vectorized exp2, vector length is 2. + 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/>. */ + +#define _ZGVbN2v_exp2 _ZGVbN2v_exp2_sse2 +#include "../svml_d_exp22_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core.c new file mode 100644 index 0000000000..e0cf198030 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core.c @@ -0,0 +1,27 @@ +/* Multiple versions of vectorized exp2, vector length is 2. + 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/>. */ + +#define SYMBOL_NAME _ZGVbN2v_exp2 +#include "ifunc-mathvec-sse4_1.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVbN2v_exp2, __GI__ZGVbN2v_exp2, __redirect__ZGVbN2v_exp2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core_sse4.S new file mode 100644 index 0000000000..7388c242f6 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp22_core_sse4.S @@ -0,0 +1,325 @@ +/* Function exp2 vectorized with SSE4. + 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: + * + * exp2(x) = 2^n * T[j] * (1 + P(y)) + * where + * x = m*(1/K) + y, y in [-1/K..1/K] + * m = n*K + j, m,n,j - signed integer, j in [-K/2..K/2] + * + * values of 2^j/K are tabulated + * + * P(y) is a minimax polynomial approximation of exp2(x)-1 + * on small interval [-1/K..1/K] + * + * Special cases: + * + * exp2(NaN) = NaN + * exp2(+INF) = +INF + * exp2(-INF) = 0 + * exp2(x) = 1 for subnormals + * For IEEE double + * if x >= 1024.0 then exp2(x) overflows + * if x < -1076.0 then exp2(x) underflows + * + */ + +/* Offsets for data table __svml_dexp2_data_internal + */ +#define _dbT 0 +#define _dbShifter 1024 +#define _dPC1 1040 +#define _dPC2 1056 +#define _dPC3 1072 +#define _dPC4 1088 +#define _lIndexMask 1104 +#define _iAbsMask 1120 +#define _iDomainRange 1136 + +#include <sysdep.h> + + .text + .section .text.sse4,"ax",@progbits +ENTRY(_ZGVbN2v_exp2_sse4) + subq $72, %rsp + cfi_def_cfa_offset(80) + +/* R */ + movaps %xmm0, %xmm7 + movups _dbShifter+__svml_dexp2_data_internal(%rip), %xmm1 + +/* out, basePtr, iIndex, iBaseOfs, iSize, iGran, iOfs */ + lea __svml_dexp2_data_internal(%rip), %rsi + +/* Load arument */ + movaps %xmm1, %xmm10 + addpd %xmm0, %xmm10 + movaps %xmm10, %xmm6 + subpd %xmm1, %xmm6 + subpd %xmm6, %xmm7 + +/* + * Polynomial + * poly(dN) = a1*dR+...+a4*dR^4 + */ + movups _dPC4+__svml_dexp2_data_internal(%rip), %xmm8 + mulpd %xmm7, %xmm8 + addpd _dPC3+__svml_dexp2_data_internal(%rip), %xmm8 + mulpd %xmm7, %xmm8 + addpd _dPC2+__svml_dexp2_data_internal(%rip), %xmm8 + movdqu _lIndexMask+__svml_dexp2_data_internal(%rip), %xmm9 + +/* Index and lookup */ + movdqa %xmm9, %xmm5 + pandn %xmm10, %xmm9 + pand %xmm10, %xmm5 + +/* 2^N */ + psllq $45, %xmm9 + movd %xmm5, %eax + movq _iAbsMask+__svml_dexp2_data_internal(%rip), %xmm2 + +/* Check for overflow\underflow */ + pshufd $221, %xmm0, %xmm4 + pextrw $4, %xmm5, %ecx + +/* a1+...+a4*dR^3 ! */ + mulpd %xmm7, %xmm8 + shll $3, %eax + pand %xmm2, %xmm4 + shll $3, %ecx + movq (%rsi,%rax), %xmm1 + movhpd (%rsi,%rcx), %xmm1 + +/* dR=dR*dT */ + mulpd %xmm1, %xmm7 + addpd _dPC1+__svml_dexp2_data_internal(%rip), %xmm8 + +/* + * Reconstruction + * exp2 = {2^N later}*(Tj+Tj*poly) + * dN = dT+dT*dR*(a1+...+a4*dR^3) + */ + mulpd %xmm7, %xmm8 + addpd %xmm8, %xmm1 + movq _iDomainRange+__svml_dexp2_data_internal(%rip), %xmm3 + pcmpgtd %xmm3, %xmm4 + movmskps %xmm4, %edx + +/* quick 2^N */ + paddq %xmm9, %xmm1 + andl $3, %edx + +/* Finish */ + +/* 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 + */ + +L(SPECIAL_VALUES_BRANCH): + movups %xmm0, 32(%rsp) + movups %xmm1, 48(%rsp) + # LOE rbx rbp r12 r13 r14 r15 edx xmm1 + + 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 + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx rbp r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $2, %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 + movsd 32(%rsp,%r14,8), %xmm0 + call exp2@PLT + # LOE rbx rbp r14 r15 r12d r13d xmm0 + + movsd %xmm0, 48(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx rbp r15 r12d r13d +END(_ZGVbN2v_exp2_sse4) + + .section .rodata, "a" + .align 16 + +#ifdef __svml_dexp2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct +{ + __declspec(align(16)) VUINT32 _dbT[(1<<7)][2]; + __declspec(align(16)) VUINT32 _dbShifter[2][2]; + __declspec(align(16)) VUINT32 _dPC1[2][2]; + __declspec(align(16)) VUINT32 _dPC2[2][2]; + __declspec(align(16)) VUINT32 _dPC3[2][2]; + __declspec(align(16)) VUINT32 _dPC4[2][2]; + __declspec(align(16)) VUINT32 _lIndexMask[2][2]; + __declspec(align(16)) VUINT32 _iAbsMask[4][1]; + __declspec(align(16)) VUINT32 _iDomainRange[4][1]; +} __svml_dexp2_data_internal; +#endif +__svml_dexp2_data_internal: + /*== _dbT ==*/ + .quad 0x3ff0000000000000, 0x3ff0163da9fb3335 /*2^( 0 /128),2^( 1 /128)*/ + .quad 0x3ff02c9a3e778061, 0x3ff04315e86e7f85 /*2^( 2 /128),2^( 3 /128)*/ + .quad 0x3ff059b0d3158574, 0x3ff0706b29ddf6de /*2^( 4 /128),2^( 5 /128)*/ + .quad 0x3ff0874518759bc8, 0x3ff09e3ecac6f383 /*2^( 6 /128),2^( 7 /128)*/ + .quad 0x3ff0b5586cf9890f, 0x3ff0cc922b7247f7 /*2^( 8 /128),2^( 9 /128)*/ + .quad 0x3ff0e3ec32d3d1a2, 0x3ff0fb66affed31b /*2^( 10 /128),2^( 11 /128)*/ + .quad 0x3ff11301d0125b51, 0x3ff12abdc06c31cc /*2^( 12 /128),2^( 13 /128)*/ + .quad 0x3ff1429aaea92de0, 0x3ff15a98c8a58e51 /*2^( 14 /128),2^( 15 /128)*/ + .quad 0x3ff172b83c7d517b, 0x3ff18af9388c8dea /*2^( 16 /128),2^( 17 /128)*/ + .quad 0x3ff1a35beb6fcb75, 0x3ff1bbe084045cd4 /*2^( 18 /128),2^( 19 /128)*/ + .quad 0x3ff1d4873168b9aa, 0x3ff1ed5022fcd91d /*2^( 20 /128),2^( 21 /128)*/ + .quad 0x3ff2063b88628cd6, 0x3ff21f49917ddc96 /*2^( 22 /128),2^( 23 /128)*/ + .quad 0x3ff2387a6e756238, 0x3ff251ce4fb2a63f /*2^( 24 /128),2^( 25 /128)*/ + .quad 0x3ff26b4565e27cdd, 0x3ff284dfe1f56381 /*2^( 26 /128),2^( 27 /128)*/ + .quad 0x3ff29e9df51fdee1, 0x3ff2b87fd0dad990 /*2^( 28 /128),2^( 29 /128)*/ + .quad 0x3ff2d285a6e4030b, 0x3ff2ecafa93e2f56 /*2^( 30 /128),2^( 31 /128)*/ + .quad 0x3ff306fe0a31b715, 0x3ff32170fc4cd831 /*2^( 32 /128),2^( 33 /128)*/ + .quad 0x3ff33c08b26416ff, 0x3ff356c55f929ff1 /*2^( 34 /128),2^( 35 /128)*/ + .quad 0x3ff371a7373aa9cb, 0x3ff38cae6d05d866 /*2^( 36 /128),2^( 37 /128)*/ + .quad 0x3ff3a7db34e59ff7, 0x3ff3c32dc313a8e5 /*2^( 38 /128),2^( 39 /128)*/ + .quad 0x3ff3dea64c123422, 0x3ff3fa4504ac801c /*2^( 40 /128),2^( 41 /128)*/ + .quad 0x3ff4160a21f72e2a, 0x3ff431f5d950a897 /*2^( 42 /128),2^( 43 /128)*/ + .quad 0x3ff44e086061892d, 0x3ff46a41ed1d0057 /*2^( 44 /128),2^( 45 /128)*/ + .quad 0x3ff486a2b5c13cd0, 0x3ff4a32af0d7d3de /*2^( 46 /128),2^( 47 /128)*/ + .quad 0x3ff4bfdad5362a27, 0x3ff4dcb299fddd0d /*2^( 48 /128),2^( 49 /128)*/ + .quad 0x3ff4f9b2769d2ca7, 0x3ff516daa2cf6642 /*2^( 50 /128),2^( 51 /128)*/ + .quad 0x3ff5342b569d4f82, 0x3ff551a4ca5d920f /*2^( 52 /128),2^( 53 /128)*/ + .quad 0x3ff56f4736b527da, 0x3ff58d12d497c7fd /*2^( 54 /128),2^( 55 /128)*/ + .quad 0x3ff5ab07dd485429, 0x3ff5c9268a5946b7 /*2^( 56 /128),2^( 57 /128)*/ + .quad 0x3ff5e76f15ad2148, 0x3ff605e1b976dc09 /*2^( 58 /128),2^( 59 /128)*/ + .quad 0x3ff6247eb03a5585, 0x3ff6434634ccc320 /*2^( 60 /128),2^( 61 /128)*/ + .quad 0x3ff6623882552225, 0x3ff68155d44ca973 /*2^( 62 /128),2^( 63 /128)*/ + .quad 0x3ff6a09e667f3bcd, 0x3ff6c012750bdabf /*2^( 64 /128),2^( 65 /128)*/ + .quad 0x3ff6dfb23c651a2f, 0x3ff6ff7df9519484 /*2^( 66 /128),2^( 67 /128)*/ + .quad 0x3ff71f75e8ec5f74, 0x3ff73f9a48a58174 /*2^( 68 /128),2^( 69 /128)*/ + .quad 0x3ff75feb564267c9, 0x3ff780694fde5d3f /*2^( 70 /128),2^( 71 /128)*/ + .quad 0x3ff7a11473eb0187, 0x3ff7c1ed0130c132 /*2^( 72 /128),2^( 73 /128)*/ + .quad 0x3ff7e2f336cf4e62, 0x3ff80427543e1a12 /*2^( 74 /128),2^( 75 /128)*/ + .quad 0x3ff82589994cce13, 0x3ff8471a4623c7ad /*2^( 76 /128),2^( 77 /128)*/ + .quad 0x3ff868d99b4492ed, 0x3ff88ac7d98a6699 /*2^( 78 /128),2^( 79 /128)*/ + .quad 0x3ff8ace5422aa0db, 0x3ff8cf3216b5448c /*2^( 80 /128),2^( 81 /128)*/ + .quad 0x3ff8f1ae99157736, 0x3ff9145b0b91ffc6 /*2^( 82 /128),2^( 83 /128)*/ + .quad 0x3ff93737b0cdc5e5, 0x3ff95a44cbc8520f /*2^( 84 /128),2^( 85 /128)*/ + .quad 0x3ff97d829fde4e50, 0x3ff9a0f170ca07ba /*2^( 86 /128),2^( 87 /128)*/ + .quad 0x3ff9c49182a3f090, 0x3ff9e86319e32323 /*2^( 88 /128),2^( 89 /128)*/ + .quad 0x3ffa0c667b5de565, 0x3ffa309bec4a2d33 /*2^( 90 /128),2^( 91 /128)*/ + .quad 0x3ffa5503b23e255d, 0x3ffa799e1330b358 /*2^( 92 /128),2^( 93 /128)*/ + .quad 0x3ffa9e6b5579fdbf, 0x3ffac36bbfd3f37a /*2^( 94 /128),2^( 95 /128)*/ + .quad 0x3ffae89f995ad3ad, 0x3ffb0e07298db666 /*2^( 96 /128),2^( 97 /128)*/ + .quad 0x3ffb33a2b84f15fb, 0x3ffb59728de5593a /*2^( 98 /128),2^( 99 /128)*/ + .quad 0x3ffb7f76f2fb5e47, 0x3ffba5b030a1064a /*2^( 100 /128),2^( 101 /128)*/ + .quad 0x3ffbcc1e904bc1d2, 0x3ffbf2c25bd71e09 /*2^( 102 /128),2^( 103 /128)*/ + .quad 0x3ffc199bdd85529c, 0x3ffc40ab5fffd07a /*2^( 104 /128),2^( 105 /128)*/ + .quad 0x3ffc67f12e57d14b, 0x3ffc8f6d9406e7b5 /*2^( 106 /128),2^( 107 /128)*/ + .quad 0x3ffcb720dcef9069, 0x3ffcdf0b555dc3fa /*2^( 108 /128),2^( 109 /128)*/ + .quad 0x3ffd072d4a07897c, 0x3ffd2f87080d89f2 /*2^( 110 /128),2^( 111 /128)*/ + .quad 0x3ffd5818dcfba487, 0x3ffd80e316c98398 /*2^( 112 /128),2^( 113 /128)*/ + .quad 0x3ffda9e603db3285, 0x3ffdd321f301b460 /*2^( 114 /128),2^( 115 /128)*/ + .quad 0x3ffdfc97337b9b5f, 0x3ffe264614f5a129 /*2^( 116 /128),2^( 117 /128)*/ + .quad 0x3ffe502ee78b3ff6, 0x3ffe7a51fbc74c83 /*2^( 118 /128),2^( 119 /128)*/ + .quad 0x3ffea4afa2a490da, 0x3ffecf482d8e67f1 /*2^( 120 /128),2^( 121 /128)*/ + .quad 0x3ffefa1bee615a27, 0x3fff252b376bba97 /*2^( 122 /128),2^( 123 /128)*/ + .quad 0x3fff50765b6e4540, 0x3fff7bfdad9cbe14 /*2^( 124 /128),2^( 125 /128)*/ + .quad 0x3fffa7c1819e90d8, 0x3fffd3c22b8f71f1 /*2^( 126 /128),2^( 127 /128)*/ + .align 16 + .quad 0x42c8000000000000, 0x42c8000000000000 /* _dbShifter - 0x433-7=0x42c shifted right on K!*/ + //log2(relerr) = -53.547756365162 + .align 16 + .quad 0x3fe62e42fefa3685, 0x3fe62e42fefa3685 /* _dPC1 */ + .align 16 + .quad 0x3fcebfbdff82ca48, 0x3fcebfbdff82ca48 /* _dPC2 */ + .align 16 + .quad 0x3fac6b09b180f045, 0x3fac6b09b180f045 /* _dPC3 */ + .align 16 + .quad 0x3f83b2ab5bb1268f, 0x3f83b2ab5bb1268f /* _dPC4 */ + .align 16 + .quad 0x000000000000007f, 0x000000000000007f /* _lIndexMask =(2^K-1)*/ + .align 16 + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* _iAbsMask */ + .align 16 + .long 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff /* _iDomainRange */ + .align 16 + .type __svml_dexp2_data_internal,@object + .size __svml_dexp2_data_internal,.-__svml_dexp2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core-sse.S new file mode 100644 index 0000000000..51c5de1100 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core-sse.S @@ -0,0 +1,20 @@ +/* SSE version of vectorized exp2, vector length is 4. + 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/>. */ + +#define _ZGVdN4v_exp2 _ZGVdN4v_exp2_sse_wrapper +#include "../svml_d_exp24_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core.c new file mode 100644 index 0000000000..bb979afde6 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core.c @@ -0,0 +1,27 @@ +/* Multiple versions of vectorized exp2, vector length is 4. + 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/>. */ + +#define SYMBOL_NAME _ZGVdN4v_exp2 +#include "ifunc-mathvec-avx2.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVdN4v_exp2, __GI__ZGVdN4v_exp2, __redirect__ZGVdN4v_exp2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core_avx2.S new file mode 100644 index 0000000000..6aaadafeeb --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp24_core_avx2.S @@ -0,0 +1,341 @@ +/* Function exp2 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: + * + * exp2(x) = 2^n * T[j] * (1 + P(y)) + * where + * x = m*(1/K) + y, y in [-1/K..1/K] + * m = n*K + j, m,n,j - signed integer, j in [-K/2..K/2] + * + * values of 2^j/K are tabulated + * + * P(y) is a minimax polynomial approximation of exp2(x)-1 + * on small interval [-1/K..1/K] + * + * Special cases: + * + * exp2(NaN) = NaN + * exp2(+INF) = +INF + * exp2(-INF) = 0 + * exp2(x) = 1 for subnormals + * For IEEE double + * if x >= 1024.0 then exp2(x) overflows + * if x < -1076.0 then exp2(x) underflows + * + */ + +/* Offsets for data table __svml_dexp2_data_internal + */ +#define _dbT 0 +#define _dbShifter 1024 +#define _dPC1 1056 +#define _dPC2 1088 +#define _dPC3 1120 +#define _dPC4 1152 +#define _lIndexMask 1184 +#define _iAbsMask 1216 +#define _iDomainRange 1248 + +#include <sysdep.h> + + .text + .section .text.avx2,"ax",@progbits +ENTRY(_ZGVdN4v_exp2_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 + +/* out, basePtr, iIndex, iBaseOfs, iSize, iGran, iOfs */ + lea __svml_dexp2_data_internal(%rip), %r8 + vmovupd _dbShifter+__svml_dexp2_data_internal(%rip), %ymm4 + vmovupd _lIndexMask+__svml_dexp2_data_internal(%rip), %ymm3 + vmovapd %ymm0, %ymm1 + +/* Load arument */ + vaddpd %ymm4, %ymm1, %ymm2 + vsubpd %ymm4, %ymm2, %ymm0 + +/* Index and lookup */ + vandps %ymm3, %ymm2, %ymm9 + vpandn %ymm2, %ymm3, %ymm2 + +/* 2^N */ + vpsllq $45, %ymm2, %ymm3 + +/* R */ + vsubpd %ymm0, %ymm1, %ymm15 + +/* Check for overflow\underflow */ + vextractf128 $1, %ymm1, %xmm5 + +/* + * Polynomial + * poly(dN) = a1*dR+...+a4*dR^4 + */ + vmovupd _dPC4+__svml_dexp2_data_internal(%rip), %ymm0 + vshufps $221, %xmm5, %xmm1, %xmm6 + vandps _iAbsMask+__svml_dexp2_data_internal(%rip), %xmm6, %xmm7 + vpcmpgtd _iDomainRange+__svml_dexp2_data_internal(%rip), %xmm7, %xmm8 + vfmadd213pd _dPC3+__svml_dexp2_data_internal(%rip), %ymm15, %ymm0 + vmovmskps %xmm8, %eax + vfmadd213pd _dPC2+__svml_dexp2_data_internal(%rip), %ymm15, %ymm0 + +/* a1+...+a4*dR^3 ! */ + vfmadd213pd _dPC1+__svml_dexp2_data_internal(%rip), %ymm15, %ymm0 + vextractf128 $1, %ymm9, %xmm12 + vmovd %xmm9, %edx + vmovd %xmm12, %esi + shll $3, %edx + vpextrd $2, %xmm9, %ecx + shll $3, %esi + vpextrd $2, %xmm12, %edi + shll $3, %ecx + vmovq (%r8,%rdx), %xmm10 + shll $3, %edi + vmovq (%r8,%rsi), %xmm13 + vmovhpd (%r8,%rcx), %xmm10, %xmm11 + vmovhpd (%r8,%rdi), %xmm13, %xmm14 + vinsertf128 $1, %xmm14, %ymm11, %ymm4 + +/* dR=dR*dT */ + vmulpd %ymm15, %ymm4, %ymm15 + +/* + * Reconstruction + * exp2 = {2^N later}*(Tj+Tj*poly) + * dN = dT+dT*dR*(a1+...