diff options
Diffstat (limited to 'gmp-6.3.0/mpn/arm64/cora53')
| -rw-r--r-- | gmp-6.3.0/mpn/arm64/cora53/cnd_aors_n.asm | 99 | ||||
| -rw-r--r-- | gmp-6.3.0/mpn/arm64/cora53/gmp-mparam.h | 242 |
2 files changed, 341 insertions, 0 deletions
diff --git a/gmp-6.3.0/mpn/arm64/cora53/cnd_aors_n.asm b/gmp-6.3.0/mpn/arm64/cora53/cnd_aors_n.asm new file mode 100644 index 0000000..1b227da --- /dev/null +++ b/gmp-6.3.0/mpn/arm64/cora53/cnd_aors_n.asm @@ -0,0 +1,99 @@ +dnl ARM64 mpn_cnd_add_n, mpn_cnd_sub_n + +dnl Contributed to the GNU project by Torbjörn Granlund. + +dnl Copyright 2012, 2013, 2017 Free Software Foundation, Inc. + +dnl This file is part of the GNU MP Library. +dnl +dnl The GNU MP Library is free software; you can redistribute it and/or modify +dnl it under the terms of either: +dnl +dnl * the GNU Lesser General Public License as published by the Free +dnl Software Foundation; either version 3 of the License, or (at your +dnl option) any later version. +dnl +dnl or +dnl +dnl * the GNU General Public License as published by the Free Software +dnl Foundation; either version 2 of the License, or (at your option) any +dnl later version. +dnl +dnl or both in parallel, as here. +dnl +dnl The GNU MP Library is distributed in the hope that it will be useful, but +dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY +dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +dnl for more details. +dnl +dnl You should have received copies of the GNU General Public License and the +dnl GNU Lesser General Public License along with the GNU MP Library. If not, +dnl see https://www.gnu.org/licenses/. + +include(`../config.m4') + +C cycles/limb +C Cortex-A53 3.5-4 +C Cortex-A57 2.25 +C X-Gene 3.5 + +changecom(blah) + +define(`cnd', `x0') +define(`rp', `x1') +define(`up', `x2') +define(`vp', `x3') +define(`n', `x4') + +ifdef(`OPERATION_cnd_add_n', ` + define(`ADDSUBC', adcs) + define(`CLRCY', `cmn xzr, xzr') + define(`RETVAL', `cset x0, cs') + define(`func', mpn_cnd_add_n)') +ifdef(`OPERATION_cnd_sub_n', ` + define(`ADDSUBC', sbcs) + define(`CLRCY', `cmp xzr, xzr') + define(`RETVAL', `cset x0, cc') + define(`func', mpn_cnd_sub_n)') + +MULFUNC_PROLOGUE(mpn_cnd_add_n mpn_cnd_sub_n) + +ASM_START() +PROLOGUE(func) + cmp cnd, #1 + sbc cnd, cnd, cnd + + CLRCY C really only needed for n = 0 (mod 4) + + tbz n, #0, L(1) + ldr x10, [up], #8 + ldr x12, [vp], #8 + bic x6, x12, cnd + ADDSUBC x8, x10, x6 + sub n, n, #1 + str x8, [rp], #8 + cbz n, L(rt) + +L(1): ldp