[Gmp-commit] /var/hg/gmp: 3 new changesets
mercurial at gmplib.org
mercurial at gmplib.org
Fri Dec 4 16:40:37 UTC 2020
details: /var/hg/gmp/rev/143fde6fef5a
changeset: 18171:143fde6fef5a
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Dec 04 17:35:08 2020 +0100
description:
Use sub insn for subtracting constants.
details: /var/hg/gmp/rev/ac37c69b7304
changeset: 18172:ac37c69b7304
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Dec 04 17:35:54 2020 +0100
description:
Provide add_mssaaaa for arm64.
details: /var/hg/gmp/rev/97959c7c766d
changeset: 18173:97959c7c766d
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Dec 04 17:40:04 2020 +0100
description:
Provide tuneup parameters for Apple M1.
diffstat:
mpn/arm64/applem1/gmp-mparam.h | 186 +++++++++++++++++++++++++++++++++++++++++
mpn/arm64/divrem_1.asm | 4 +-
mpn/generic/div_qr_1n_pi1.c | 11 ++-
mpn/generic/mod_1_1.c | 9 +
4 files changed, 207 insertions(+), 3 deletions(-)
diffs (258 lines):
diff -r 1a5f0f394f98 -r 97959c7c766d mpn/arm64/applem1/gmp-mparam.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/mpn/arm64/applem1/gmp-mparam.h Fri Dec 04 17:40:04 2020 +0100
@@ -0,0 +1,186 @@
+/* gmp-mparam.h -- Compiler/machine parameter header file.
+
+Copyright 2020 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
+
+/* 3200 MHz Apple M1 */
+/* FFT tuning limit = 1 M */
+/* Generated by tuneup.c, 2020-11-30, clang */
+
+#define MOD_1_1P_METHOD 2 /* 8.96% faster than 1 */
+#define MOD_1_NORM_THRESHOLD 0 /* always */
+#define MOD_1_UNNORM_THRESHOLD 0 /* always */
+#define MOD_1N_TO_MOD_1_1_THRESHOLD 13
+#define MOD_1U_TO_MOD_1_1_THRESHOLD 5
+#define MOD_1_1_TO_MOD_1_2_THRESHOLD 6
+#define MOD_1_2_TO_MOD_1_4_THRESHOLD 18
+#define PREINV_MOD_1_TO_MOD_1_THRESHOLD 21
+#define USE_PREINV_DIVREM_1 1 /* native */
+#define DIV_QR_1N_PI1_METHOD 1 /* 36.98% faster than 2 */
+#define DIV_QR_1_NORM_THRESHOLD 2
+#define DIV_QR_1_UNNORM_THRESHOLD 1
+#define DIV_QR_2_PI2_THRESHOLD 9
+#define DIVEXACT_1_THRESHOLD 0 /* always */
+#define BMOD_1_TO_MOD_1_THRESHOLD 28
+
+#define DIV_1_VS_MUL_1_PERCENT 659
+
+#define MUL_TOOM22_THRESHOLD 26
+#define MUL_TOOM33_THRESHOLD 63
+#define MUL_TOOM44_THRESHOLD 190
+#define MUL_TOOM6H_THRESHOLD 351
+#define MUL_TOOM8H_THRESHOLD 537
+
+#define MUL_TOOM32_TO_TOOM43_THRESHOLD 94
+#define MUL_TOOM32_TO_TOOM53_THRESHOLD 81
+#define MUL_TOOM42_TO_TOOM53_THRESHOLD 79
+#define MUL_TOOM42_TO_TOOM63_THRESHOLD 99
+#define MUL_TOOM43_TO_TOOM54_THRESHOLD 135
+
+#define SQR_BASECASE_THRESHOLD 0 /* always */
+#define SQR_TOOM2_THRESHOLD 53
+#define SQR_TOOM3_THRESHOLD 87
+#define SQR_TOOM4_THRESHOLD 387
+#define SQR_TOOM6_THRESHOLD 442
+#define SQR_TOOM8_THRESHOLD 502
+
+#define MULMID_TOOM42_THRESHOLD 44
+
+#define MULMOD_BNM1_THRESHOLD 10
+#define SQRMOD_BNM1_THRESHOLD 11
+
+#define MUL_FFT_MODF_THRESHOLD 216 /* k = 5 */