+a4*dR^3) + */ + vfmadd213pd %ymm4, %ymm15, %ymm0 + +/* quick 2^N */ + vpaddq %ymm3, %ymm0, %ymm0 + +/* Finish */ + testl %eax, %eax + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 eax ymm0 ymm1 + +/* 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): + vmovupd %ymm1, 32(%rsp) + vmovupd %ymm0, 64(%rsp) + # LOE rbx r12 r13 r14 r15 eax ymm0 + + xorl %edx, %edx + # 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 %edx, %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 %eax, %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 $4, %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) + vmovupd 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 + movsd 32(%rsp,%r14,8), %xmm0 + call exp2@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movsd %xmm0, 64(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx r15 r12d r13d +END(_ZGVdN4v_exp2_avx2) + + .section .rodata, "a" + .align 32 + +#ifdef __svml_dexp2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct +{ + __declspec(align(32)) VUINT32 _dbT[(1<<7)][2]; + __declspec(align(32)) VUINT32 _dbShifter[4][2]; + __declspec(align(32)) VUINT32 _dPC1[4][2]; + __declspec(align(32)) VUINT32 _dPC2[4][2]; + __declspec(align(32)) VUINT32 _dPC3[4][2]; + __declspec(align(32)) VUINT32 _dPC4[4][2]; + __declspec(align(32)) VUINT32 _lIndexMask[4][2]; + __declspec(align(32)) VUINT32 _iAbsMask[8][1]; + __declspec(align(32)) VUINT32 _iDomainRange[8][1]; +} __svml_dexp2_data_internal; +#endif +__svml_dexp2_data_internal: + /*== _dbT ==*/ + .quad 0x3ff0000000000000, 0x3ff0163da9fb3335 /*2^( 0 /128),2^( 1 /128)*/ + .quad 0x3ff02c9a3e778061, 0x3ff04315e86e7f85 /*2^( 2 /128),2^( 3 /128)*/ + .quad 0x3ff059b0d3158574, 0x3ff0706b29ddf6de /*2^( 4 /128),2^( 5 /128)*/ + .quad 0x3ff0874518759bc8, 0x3ff09e3ecac6f383 /*2^( 6 /128),2^( 7 /128)*/ + .quad 0x3ff0b5586cf9890f, 0x3ff0cc922b7247f7 /*2^( 8 /128),2^( 9 /128)*/ + .quad 0x3ff0e3ec32d3d1a2, 0x3ff0fb66affed31b /*2^( 10 /128),2^( 11 /128)*/ + .quad 0x3ff11301d0125b51, 0x3ff12abdc06c31cc /*2^( 12 /128),2^( 13 /128)*/ + .quad 0x3ff1429aaea92de0, 0x3ff15a98c8a58e51 /*2^( 14 /128),2^( 15 /128)*/ + .quad 0x3ff172b83c7d517b, 0x3ff18af9388c8dea /*2^( 16 /128),2^( 17 /128)*/ + .quad 0x3ff1a35beb6fcb75, 0x3ff1bbe084045cd4 /*2^( 18 /128),2^( 19 /128)*/ + .quad 0x3ff1d4873168b9aa, 0x3ff1ed5022fcd91d /*2^( 20 /128),2^( 21 /128)*/ + .quad 0x3ff2063b88628cd6, 0x3ff21f49917ddc96 /*2^( 22 /128),2^( 23 /128)*/ + .quad 0x3ff2387a6e756238, 0x3ff251ce4fb2a63f /*2^( 24 /128),2^( 25 /128)*/ + .quad 0x3ff26b4565e27cdd, 0x3ff284dfe1f56381 /*2^( 26 /128),2^( 27 /128)*/ + .quad 0x3ff29e9df51fdee1, 0x3ff2b87fd0dad990 /*2^( 28 /128),2^( 29 /128)*/ + .quad 0x3ff2d285a6e4030b, 0x3ff2ecafa93e2f56 /*2^( 30 /128),2^( 31 /128)*/ + .quad 0x3ff306fe0a31b715, 0x3ff32170fc4cd831 /*2^( 32 /128),2^( 33 /128)*/ + .quad 0x3ff33c08b26416ff, 0x3ff356c55f929ff1 /*2^( 34 /128),2^( 35 /128)*/ + .quad 0x3ff371a7373aa9cb, 0x3ff38cae6d05d866 /*2^( 36 /128),2^( 37 /128)*/ + .quad 0x3ff3a7db34e59ff7, 0x3ff3c32dc313a8e5 /*2^( 38 /128),2^( 39 /128)*/ + .quad 0x3ff3dea64c123422, 0x3ff3fa4504ac801c /*2^( 40 /128),2^( 41 /128)*/ + .quad 0x3ff4160a21f72e2a, 0x3ff431f5d950a897 /*2^( 42 /128),2^( 43 /128)*/ + .quad 0x3ff44e086061892d, 0x3ff46a41ed1d0057 /*2^( 44 /128),2^( 45 /128)*/ + .quad 0x3ff486a2b5c13cd0, 0x3ff4a32af0d7d3de /*2^( 46 /128),2^( 47 /128)*/ + .quad 0x3ff4bfdad5362a27, 0x3ff4dcb299fddd0d /*2^( 48 /128),2^( 49 /128)*/ + .quad 0x3ff4f9b2769d2ca7, 0x3ff516daa2cf6642 /*2^( 50 /128),2^( 51 /128)*/ + .quad 0x3ff5342b569d4f82, 0x3ff551a4ca5d920f /*2^( 52 /128),2^( 53 /128)*/ + .quad 0x3ff56f4736b527da, 0x3ff58d12d497c7fd /*2^( 54 /128),2^( 55 /128)*/ + .quad 0x3ff5ab07dd485429, 0x3ff5c9268a5946b7 /*2^( 56 /128),2^( 57 /128)*/ + .quad 0x3ff5e76f15ad2148, 0x3ff605e1b976dc09 /*2^( 58 /128),2^( 59 /128)*/ + .quad 0x3ff6247eb03a5585, 0x3ff6434634ccc320 /*2^( 60 /128),2^( 61 /128)*/ + .quad 0x3ff6623882552225, 0x3ff68155d44ca973 /*2^( 62 /128),2^( 63 /128)*/ + .quad 0x3ff6a09e667f3bcd, 0x3ff6c012750bdabf /*2^( 64 /128),2^( 65 /128)*/ + .quad 0x3ff6dfb23c651a2f, 0x3ff6ff7df9519484 /*2^( 66 /128),2^( 67 /128)*/ + .quad 0x3ff71f75e8ec5f74, 0x3ff73f9a48a58174 /*2^( 68 /128),2^( 69 /128)*/ + .quad 0x3ff75feb564267c9, 0x3ff780694fde5d3f /*2^( 70 /128),2^( 71 /128)*/ + .quad 0x3ff7a11473eb0187, 0x3ff7c1ed0130c132 /*2^( 72 /128),2^( 73 /128)*/ + .quad 0x3ff7e2f336cf4e62, 0x3ff80427543e1a12 /*2^( 74 /128),2^( 75 /128)*/ + .quad 0x3ff82589994cce13, 0x3ff8471a4623c7ad /*2^( 76 /128),2^( 77 /128)*/ + .quad 0x3ff868d99b4492ed, 0x3ff88ac7d98a6699 /*2^( 78 /128),2^( 79 /128)*/ + .quad 0x3ff8ace5422aa0db, 0x3ff8cf3216b5448c /*2^( 80 /128),2^( 81 /128)*/ + .quad 0x3ff8f1ae99157736, 0x3ff9145b0b91ffc6 /*2^( 82 /128),2^( 83 /128)*/ + .quad 0x3ff93737b0cdc5e5, 0x3ff95a44cbc8520f /*2^( 84 /128),2^( 85 /128)*/ + .quad 0x3ff97d829fde4e50, 0x3ff9a0f170ca07ba /*2^( 86 /128),2^( 87 /128)*/ + .quad 0x3ff9c49182a3f090, 0x3ff9e86319e32323 /*2^( 88 /128),2^( 89 /128)*/ + .quad 0x3ffa0c667b5de565, 0x3ffa309bec4a2d33 /*2^( 90 /128),2^( 91 /128)*/ + .quad 0x3ffa5503b23e255d, 0x3ffa799e1330b358 /*2^( 92 /128),2^( 93 /128)*/ + .quad 0x3ffa9e6b5579fdbf, 0x3ffac36bbfd3f37a /*2^( 94 /128),2^( 95 /128)*/ + .quad 0x3ffae89f995ad3ad, 0x3ffb0e07298db666 /*2^( 96 /128),2^( 97 /128)*/ + .quad 0x3ffb33a2b84f15fb, 0x3ffb59728de5593a /*2^( 98 /128),2^( 99 /128)*/ + .quad 0x3ffb7f76f2fb5e47, 0x3ffba5b030a1064a /*2^( 100 /128),2^( 101 /128)*/ + .quad 0x3ffbcc1e904bc1d2, 0x3ffbf2c25bd71e09 /*2^( 102 /128),2^( 103 /128)*/ + .quad 0x3ffc199bdd85529c, 0x3ffc40ab5fffd07a /*2^( 104 /128),2^( 105 /128)*/ + .quad 0x3ffc67f12e57d14b, 0x3ffc8f6d9406e7b5 /*2^( 106 /128),2^( 107 /128)*/ + .quad 0x3ffcb720dcef9069, 0x3ffcdf0b555dc3fa /*2^( 108 /128),2^( 109 /128)*/ + .quad 0x3ffd072d4a07897c, 0x3ffd2f87080d89f2 /*2^( 110 /128),2^( 111 /128)*/ + .quad 0x3ffd5818dcfba487, 0x3ffd80e316c98398 /*2^( 112 /128),2^( 113 /128)*/ + .quad 0x3ffda9e603db3285, 0x3ffdd321f301b460 /*2^( 114 /128),2^( 115 /128)*/ + .quad 0x3ffdfc97337b9b5f, 0x3ffe264614f5a129 /*2^( 116 /128),2^( 117 /128)*/ + .quad 0x3ffe502ee78b3ff6, 0x3ffe7a51fbc74c83 /*2^( 118 /128),2^( 119 /128)*/ + .quad 0x3ffea4afa2a490da, 0x3ffecf482d8e67f1 /*2^( 120 /128),2^( 121 /128)*/ + .quad 0x3ffefa1bee615a27, 0x3fff252b376bba97 /*2^( 122 /128),2^( 123 /128)*/ + .quad 0x3fff50765b6e4540, 0x3fff7bfdad9cbe14 /*2^( 124 /128),2^( 125 /128)*/ + .quad 0x3fffa7c1819e90d8, 0x3fffd3c22b8f71f1 /*2^( 126 /128),2^( 127 /128)*/ + .align 32 + .quad 0x42c8000000000000, 0x42c8000000000000, 0x42c8000000000000, 0x42c8000000000000 /* _dbShifter - 0x433-7=0x42c shifted right on K!*/ + //log2(relerr) = -53.547756365162 + .align 32 + .quad 0x3fe62e42fefa3685, 0x3fe62e42fefa3685, 0x3fe62e42fefa3685, 0x3fe62e42fefa3685 /* _dPC1 */ + .align 32 + .quad 0x3fcebfbdff82ca48, 0x3fcebfbdff82ca48, 0x3fcebfbdff82ca48, 0x3fcebfbdff82ca48 /* _dPC2 */ + .align 32 + .quad 0x3fac6b09b180f045, 0x3fac6b09b180f045, 0x3fac6b09b180f045, 0x3fac6b09b180f045 /* _dPC3 */ + .align 32 + .quad 0x3f83b2ab5bb1268f, 0x3f83b2ab5bb1268f, 0x3f83b2ab5bb1268f, 0x3f83b2ab5bb1268f /* _dPC4 */ + .align 32 + .quad 0x000000000000007f, 0x000000000000007f, 0x000000000000007f, 0x000000000000007f /* _lIndexMask =(2^K-1)*/ + .align 32 + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* _iAbsMask */ + .align 32 + .long 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff, 0x408fefff /* _iDomainRange */ + .align 32 + .type __svml_dexp2_data_internal,@object + .size __svml_dexp2_data_internal,.