x10, x11, [up], #16 + ldp x12, x13, [vp], #16 + sub n, n, #2 + cbz n, L(end) + +L(top): bic x6, x12, cnd + bic x7, x13, cnd + ldp x12, x13, [vp], #16 + ADDSUBC x8, x10, x6 + ADDSUBC x9, x11, x7 + ldp x10, x11, [up], #16 + sub n, n, #2 + stp x8, x9, [rp], #16 + cbnz n, L(top) + +L(end): bic x6, x12, cnd + bic x7, x13, cnd + ADDSUBC x8, x10, x6 + ADDSUBC x9, x11, x7 + stp x8, x9, [rp] +L(rt): RETVAL + ret +EPILOGUE() diff --git a/gmp-6.3.0/mpn/arm64/cora53/gmp-mparam.h b/gmp-6.3.0/mpn/arm64/cora53/gmp-mparam.h new file mode 100644 index 0000000..f4e258d --- /dev/null +++ b/gmp-6.3.0/mpn/arm64/cora53/gmp-mparam.h @@ -0,0 +1,242 @@ +/* gmp-mparam.h -- Compiler/machine parameter header file for a53. + +Copyright 2019 Free Software Foundation, Inc. + +This file is part of the GNU MP Library. + +The GNU MP Library is free software; you can redistribute it and/or modify +it under the terms of either: + + * the GNU Lesser General Public License as published by the Free + Software Foundation; either version 3 of the License, or (at your + option) any later version. + +or + + * the GNU General Public License as published by the Free Software + Foundation; either version 2 of the License, or (at your option) any + later version. + +or both in parallel, as here. + +The GNU MP 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 General Public License +for more details. + +You should have received copies of the GNU General Public License and the +GNU Lesser General Public License along with the GNU MP Library. If not, +see https://www.gnu.org/licenses/. */ + +#define GMP_LIMB_BITS 64 +#define GMP_LIMB_BYTES 8 + +/* 1536 MHz Cortex-A53 */ +/* FFT tuning limit = 21,583,800 */ +/* Generated by tuneup.c, 2019-10-22, gcc 5.4 */ + +#define DIVREM_1_NORM_THRESHOLD 3 +#define DIVREM_1_UNNORM_THRESHOLD 4 +#define MOD_1_1P_METHOD 2 /* 4.84% faster than 1 */ +#define MOD_1_NORM_THRESHOLD 3 +#define MOD_1_UNNORM_THRESHOLD 4 +#define MOD_1N_TO_MOD_1_1_THRESHOLD 7 +#define MOD_1U_TO_MOD_1_1_THRESHOLD 7 +#define MOD_1_1_TO_MOD_1_2_THRESHOLD 12 +#define MOD_1_2_TO_MOD_1_4_THRESHOLD 18 +#define PREINV_MOD_1_TO_MOD_1_THRESHOLD 22 +#define USE_PREINV_DIVREM_1 1 +#define DIV_QR_1N_PI1_METHOD 1 /* 39.05% faster than 2 */ +#define DIV_QR_1_NORM_THRESHOLD 21 +#define DIV_QR_1_UNNORM_THRESHOLD 21 +#define DIV_QR_2_PI2_THRESHOLD MP_SIZE_T_MAX /* never */ +#define DIVEXACT_1_THRESHOLD 0 /* always */ +#define BMOD_1_TO_MOD_1_THRESHOLD 38 + +#define DIV_1_VS_MUL_1_PERCENT 