+#define MUL_FFT_TABLE3 \
+ { { 216, 5}, { 7, 4}, { 17, 5}, { 9, 4}, \
+ { 19, 5}, { 19, 6}, { 21, 7}, { 11, 6}, \
+ { 23, 7}, { 21, 8}, { 11, 7}, { 24, 8}, \
+ { 13, 7}, { 27, 8}, { 21, 9}, { 11, 8}, \
+ { 27, 9}, { 15, 8}, { 33, 9}, { 19, 8}, \
+ { 39, 9}, { 27,10}, { 15, 9}, { 39,10}, \
+ { 23, 9}, { 47,11}, { 15,10}, { 31, 9}, \
+ { 63,10}, { 39, 9}, { 79,10}, { 47,11}, \
+ { 31,10}, { 79,11}, { 47,12}, { 31,11}, \
+ { 63,10}, { 127, 9}, { 255, 8}, { 511,11}, \
+ { 79,10}, { 159, 9}, { 319, 8}, { 639,11}, \
+ { 95,10}, { 191, 9}, { 383,12}, { 63,11}, \
+ { 127,10}, { 255, 9}, { 511, 8}, { 1023,11}, \
+ { 143,10}, { 287, 9}, { 575, 8}, { 1151,11}, \
+ { 159,10}, { 319, 9}, { 639,12}, { 95,11}, \
+ { 191,10}, { 383,13}, { 63,12}, { 127,11}, \
+ { 255,10}, { 511, 9}, { 1023,11}, { 271,10}, \
+ { 543, 9}, { 1087, 8}, { 2175,11}, { 287,10}, \
+ { 575, 9}, { 1151,11}, { 303,10}, { 607,12}, \
+ { 159,11}, { 319,10}, { 639, 9}, { 1279,11}, \
+ { 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, 9}, { 1919,13}, { 8192,14}, { 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 105
+#define MUL_FFT_THRESHOLD 2240
+
+#define SQR_FFT_MODF_THRESHOLD 300 /* k = 5 */
+#define SQR_FFT_TABLE3 \
+ { { 300, 5}, { 10, 4}, { 22, 5}, { 21, 6}, \
+ { 21, 7}, { 11, 6}, { 23, 7}, { 21, 8}, \
+ { 11, 7}, { 24, 8}, { 15, 7}, { 31, 8}, \
+ { 21, 9}, { 11, 8}, { 25, 9}, { 15, 8}, \
+ { 33, 9}, { 19, 8}, { 39, 9}, { 23, 8}, \
+ { 47,10}, { 15, 9}, { 39,10}, { 23, 9}, \
+ { 47,11}, { 15,10}, { 31, 9}, { 63,10}, \
+ { 47,11}, { 31,10}, { 79,11}, { 47,12}, \
+ { 31,11}, { 63,10}, { 127, 9}, { 255, 8}, \
+ { 511, 7}, { 1023,11}, { 79,10}, { 159, 9}, \
+ { 319, 8}, { 639,11}, { 95,10}, { 191, 9}, \
+ { 383,12}, { 63,11}, { 127,10}, { 255, 9}, \
+ { 511, 8}, { 1023,10}, { 271, 9}, { 543, 8}, \
+ { 1087,10}, { 287, 9}, { 575, 8}, { 1151,10}, \
+ { 303,11}, { 159,10}, { 319, 9}, { 639,12}, \
+ { 95,11}, { 191,10}, { 383,13}, { 63,12}, \
+ { 127,10}, { 511, 9}, { 1023,11}, { 271,10}, \
+ { 543, 9}, { 1087,11}, { 287,10}, { 575, 9}, \
+ { 1151,11}, { 303,10}, { 607,12}, { 159,11}, \
+ { 319,10}, { 639, 9}, { 1279,11}, { 335,10}, \
+ { 671,11}, { 351,10}, { 703,12}, { 191,11}, \
+ { 383,10}, { 767,11}, { 415,10}, { 831, 9}, \
+ { 1663,12}, { 223,11}, { 447,10}, { 895,11}, \
+ { 479,10}, { 959, 9}, { 1919, 8}, { 3839,12}, \
+ { 4096,13}, { 8192,14}, { 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 108
+#define SQR_FFT_THRESHOLD 1856
+
+#define MULLO_BASECASE_THRESHOLD 0 /* always */
+#define MULLO_DC_THRESHOLD 106
+#define MULLO_MUL_N_THRESHOLD 4292
+#define SQRLO_BASECASE_THRESHOLD 4
+#define SQRLO_DC_THRESHOLD 177
+#define SQRLO_SQR_THRESHOLD 3690
+
+#define DC_DIV_QR_THRESHOLD 67
+#define DC_DIVAPPR_Q_THRESHOLD 242
+#define DC_BDIV_QR_THRESHOLD 68
+#define DC_BDIV_Q_THRESHOLD 129
+
+#define INV_MULMOD_BNM1_THRESHOLD 82