-__svml_dexp2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core-avx2.S new file mode 100644 index 0000000000..c9c17f0aaa --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core-avx2.S @@ -0,0 +1,20 @@ +/* AVX2 version of vectorized exp2, vector length is 8. + 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/>. */ + +#define _ZGVeN8v_exp2 _ZGVeN8v_exp2_avx2_wrapper +#include "../svml_d_exp28_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core.c new file mode 100644 index 0000000000..3be9e88e98 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core.c @@ -0,0 +1,27 @@ +/* Multiple versions of vectorized exp2, vector length is 8. + 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/>. */ + +#define SYMBOL_NAME _ZGVeN8v_exp2 +#include "ifunc-mathvec-avx512-skx.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVeN8v_exp2, __GI__ZGVeN8v_exp2, __redirect__ZGVeN8v_exp2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core_avx512.S new file mode 100644 index 0000000000..90f21695f0 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_exp28_core_avx512.S @@ -0,0 +1,301 @@ +/* Function exp2 vectorized with AVX-512. + 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: + * + * Double precision mantissa represented as: 1.b1b2b3 ... b52 + * Constant for double precision: S = 2^48 x 1.5 + * + * 2^X = 2^Xo x 2^{X-Xo} + * 2^X = 2^K x 2^fo x 2^{X-Xo} + * 2^X = 2^K x 2^fo x 2^r + * + * 2^K --> Manual scaling + * 2^fo --> Table lookup + * r --> 1 + poly (r = X - Xo) + * + * Xo = K + fo + * Xo = K + 0.x1x2x3x4 + * + * r = X - Xo + * = Vreduce(X, imm) + * = X - VRndScale(X, imm), where Xo = VRndScale(X, imm) + * + * Rnd(S + X) = S + Xo, where S is selected as S = 2^19 x 1.5 + * S + X = S + floor(X) + 0.x1x2x3x4 + * Rnd(S + X) = Rnd(2^48 x 1.5 + X) + * (Note: 2^exp x 1.b1b2b3 ... b52, 2^{exp-52} = 2^-4 for exp=48) + * + * exp2(x) = 2^K x 2^fo x (1 + poly(r)), where 2^r = 1 + poly(r) + * + * Scale back: + * dest = src1 x 2^floor(src2) + * + * + */ + +/* Offsets for data table __svml_dexp2_data_internal_avx512 + */ +#define Frac_PowerD0 0 +#define poly_coeff1 128 +#define poly_coeff2 192 +#define poly_coeff3 256 +#define poly_coeff4 320 +#define poly_coeff5 384 +#define poly_coeff6 448 +#define add_const 512 +#define AbsMask 576 +#define Threshold 640 +#define _lIndexMask 704 + +#include <sysdep.h> + + .text + .section .text.evex512,"ax",@progbits +ENTRY(_ZGVeN8v_exp2_skx) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-64, %rsp + subq $192, %rsp + vmovups poly_coeff5+__svml_dexp2_data_internal_avx512(%rip), %zmm14 + vmovups poly_coeff6+__svml_dexp2_data_internal_avx512(%rip), %zmm6 + +/* + * Reduced argument + * where VREDUCE is available + */ + vreducepd $65, {sae}, %zmm0, %zmm10 + vmovups poly_coeff4+__svml_dexp2_data_internal_avx512(%rip), %zmm7 + vmovups add_const+__svml_dexp2_data_internal_avx512(%rip), %zmm3 + vmovups poly_coeff3+__svml_dexp2_data_internal_avx512(%rip), %zmm8 + vmovups __svml_dexp2_data_internal_avx512(%rip), %zmm13 + +/* c6*r + c5 */ + vfmadd231pd {rn-sae}, %zmm10, %zmm6, %zmm14 + vmovups poly_coeff2+__svml_dexp2_data_internal_avx512(%rip), %zmm9 + vmovups Threshold+__svml_dexp2_data_internal_avx512(%rip), %zmm2 + +/* + * + * HA + * Variables and constants + * Load constants and vector(s) + */ + vmovups poly_coeff1+__svml_dexp2_data_internal_avx512(%rip), %zmm11 + +/* c6*r^2 + c5*r + c4 */ + vfmadd213pd {rn-sae}, %zmm7, %zmm10, %zmm14 + +/* + * Integer form of K+0.b1b2b3b4 in lower bits - call K_plus_f0 + * Mantisssa of normalized double precision FP: 1.b1b2...b52 + */ + vaddpd {rd-sae}, %zmm3, %zmm0, %zmm4 + vandpd AbsMask+__svml_dexp2_data_internal_avx512(%rip), %zmm0, %zmm1 + +/* c6*r^3 + c5*r^2 + c4*r + c3 */ + vfmadd213pd {rn-sae}, %zmm8, %zmm10, %zmm14 + vcmppd $29, {sae}, %zmm2, %zmm1, %k0 + +/* c6*r^4 + c5*r^3 + c4*r^2 + c3*r + c2 */ + vfmadd213pd {rn-sae}, %zmm9, %zmm10, %zmm14 + kmovw %k0, %edx + +/* c6*r^5 + c5*r^4 + c4*r^3 + c3*r^2 + c2*r + c1 */ + vfmadd213pd {rn-sae}, %zmm11, %zmm10, %zmm14 + +/* Table value: 2^(0.b1b2b3b4) */ + vpandq _lIndexMask+__svml_dexp2_data_internal_avx512(%rip), %zmm4, %zmm5 + vpermt2pd Frac_PowerD0+64+__svml_dexp2_data_internal_avx512(%rip), %zmm5, %zmm13 + +/* T*r */ + vmulpd {rn-sae}, %zmm10, %zmm13, %zmm12 + +/* T + (T*r*(c6*r^5 + c5*r^4 + c4*r^3 + c3*r^2 + c2*r + c1)) */ + vfmadd213pd {rn-sae}, %zmm13, %zmm12, %zmm14 + +/* Scaling placed at the end to avoid accuracy loss when T*r*scale underflows */ + vscalefpd {rn-sae}, %zmm0, %zmm14, %zmm1 + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %zmm1, %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 %zmm1, 128(%rsp) + # LOE rbx r12 r13 r14 r15 edx zmm1 + + 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: -176; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 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: -184; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 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: -192; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 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 128(%rsp), %zmm1 + +/* 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: -176; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 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: -184; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 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: -192; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 zmm1 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movsd 64(%rsp,%r14,8), %xmm0 + call exp2@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movsd %xmm0, 128(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx r15 r12d r13d +END(_ZGVeN8v_exp2_skx) + + .section .rodata, "a" + .align 64 + +#ifdef __svml_dexp2_data_internal_avx512_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(64)) VUINT32 Frac_PowerD0[16][2]; + __declspec(align(64)) VUINT32 poly_coeff1[8][2]; + __declspec(align(64)) VUINT32 poly_coeff2[8][2]; + __declspec(align(64)) VUINT32 poly_coeff3[8][2]; + __declspec(align(64)) VUINT32 poly_coeff4[8][2]; + __declspec(align(64)) VUINT32 poly_coeff5[8][2]; + __declspec(align(64)) VUINT32 poly_coeff6[8][2]; + __declspec(align(64)) VUINT32 add_const[8][2]; + __declspec(align(64)) VUINT32 AbsMask[8][2]; + __declspec(align(64)) VUINT32 Threshold[8][2]; + __declspec(align(64)) VUINT32 _lIndexMask[8][2]; +} __svml_dexp2_data_internal_avx512; +#endif +__svml_dexp2_data_internal_avx512: + /*== Frac_PowerD0 ==*/ + .quad 0x3FF0000000000000 + .quad 0x3FF0B5586CF9890F + .quad 0x3FF172B83C7D517B + .quad 0x3FF2387A6E756238 + .quad 0x3FF306FE0A31B715 + .quad 0x3FF3DEA64C123422 + .quad 0x3FF4BFDAD5362A27 + .quad 0x3FF5AB07DD485429 + .quad 0x3FF6A09E667F3BCD + .quad 0x3FF7A11473EB0187 + .quad 0x3FF8ACE5422AA0DB + .quad 0x3FF9C49182A3F090 + .quad 0x3FFAE89F995AD3AD + .quad 0x3FFC199BDD85529C + .quad 0x3FFD5818DCFBA487 + .quad 0x3FFEA4AFA2A490DA + .align 64 + .quad 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B, 0x3FE62E42FEFA398B /*== poly_coeff1 ==*/ + .align 64 + .quad 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A, 0x3FCEBFBDFF84555A /*== poly_coeff2 ==*/ + .align 64 + .quad 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9, 0x3FAC6B08D4AD86B9 /*== poly_coeff3 ==*/ + .align 64 + .quad 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252, 0x3F83B2AD1B172252 /*== poly_coeff4 ==*/ + .align 64 + .quad 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19, 0x3F55D7472713CD19 /*== poly_coeff5 ==*/ + .align 64 + .quad 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B, 0x3F24A1D7F526371B /*== poly_coeff6 ==*/ + .align 64 + .quad 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000, 0x42F8000000000000 /* add_const */ + .align 64 + .