161 + +#define MUL_TOOM22_THRESHOLD 14 +#define MUL_TOOM33_THRESHOLD 49 +#define MUL_TOOM44_THRESHOLD 73 +#define MUL_TOOM6H_THRESHOLD 173 +#define MUL_TOOM8H_THRESHOLD 236 + +#define MUL_TOOM32_TO_TOOM43_THRESHOLD 81 +#define MUL_TOOM32_TO_TOOM53_THRESHOLD 77 +#define MUL_TOOM42_TO_TOOM53_THRESHOLD 81 +#define MUL_TOOM42_TO_TOOM63_THRESHOLD 88 +#define MUL_TOOM43_TO_TOOM54_THRESHOLD 65 + +#define SQR_BASECASE_THRESHOLD 0 /* always */ +#define SQR_TOOM2_THRESHOLD 18 +#define SQR_TOOM3_THRESHOLD 68 +#define SQR_TOOM4_THRESHOLD 183 +#define SQR_TOOM6_THRESHOLD 230 +#define SQR_TOOM8_THRESHOLD 357 + +#define MULMID_TOOM42_THRESHOLD 23 + +#define MULMOD_BNM1_THRESHOLD 9 +#define SQRMOD_BNM1_THRESHOLD 11 + +#define MUL_FFT_MODF_THRESHOLD 316 /* k = 5 */ +#define MUL_FFT_TABLE3 \ + { { 316, 5}, { 13, 6}, { 7, 5}, { 15, 6}, \ + { 13, 7}, { 7, 6}, { 15, 7}, { 8, 6}, \ + { 17, 7}, { 9, 6}, { 19, 7}, { 17, 8}, \ + { 9, 7}, { 20, 8}, { 11, 7}, { 23, 8}, \ + { 13, 9}, { 7, 8}, { 19, 9}, { 11, 8}, \ + { 27, 9}, { 15, 8}, { 33, 9}, { 19, 8}, \ + { 41, 9}, { 23, 8}, { 49, 9}, { 27,10}, \ + { 15, 9}, { 39,10}, { 23, 9}, { 51,11}, \ + { 15,10}, { 31, 9}, { 71,10}, { 39, 9}, \ + { 83,10}, { 47, 9}, { 99,10}, { 55,11}, \ + { 31,10}, { 63, 8}, { 255,10}, { 71, 8}, \ + { 287,10}, { 79, 9}, { 159, 8}, { 319,10}, \ + { 87,11}, { 47,10}, { 95, 9}, { 191, 8}, \ + { 383,10}, { 103, 9}, { 207, 8}, { 415,10}, \ + { 111, 9}, { 223,12}, { 31,11}, { 63, 9}, \ + { 255, 8}, { 511,10}, { 135, 9}, { 287, 8}, \ + { 575,11}, { 79,10}, { 159, 9}, { 319, 8}, \ + { 639,10}, { 175, 9}, { 351, 8}, { 703,11}, \ + { 95,10}, { 191, 9}, { 383, 8}, { 767,10}, \ + { 207, 9}, { 415, 8}, { 831,10}, { 223, 9}, \ + { 447,12}, { 63,10}, { 255, 9}, { 511, 8}, \ + { 1023, 9}, { 543,10}, { 287, 9}, { 575, 8}, \ + { 1151,11}, { 159,10}, { 319, 9}, { 639,11}, \ + { 175,10}, { 351, 9}, { 703, 8}, { 1407,12}, \ + { 95,11}, { 191,10}, { 383, 9}, { 767,11}, \ + { 207,10}, { 415, 9}, { 831,11}, { 223,10}, \ + { 447,13}, { 63,11}, { 255,10}, { 543,11}, \ + { 287,10}, { 575, 9}, { 1151,12}, { 159,11}, \ + { 319,10}, { 639,11}, { 351,10}, { 703, 9}, \ + { 1407, 8}, { 2815,12}, { 191,11}, { 383,10}, \ + { 767,11}, { 415,10}, { 831,12}, { 223,11}, \ + { 447,10}, { 895,11}, { 479,10}, { 959, 9}, \ + { 1919,12}, { 255,11}, { 511,10}, { 1023,11}, \ + { 543,10}, { 1087,12}, { 287,11}, { 575,10}, \ + { 1151,12}, { 319,11}, { 639,12}, { 351,11}, \ + { 703,10}, { 1407, 9}, { 2815,13}, { 191,12}, \ + { 383,11}, { 767,12}, { 415,11}, { 831,10}, \ + { 1663,12}, { 447,11}, { 895,10}, { 1791,12}, \ + { 479,11}, { 959,13}, { 255,12}, { 511,11}, \ + { 1023,12}, { 543,11}, { 1087,12}, { 575,11}, \ + { 1151,13}, { 319,12}, { 703,11}, { 1407,10}, \ + { 2815,13}, { 383,12}, { 831,11}, { 1663,13}, \ + { 447,12}, { 895,11}, { 1791,12}, { 959,11}, \ + { 1919,14}, { 255,13}, { 511,12}, { 1087,13}, \ + { 575,12}, { 1151,13}, { 703,12}, { 1407,11}, \ + { 2815,14}, { 383,13}, { 831,12}, { 1663,13}, \ + { 895,12}, { 1791,13}, { 959,12}, { 1919,15}, \ + { 255,14}, { 511,13}, { 1087,12}, { 2175,13}, \ + { 1215,14}, { 639,13}, { 1407,12}, { 2815,14}, \ + { 767,13}, { 1663,14}, { 895,13}, { 1919,12}, \ + { 3839,15}, { 511,14}, { 1023,13}, { 2175,14}, \ + { 1151,13}, { 2431,12}, { 4863,14}, { 16384,15}, \ + { 32768,16}, { 65536,17}, { 131072,18}, { 262144,19}, \ + { 524288,20}, {1048576,21}, {2097152,22}, {4194304,23}, \ + {8388608,24} } +#define MUL_FFT_TABLE3_SIZE 217 +#define MUL_FFT_THRESHOLD 3200 + +#define SQR_FFT_MODF_THRESHOLD 276 /* k = 5 */ +#define SQR_FFT_TABLE3 \ + { { 276, 5}, { 13, 6}, { 7, 5}, { 15, 6}, \ + { 17, 7}, { 17, 8}, { 9, 7}, { 20, 8}, \ + { 11, 7}, { 23, 8}, { 13, 9}, { 7, 8}, \ + { 15, 7}, { 31, 8}, { 19, 9}, { 11, 8}, \ + { 27, 9}, { 15, 8}, { 33, 9}, { 19, 8}, \ + { 39, 9}, { 23, 8}, { 47, 9}, { 27,10}, \ + { 15, 9}, { 39,10}, { 23, 9}, { 47,11}, \ + { 15,10}, { 31, 9}, { 67,10}, { 39, 9}, \ + { 79,10}, { 47, 9}, { 95, 8}, { 191,10}, \ + { 55,11}, { 31,10}, { 63, 8}, { 255,10}, \ + { 71, 9}, { 143, 8}, { 287,10}, { 79, 9}, \ + { 159,11}, { 47,10}, { 95, 9}, { 191, 8}, \ + { 383, 7}, { 767,10}, { 103,12}, { 31,11}, \ + { 63, 9}, { 255, 8}, { 511, 7}, { 1023,10}, \ + { 143, 9}, { 287,11}, { 79,10}, { 159, 9}, \ + { 319, 8}, { 639,10}, { 175, 9}, { 351, 8}, \ + { 703,11}, { 95,10}, { 191, 9}, { 383, 8}, \ + { 767,10}, { 207, 9}, { 415, 8}, { 831,10}, \ + { 223, 9}, { 447,12}, { 63,10}, { 255, 9}, \ + { 511, 8}, { 1023,11}, { 143,10}, { 287, 9}, \ + { 575, 8}, { 1151,11}, { 159,10}, { 319, 9}, \ + { 639,11}, { 175,10}, { 351, 9}, { 703,12}, \ + { 95,11}, { 191,10}, { 383, 9}, { 767,11}, \ + { 207,10}, { 415, 9}, { 831,11}, { 223,10}, \ + { 447,13}, { 63,11}, { 255,10}, { 511, 9}, \ + { 1023,11}, { 287,10}, { 575, 9}, { 1151,12}, \ + { 159,11}, { 319,10}, { 639,11}, { 351,10}, \ + { 703, 9}, { 1407,12}, { 191,11}, { 383,10}, \ + { 767,11}, { 415,10}, { 831,12}, { 223,11}, \ + { 447,10}, { 895,11}, { 479,10}, { 959,12}, \ + { 255,11}, { 511,10}, { 1023,12}, { 287,11}, \ + { 575,10}, { 1151,12}, { 319,11}, { 639,12}, \ + { 351,11}, { 703,10}, { 1407,13}, { 191,12}, \ + { 383,11}, { 767,12}, { 415,11}, { 831,10}, \ + { 1663,12}, { 447,11}, { 895,12}, { 479,11}, \ + { 959,10}, { 1919,13}, { 255,12}, { 511,11}, \ + { 1023,12}, { 543,11}, { 1087,12}, { 575,11}, \ + { 1151,13}, { 319,12}, { 703,11}, { 1407,10}, \ + { 2815,13}, { 383,12}, { 831,11}, { 1663,13}, \ + { 447,12}, { 895,11}, { 1791,12}, { 959,14}, \ + { 255,13}, { 511,12}, { 1087,13}, { 575,12}, \ + { 1151,13}, { 703,12}, { 1407,11}, { 2815,14}, \ + { 383,13}, { 831,12}, { 1663,13}, { 895,12}, \ + { 1791,13}, { 959,12}, { 1919,15}, { 255,14}, \ + { 511,13}, { 1087,12}, { 2175,13}, { 1151,14}, \ + { 639,13}, { 1407,12}, { 2815,14}, { 767,13}, \ + { 1663,14}, { 895,13}, { 1919,12}, { 3839,15}, \ + { 511,14}, { 1023,13}, { 2175,14}, { 1151,13}, \ + { 2431,12}, { 4863,14}, { 16384,15}, { 32768,16}, \ + { 65536,17}, { 131072,18}, { 262144,19}, { 524288,20}, \ + {1048576,21}, {2097152,22}, {4194304,23}, {8388608,24} } +#define SQR_FFT_TABLE3_SIZE 204 +#define SQR_FFT_THRESHOLD 2688 + +#define MULLO_BASECASE_THRESHOLD 0 /* always */ +#define MULLO_DC_THRESHOLD 38 +#define MULLO_MUL_N_THRESHOLD 6253 +#define SQRLO_BASECASE_THRESHOLD 4 +#define SQRLO_DC_THRESHOLD 67 +#define SQRLO_SQR_THRESHOLD 5240 + +#define DC_DIV_QR_THRESHOLD 43 +#define DC_DIVAPPR_Q_THRESHOLD 155 +#define DC_BDIV_QR_THRESHOLD 39 +#define DC_BDIV_Q_THRESHOLD 89 + +#define INV_MULMOD_BNM1_THRESHOLD 34 +#define INV_NEWTON_THRESHOLD 163 +#define INV_APPR_THRESHOLD 161 + +#define BINV_NEWTON_THRESHOLD 196 +#define REDC_1_TO_REDC_N_THRESHOLD 43 + +#define MU_DIV_QR_THRESHOLD 998 +#define MU_DIVAPPR_Q_THRESHOLD 998 +#define MUPI_DIV_QR_THRESHOLD 91 +#define MU_BDIV_QR_THRESHOLD 807 +#define MU_BDIV_Q_THRESHOLD 924 + +#define POWM_SEC_TABLE 6,30,125,579,1730 + +#define GET_STR_DC_THRESHOLD 15 +#define GET_STR_PRECOMPUTE_THRESHOLD 30 +#define SET_STR_DC_THRESHOLD 802 +#define SET_STR_PRECOMPUTE_THRESHOLD 1815 + +#define FAC_DSC_THRESHOLD 258 +#define FAC_ODD_THRESHOLD 24 + +#define MATRIX22_STRASSEN_THRESHOLD 10 +#define HGCD2_DIV1_METHOD 1 /* 7.05% faster than 3 */ +#define HGCD_THRESHOLD 107 +#define HGCD_APPR_THRESHOLD 112 +#define HGCD_REDUCE_THRESHOLD 1679 +#define GCD_DC_THRESHOLD 324 +#define GCDEXT_DC_THRESHOLD 242 +#define JACOBI_BASE_METHOD 4 /* 22.41% faster than 1 */ + +/* Tuneup completed successfully, took 66624 seconds */ |