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff /* AbsMask */ + .align 64 + .quad 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000, 0x408fefff00000000 /* Threshold */ + .align 64 + .quad 0x000000000000000F, 0x000000000000000F, 0x000000000000000F, 0x000000000000000F, 0x000000000000000F, 0x000000000000000F, 0x000000000000000F, 0x000000000000000F /* _lIndexMask */ + .align 64 + .type __svml_dexp2_data_internal_avx512,@object + .size __svml_dexp2_data_internal_avx512,.-__svml_dexp2_data_internal_avx512 diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core-avx2.S new file mode 100644 index 0000000000..4daa687852 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core-avx2.S @@ -0,0 +1,20 @@ +/* AVX2 version of vectorized exp2f. + 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/>. */ + +#define _ZGVeN16v_exp2f _ZGVeN16v_exp2f_avx2_wrapper +#include "../svml_s_exp2f16_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core.c new file mode 100644 index 0000000000..e90d9d8684 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized exp2f, vector length is 16. + 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/>. */ + +#define SYMBOL_NAME _ZGVeN16v_exp2f +#include "ifunc-mathvec-avx512-skx.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVeN16v_exp2f, __GI__ZGVeN16v_exp2f, + __redirect__ZGVeN16v_exp2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core_avx512.S new file mode 100644 index 0000000000..6b512159bc --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f16_core_avx512.S @@ -0,0 +1,271 @@ +/* Function exp2f vectorized with AVX-512. + 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: + * + * Single precision mantissa represented as: 1.b1b2b3 ... b23 + * Constant for single precision: S = 2^19 x 1.5 + * + * 2^X = 2^Xo x 2^{X-Xo} + * 2^X = 2^K x 2^fo x 2^{X-Xo} + * 2^X = 2^K x 2^fo x 2^r + * + * 2^K --> Manual scaling + * 2^fo --> Table lookup + * r --> 1 + poly (r = X - Xo) + * + * Xo = K + fo + * Xo = K + 0.x1x2x3x4 + * + * r = X - Xo + * = Vreduce(X, imm) + * = X - VRndScale(X, imm), where Xo = VRndScale(X, imm) + * + * Rnd(S + X) = S + Xo, where S is selected as S = 2^19 x 1.5 + * S + X = S + floor(X) + 0.x1x2x3x4 + * Rnd(S + X) = Rnd(2^19 x 1.5 + X) + * (Note: 2^exp x 1.b1b2b3 ... b23, 2^{exp-23} = 2^-4 for exp=19) + * + * exp2(x) = 2^K x 2^fo x (1 + poly(r)), where 2^r = 1 + poly(r) + * + * Scale back: + * dest = src1 x 2^floor(src2) + * + * + */ + +/* Offsets for data table __svml_sexp2_data_internal_avx512 + */ +#define Frac_PowerS0 0 +#define poly_coeff1 64 +#define poly_coeff2 128 +#define poly_coeff3 192 +#define add_const 256 +#define AbsMask 320 +#define Threshold 384 + +#include <sysdep.h> + + .text + .section .text.exex512,"ax",@progbits +ENTRY(_ZGVeN16v_exp2f_skx) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-64, %rsp + subq $192, %rsp + vmovups add_const+__svml_sexp2_data_internal_avx512(%rip), %zmm3 + +/* + * Reduced argument + * where VREDUCE is available + */ + vreduceps $65, {sae}, %zmm0, %zmm6 + vmovups poly_coeff3+__svml_sexp2_data_internal_avx512(%rip), %zmm5 + vmovups poly_coeff2+__svml_sexp2_data_internal_avx512(%rip), %zmm10 + vmovups Threshold+__svml_sexp2_data_internal_avx512(%rip), %zmm2 + +/* + * + * HA + * Variables and constants + * Load constants and vector(s) + */ + vmovups poly_coeff1+__svml_sexp2_data_internal_avx512(%rip), %zmm7 + +/* + * Integer form of K+0.b1b2b3b4 in lower bits - call K_plus_f0 + * Mantisssa of normalized single precision FP: 1.b1b2...b23 + */ + vaddps {rd-sae}, %zmm3, %zmm0, %zmm4 + vandps AbsMask+__svml_sexp2_data_internal_avx512(%rip), %zmm0, %zmm1 + +/* c3*r + c2 */ + vfmadd231ps {rn-sae}, %zmm6, %zmm5, %zmm10 + vcmpps $30, {sae}, %zmm2, %zmm1, %k0 + +/* c3*r^2 + c2*r + c1 */ + vfmadd213ps {rn-sae}, %zmm7, %zmm6, %zmm10 + +/* Table value: 2^(0.b1b2b3b4) */ + vpermps __svml_sexp2_data_internal_avx512(%rip), %zmm4, %zmm9 + kmovw %k0, %edx + +/* T*r */ + vmulps {rn-sae}, %zmm6, %zmm9, %zmm8 + +/* T + (T*r*(c3*r^2 + c2*r + c1) */ + vfmadd213ps {rn-sae}, %zmm9, %zmm8, %zmm10 + +/* Scaling placed at the end to avoid accuracy loss when T*r*scale underflows */ + vscalefps {rn-sae}, %zmm0, %zmm10, %zmm1 + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %zmm1, %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 %zmm1, 128(%rsp) + # LOE rbx r12 r13 r14 r15 edx zmm1 + + 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: -176; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 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: -184; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 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: -192; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 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 $16, %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 128(%rsp), %zmm1 + +/* 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: -176; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x50, 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: -184; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x48, 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: -192; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x40, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 zmm1 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 64(%rsp,%r14,4), %xmm0 + call exp2f@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movss %xmm0, 128(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx r15 r12d r13d +END(_ZGVeN16v_exp2f_skx) + + .section .rodata, "a" + .align 64 + +#ifdef __svml_sexp2_data_internal_avx512_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(64)) VUINT32 Frac_PowerS0[16][1]; + __declspec(align(64)) VUINT32 poly_coeff1[16][1]; + __declspec(align(64)) VUINT32 poly_coeff2[16][1]; + __declspec(align(64)) VUINT32 poly_coeff3[16][1]; + __declspec(align(64)) VUINT32 add_const[16][1]; + __declspec(align(64)) VUINT32 AbsMask[16][1]; + __declspec(align(64)) VUINT32 Threshold[16][1]; +} __svml_sexp2_data_internal_avx512; +#endif +__svml_sexp2_data_internal_avx512: + /*== Frac_PowerS0 ==*/ + .long 0x3F800000 + .long 0x3F85AAC3 + .long 0x3F8B95C2 + .long 0x3F91C3D3 + .long 0x3F9837F0 + .long 0x3F9EF532 + .long 0x3FA5FED7 + .long 0x3FAD583F + .long 0x3FB504F3 + .long 0x3FBD08A4 + .long 0x3FC5672A + .long 0x3FCE248C + .long 0x3FD744FD + .long 0x3FE0CCDF + .long 0x3FEAC0C7 + .long 0x3FF5257D + .align 64 + .long 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222, 0x3F317222 /*== poly_coeff1 ==*/ + .align 64 + .long 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B, 0x3E75F16B /*== poly_coeff2 ==*/ + .align 64 + .long 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA, 0x3D6854CA /*== poly_coeff3 ==*/ + .align 64 + .long 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000, 0x49400000 /* add_const */ + .align 64 + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* AbsMask */ + .align 64 + .long 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000 /* Threshold=126.0 */ + .align 64 + .type __svml_sexp2_data_internal_avx512,@object + .size __svml_sexp2_data_internal_avx512,.-__svml_sexp2_data_internal_avx512 diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core-sse2.S new file mode 100644 index 0000000000..0b3fec834c --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core-sse2.S @@ -0,0 +1,20 @@ +/* SSE2 version of vectorized exp2f, vector length is 4. + 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/>. */ + +#define _ZGVbN4v_exp2f _ZGVbN4v_exp2f_sse2 +#include "../svml_s_exp2f4_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core.c new file mode 100644 index 0000000000..db47118d97 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized exp2f, vector length is 4. + 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/>. */ + +#define SYMBOL_NAME _ZGVbN4v_exp2f +#include "ifunc-mathvec-sse4_1.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVbN4v_exp2f, __GI__ZGVbN4v_exp2f, + __redirect__ZGVbN4v_exp2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core_sse4.S new file mode 100644 index 0000000000..0d9f45d5c3 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f4_core_sse4.S @@ -0,0 +1,238 @@ +/* Function exp2f vectorized with SSE4. + 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: + * + * exp2(x) = 2^n * T[j] * (1 + P(y)) + * where + * x = m*(1/K) + y, y in [-1/K..1/K] + * m = n*K + j, m,n,j - signed integer, j in [-K/2..K/2] + * + * values of 2^j/K are tabulated + * + * P(y) is a minimax polynomial approximation of exp2(x)-1 + * on small interval [-1/K..1/K] + * + * Special cases: + * + * exp2(NaN) = NaN + * exp2(+INF) = +INF + * exp2(-INF) = 0 + * exp2(x) = 1 for subnormals + * For IEEE float + * if x >= 128.0 then exp2f(x) overflow + * if x < -151.0 then exp2f(x) underflow + * + */ + +/* Offsets for data table __svml_sexp2_data_internal + */ +#define _sShifter 0 +#define _sPC0 16 +#define _sPC1 32 +#define _sPC2 48 +#define _sPC3 64 +#define _sPC4 80 +#define _sPC5 96 +#define _sPC6 112 +#define _iAbsMask 128 +#define _iDomainRange 144 + +#include <sysdep.h> + + .text + .section .text.sse4,"ax",@progbits +ENTRY(_ZGVbN4v_exp2f_sse4) + subq $72, %rsp + cfi_def_cfa_offset(80) + +/* Check for overflow\underflow */ + movups __svml_sexp2_data_internal(%rip), %xmm1 + +/* Implementation */ + movaps %xmm1, %xmm5 + +/* Polynomial */ + movups _sPC6+__svml_sexp2_data_internal(%rip), %xmm4 + addps %xmm0, %xmm5 + movaps %xmm5, %xmm3 + +/* 2^N */ + pslld $23, %xmm5 + +/* Check for overflow\underflow */ + movdqu _iAbsMask+__svml_sexp2_data_internal(%rip), %xmm2 + subps %xmm1, %xmm3 + +/* R */ + movaps %xmm0, %xmm1 + pand %xmm0, %xmm2 + pcmpgtd _iDomainRange+__svml_sexp2_data_internal(%rip), %xmm2 + subps %xmm3, %xmm1 + movmskps %xmm2, %edx + mulps %xmm1, %xmm4 + addps _sPC5+__svml_sexp2_data_internal(%rip), %xmm4 + mulps %xmm1, %xmm4 + addps _sPC4+__svml_sexp2_data_internal(%rip), %xmm4 + mulps %xmm1, %xmm4 + addps _sPC3+__svml_sexp2_data_internal(%rip), %xmm4 + mulps %xmm1, %xmm4 + addps _sPC2+__svml_sexp2_data_internal(%rip), %xmm4 + mulps %xmm1, %xmm4 + addps _sPC1+__svml_sexp2_data_internal(%rip), %xmm4 + mulps %xmm4, %xmm1 + addps _sPC0+__svml_sexp2_data_internal(%rip), %xmm1 + +/* Reconstruction */ + paddd %xmm5, %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 + */ + +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 + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx rbp r15 r12d r13d + +/* 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 + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 32(%rsp,%r14,4), %xmm0 + call exp2f@PLT + # LOE rbx rbp r14 r15 r12d r13d xmm0 + + movss %xmm0, 48(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + # LOE rbx rbp r15 r12d r13d +END(_ZGVbN4v_exp2f_sse4) + + .section .rodata, "a" + .align 16 + +#ifdef __svml_sexp2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct +{ + __declspec(align(16)) VUINT32 _sShifter[4][1]; + __declspec(align(16)) VUINT32 _sPC0[4][1]; + __declspec(align(16)) VUINT32 _sPC1[4][1]; + __declspec(align(16)) VUINT32 _sPC2[4][1]; + __declspec(align(16)) VUINT32 _sPC3[4][1]; + __declspec(align(16)) VUINT32 _sPC4[4][1]; + __declspec(align(16)) VUINT32 _sPC5[4][1]; + __declspec(align(16)) VUINT32 _sPC6[4][1]; + __declspec(align(16)) VUINT32 _iAbsMask[4][1]; + __declspec(align(16)) VUINT32 _iDomainRange[4][1]; +} __svml_sexp2_data_internal; +#endif +__svml_sexp2_data_internal: + .long 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000 /* _sShifter */ + .align 16 + .long 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000 /* _sPC0 */ + .align 16 + .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218 /* _sPC1 */ + .align 16 + .long 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef /* _sPC2 */ + .align 16 + .long 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf /* _sPC3 */ + .align 16 + .long 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c /* _sPC4 */ + .align 16 + .long 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51 /* _sPC5 */ + .align 16 + .long 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c /* _sPC6 */ + //common + .align 16 + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* _iAbsMask */ + .align 16 + .long 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000 /* _iDomainRange=126.0 */ + .align 16 + .type __svml_sexp2_data_internal,@object + .size __svml_sexp2_data_internal,.-__svml_sexp2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core-sse.S new file mode 100644 index 0000000000..4da2278ed8 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core-sse.S @@ -0,0 +1,20 @@ +/* SSE version of vectorized exp2f, vector length is 8. + 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/>. */ + +#define _ZGVdN8v_exp2f _ZGVdN8v_exp2f_sse_wrapper +#include "../svml_s_exp2f8_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core.c new file mode 100644 index 0000000000..dc34671263 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized exp2f, vector length is 8. + 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/>. */ + +#define SYMBOL_NAME _ZGVdN8v_exp2f +#include "ifunc-mathvec-avx2.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVdN8v_exp2f, __GI__ZGVdN8v_exp2f, + __redirect__ZGVdN8v_exp2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core_avx2.S new file mode 100644 index 0000000000..aa7af4be79 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_exp2f8_core_avx2.S @@ -0,0 +1,245 @@ +/* Function exp2f 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: + * + * exp2(x) = 2^n * T[j] * (1 + P(y)) + * where + * x = m*(1/K) + y, y in [-1/K..1/K] + * m = n*K + j, m,n,j - signed integer, j in [-K/2..K/2] + * + * values of 2^j/K are tabulated + * + * P(y) is a minimax polynomial approximation of exp2(x)-1 + * on small interval [-1/K..1/K] + * + * Special cases: + * + * exp2(NaN) = NaN + * exp2(+INF) = +INF + * exp2(-INF) = 0 + * exp2(x) = 1 for subnormals + * For IEEE float + * if x >= 128.0 then exp2f(x) overflow + * if x < -151.0 then exp2f(x) underflow + * + */ + +/* Offsets for data table __svml_sexp2_data_internal + */ +#define _sShifter 0 +#define _sPC0 32 +#define _sPC1 64 +#define _sPC2 96 +#define _sPC3 128 +#define _sPC4 160 +#define _sPC5 192 +#define _sPC6 224 +#define _iAbsMask 256 +#define _iDomainRange 288 + +#include <sysdep.h> + + .text + .section .text.avx2,"ax",@progbits +ENTRY(_ZGVdN8v_exp2f_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_sexp2_data_internal(%rip), %ymm1 + +/* Check for overflow\underflow */ + vmovups _sPC6+__svml_sexp2_data_internal(%rip), %ymm7 + +/* Implementation */ + vaddps %ymm1, %ymm0, %ymm6 + vsubps %ymm1, %ymm6, %ymm4 + +/* 2^N */ + vpslld $23, %ymm6, %ymm8 + +/* R */ + vsubps %ymm4, %ymm0, %ymm5 + +/* Polynomial */ + vfmadd213ps _sPC5+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + vfmadd213ps _sPC4+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + vfmadd213ps _sPC3+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + vfmadd213ps _sPC2+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + vfmadd213ps _sPC1+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + vfmadd213ps _sPC0+__svml_sexp2_data_internal(%rip), %ymm5, %ymm7 + +/* Check for overflow\underflow */ + vandps _iAbsMask+__svml_sexp2_data_internal(%rip), %ymm0, %ymm2 + vpcmpgtd _iDomainRange+__svml_sexp2_data_internal(%rip), %ymm2, %ymm3 + vmovmskps %ymm3, %edx + +/* Reconstruction */ + vpaddd %ymm8, %ymm7, %ymm1 + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm1 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %ymm1, %ymm0 + 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 %ymm0, 32(%rsp) + vmovups %ymm1, 64(%rsp) + # LOE rbx r12 r13 r14 r15 edx ymm1 + + 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), %ymm1 + +/* 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 ymm1 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 32(%rsp,%r14,4), %xmm0 + call exp2f@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_exp2f_avx2) + + .section .rodata, "a" + .align 32 + +#ifdef __svml_sexp2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct +{ + __declspec(align(32)) VUINT32 _sShifter[8][1]; + __declspec(align(32)) VUINT32 _sPC0[8][1]; + __declspec(align(32)) VUINT32 _sPC1[8][1]; + __declspec(align(32)) VUINT32 _sPC2[8][1]; + __declspec(align(32)) VUINT32 _sPC3[8][1]; + __declspec(align(32)) VUINT32 _sPC4[8][1]; + __declspec(align(32)) VUINT32 _sPC5[8][1]; + __declspec(align(32)) VUINT32 _sPC6[8][1]; + __declspec(align(32)) VUINT32 _iAbsMask[8][1]; + __declspec(align(32)) VUINT32 _iDomainRange[8][1]; +} __svml_sexp2_data_internal; +#endif +__svml_sexp2_data_internal: + .long 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000, 0x4b400000 /* _sShifter */ + .align 32 + .long 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000 /* _sPC0 */ + .align 32 + .long 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218, 0x3f317218 /* _sPC1 */ + .align 32 + .long 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef, 0x3e75fdef /* _sPC2 */ + .align 32 + .long 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf, 0x3d6357cf /* _sPC3 */ + .align 32 + .long 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c, 0x3c1d962c /* _sPC4 */ + .align 32 + .long 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51, 0x3aaf7a51 /* _sPC5 */ + .align 32 + .long 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c, 0x39213c8c /* _sPC6 */ + //common + .align 32 + .long 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff /* _iAbsMask */ + .align 32 + .long 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000, 0x42fc0000 /* _iDomainRange=126.0 */ + .align 32 + .type __svml_sexp2_data_internal,@object + .size __svml_sexp2_data_internal,.-__svml_sexp2_data_internal diff --git a/sysdeps/x86_64/fpu/svml_d_exp22_core.S b/sysdeps/x86_64/fpu/svml_d_exp22_core.S new file mode 100644 index 0000000000..f03080a977 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_exp22_core.S @@ -0,0 +1,29 @@ +/* Function exp2 vectorized with SSE2. + 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/>. */ + +#include <sysdep.h> +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVbN2v_exp2) +WRAPPER_IMPL_SSE2 exp2 +END (_ZGVbN2v_exp2) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVbN2v_exp2) +#endif diff --git a/sysdeps/x86_64/fpu/svml_d_exp24_core.S b/sysdeps/x86_64/fpu/svml_d_exp24_core.S new file mode 100644 index 0000000000..40475c7a94 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_exp24_core.S @@ -0,0 +1,29 @@ +/* Function exp2 vectorized with AVX2, wrapper version. + 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/>. */ + +#include <sysdep.h> +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVdN4v_exp2) +WRAPPER_IMPL_AVX _ZGVbN2v_exp2 +END (_ZGVdN4v_exp2) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVdN4v_exp2) +#endif diff --git a/sysdeps/x86_64/fpu/svml_d_exp24_core_avx.S b/sysdeps/x86_64/fpu/svml_d_exp24_core_avx.S new file mode 100644 index 0000000000..a7d22409df --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_exp24_core_avx.S @@ -0,0 +1,25 @@ +/* Function exp2 vectorized in AVX ISA as wrapper to SSE4 ISA version. + 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/>. */ + +#include <sysdep.h> +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVcN4v_exp2) +WRAPPER_IMPL_AVX _ZGVbN2v_exp2 +END (_ZGVcN4v_exp2) diff --git a/sysdeps/x86_64/fpu/svml_d_exp28_core.S b/sysdeps/x86_64/fpu/svml_d_exp28_core.S new file mode 100644 index 0000000000..f68aaed427 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_exp28_core.S @@ -0,0 +1,25 @@ +/* Function exp2 vectorized with AVX-512, wrapper to 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/>. */ + +#include <sysdep.h> +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVeN8v_exp2) +WRAPPER_IMPL_AVX512 _ZGVdN4v_exp2 +END (_ZGVeN8v_exp2) diff --git a/sysdeps/x86_64/fpu/svml_s_exp2f16_core.S b/sysdeps/x86_64/fpu/svml_s_exp2f16_core.S new file mode 100644 index 0000000000..8ba4e82272 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_exp2f16_core.S @@ -0,0 +1,25 @@ +/* Function exp2f vectorized with AVX-512. Wrapper to AVX2 version. + 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/>. */ + +#include <sysdep.h> +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVeN16v_exp2f) +WRAPPER_IMPL_AVX512 _ZGVdN8v_exp2f +END (_ZGVeN16v_exp2f) diff --git a/sysdeps/x86_64/fpu/svml_s_exp2f4_core.S b/sysdeps/x86_64/fpu/svml_s_exp2f4_core.S new file mode 100644 index 0000000000..916f176dca --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_exp2f4_core.S @@ -0,0 +1,29 @@ +/* Function exp2f vectorized with SSE2, wrapper version. + 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/>. */ + +#include <sysdep.h> +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVbN4v_exp2f) +WRAPPER_IMPL_SSE2 exp2f +END (_ZGVbN4v_exp2f) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVbN4v_exp2f) +#endif diff --git a/sysdeps/x86_64/fpu/svml_s_exp2f8_core.S b/sysdeps/x86_64/fpu/svml_s_exp2f8_core.S new file mode 100644 index 0000000000..b8821b952b --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_exp2f8_core.S @@ -0,0 +1,29 @@ +/* Function exp2f vectorized with AVX2, wrapper version. + 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/>. */ + +#include <sysdep.h> +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVdN8v_exp2f) +WRAPPER_IMPL_AVX _ZGVbN4v_exp2f +END (_ZGVdN8v_exp2f) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVdN8v_exp2f) +#endif diff --git a/sysdeps/x86_64/fpu/svml_s_exp2f8_core_avx.S b/sysdeps/x86_64/fpu/svml_s_exp2f8_core_avx.S new file mode 100644 index 0000000000..ddaaf3b59a --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_exp2f8_core_avx.S @@ -0,0 +1,25 @@ +/* Function exp2f vectorized in AVX ISA as wrapper to SSE4 ISA version. + 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/>. */ + +#include <sysdep.h> +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVcN8v_exp2f) +WRAPPER_IMPL_AVX _ZGVbN4v_exp2f +END (_ZGVcN8v_exp2f) diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx.c new file mode 100644 index 0000000000..341ec99724 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx.c @@ -0,0 +1 @@ +#include "test-double-libmvec-exp2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx2.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx2.c new file mode 100644 index 0000000000..341ec99724 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx2.c @@ -0,0 +1 @@ +#include "test-double-libmvec-exp2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx512f.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx512f.c new file mode 100644 index 0000000000..341ec99724 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp2-avx512f.c @@ -0,0 +1 @@ +#include "test-double-libmvec-exp2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-exp2.c b/sysdeps/x86_64/fpu/test-double-libmvec-exp2.c new file mode 100644 index 0000000000..b3b04f63e4 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-exp2.c @@ -0,0 +1,3 @@ +#define LIBMVEC_TYPE double +#define LIBMVEC_FUNC exp2 +#include "test-vector-abi-arg1.h" diff --git a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c index 9bc9d1dafa..2f7172bd7b 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acos), _ZGVbN2v_acos) VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVbN2v_atan) VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVbN2v_asin) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVbN2vv_hypot) +VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVbN2v_exp2) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c index c41994d90a..e2d519faac 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c @@ -34,6 +34,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acos), _ZGVdN4v_acos) VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVdN4v_atan) VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVdN4v_asin) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVdN4vv_hypot) +VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVdN4v_exp2) #ifndef __ILP32__ # define VEC_INT_TYPE __m256i diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c index 881f6c801a..1ce4d8b413 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acos), _ZGVcN4v_acos) VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVcN4v_atan) VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVcN4v_asin) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVcN4vv_hypot) +VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVcN4v_exp2) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c index 6fd106fe68..6c87cec648 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acos), _ZGVeN8v_acos) VECTOR_WRAPPER (WRAPPER_NAME (atan), _ZGVeN8v_atan) VECTOR_WRAPPER (WRAPPER_NAME (asin), _ZGVeN8v_asin) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypot), _ZGVeN8vv_hypot) +VECTOR_WRAPPER (WRAPPER_NAME (exp2), _ZGVeN8v_exp2) #ifndef __ILP32__ # define VEC_INT_TYPE __m512i diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx.c new file mode 100644 index 0000000000..0281d386fb --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx.c @@ -0,0 +1 @@ +#include "test-float-libmvec-exp2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx2.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx2.c new file mode 100644 index 0000000000..0281d386fb --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx2.c @@ -0,0 +1 @@ +#include "test-float-libmvec-exp2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx512f.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx512f.c new file mode 100644 index 0000000000..0281d386fb --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f-avx512f.c @@ -0,0 +1 @@ +#include "test-float-libmvec-exp2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-exp2f.c b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f.c new file mode 100644 index 0000000000..bf57661bee --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-exp2f.c @@ -0,0 +1,3 @@ +#define LIBMVEC_TYPE float +#define LIBMVEC_FUNC exp2f +#include "test-vector-abi-arg1.h" diff --git a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c index 4c2ea6ddfe..597d7d7598 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acosf), _ZGVeN16v_acosf) VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVeN16v_atanf) VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVeN16v_asinf) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVeN16vv_hypotf) +VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVeN16v_exp2f) #define VEC_INT_TYPE __m512i diff --git a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c index 1d5d952d07..3500eec810 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acosf), _ZGVbN4v_acosf) VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVbN4v_atanf) VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVbN4v_asinf) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVbN4vv_hypotf) +VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVbN4v_exp2f) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c index 7a750f3781..921b9c65d6 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c @@ -34,6 +34,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acosf), _ZGVdN8v_acosf) VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVdN8v_atanf) VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVdN8v_asinf) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVdN8vv_hypotf) +VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVdN8v_exp2f) /* Redefinition of wrapper to be compatible with _ZGVdN8vvv_sincosf. */ #undef VECTOR_WRAPPER_fFF diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c index af816a7789..6cbcb57521 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c @@ -31,6 +31,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (acosf), _ZGVcN8v_acosf) VECTOR_WRAPPER (WRAPPER_NAME (atanf), _ZGVcN8v_atanf) VECTOR_WRAPPER (WRAPPER_NAME (asinf), _ZGVcN8v_asinf) VECTOR_WRAPPER_ff (WRAPPER_NAME (hypotf), _ZGVcN8vv_hypotf) +VECTOR_WRAPPER (WRAPPER_NAME (exp2f), _ZGVcN8v_exp2f) #define VEC_INT_TYPE __m128i |