+#define INV_NEWTON_THRESHOLD 157
+#define INV_APPR_THRESHOLD 157
+
+#define BINV_NEWTON_THRESHOLD 99
+#define REDC_1_TO_REDC_N_THRESHOLD 67
+
+#define MU_DIV_QR_THRESHOLD 1037
+#define MU_DIVAPPR_Q_THRESHOLD 1120
+#define MUPI_DIV_QR_THRESHOLD 76
+#define MU_BDIV_QR_THRESHOLD 1037
+#define MU_BDIV_Q_THRESHOLD 1290
+
+#define POWM_SEC_TABLE 59,72,78,702,2211
+
+#define GET_STR_DC_THRESHOLD 12
+#define GET_STR_PRECOMPUTE_THRESHOLD 21
+#define SET_STR_DC_THRESHOLD 668
+#define SET_STR_PRECOMPUTE_THRESHOLD 1212
+
+#define FAC_DSC_THRESHOLD 216
+#define FAC_ODD_THRESHOLD 0 /* always */
+
+#define MATRIX22_STRASSEN_THRESHOLD 9
+#define HGCD2_DIV1_METHOD 1 /* 7.39% faster than 3 */
+#define HGCD_THRESHOLD 133
+#define HGCD_APPR_THRESHOLD 133
+#define HGCD_REDUCE_THRESHOLD 2121
+#define GCD_DC_THRESHOLD 411
+#define GCDEXT_DC_THRESHOLD 496
+#define JACOBI_BASE_METHOD 1 /* 26.16% faster than 4 */
diff -r 1a5f0f394f98 -r 97959c7c766d mpn/arm64/divrem_1.asm
--- a/mpn/arm64/divrem_1.asm Mon Nov 30 09:43:41 2020 +0100
+++ b/mpn/arm64/divrem_1.asm Fri Dec 04 17:40:04 2020 +0100
@@ -73,7 +73,7 @@
stp x21, x22, [sp, #32]
stp x23, x24, [sp, #48]
- add n, n_arg, #-1
+ sub n, n_arg, #1
add x7, n, fn_arg
add np, np_arg, n, lsl #3
add qp, qp_arg, x7, lsl #3
@@ -93,7 +93,7 @@
stp x21, x22, [sp, #32]
stp x23, x24, [sp, #48]
- add n, n_arg, #-1
+ sub n, n_arg, #1
add x7, n, fn_arg
add np, np_arg, n, lsl #3
add qp, qp_arg, x7, lsl #3
diff -r 1a5f0f394f98 -r 97959c7c766d mpn/generic/div_qr_1n_pi1.c
--- a/mpn/generic/div_qr_1n_pi1.c Mon Nov 30 09:43:41 2020 +0100
+++ b/mpn/generic/div_qr_1n_pi1.c Fri Dec 04 17:40:04 2020 +0100
@@ -115,7 +115,7 @@
"nor %0, %0, %0" \
: "=r" (m), "=r" (s1), "=&r" (s0) \
: "r" (a1), "r" (b1), "%r" (a0), "rI" (b0) \
- __CLOBBER_CC)
+ __CLOBBER_CC)
#endif
#if defined (__s390x__) && W_TYPE_SIZE == 64
@@ -139,6 +139,15 @@
: "=r" (m), "=r" (sh), "=&r" (sl) \
: "r" (ah), "rI" (bh), "%r" (al), "rI" (bl) __CLOBBER_CC)
#endif
+
+#if defined (__aarch64__) && W_TYPE_SIZE == 64
+#define add_mssaaaa(m, sh, sl, ah, al, bh, bl) \
+ __asm__ ( "adds %2, %5, %6\n\t" \
+ "adcs %1, %3, %4\n\t" \
+ "csinv %0, xzr, xzr, cs\n\t" \
+ : "=r" (m), "=r" (sh), "=&r" (sl) \
+ : "r" (ah), "rI" (bh), "%r" (al), "rI" (bl) __CLOBBER_CC)
+#endif
#endif /* defined (__GNUC__) */
#ifndef add_mssaaaa
diff -r 1a5f0f394f98 -r 97959c7c766d mpn/generic/mod_1_1.c
--- a/mpn/generic/mod_1_1.c Mon Nov 30 09:43:41 2020 +0100
+++ b/mpn/generic/mod_1_1.c Fri Dec 04 17:40:04 2020 +0100
@@ -138,6 +138,15 @@
: "=r" (m), "=r" (sh), "=&r" (sl) \
: "r" (ah), "rI" (bh), "%r" (al), "rI" (bl) __CLOBBER_CC)
#endif
+
+#if defined (__aarch64__) && W_TYPE_SIZE == 64
+#define add_mssaaaa(m, sh, sl, ah, al, bh, bl) \
+ __asm__ ( "adds %2, %5, %6\n\t" \
+ "adcs %1, %3, %4\n\t" \
+ "csinv %0, xzr, xzr, cs\n\t" \
+ : "=r" (m), "=r" (sh), "=&r" (sl) \
+ : "r" (ah), "rI" (bh), "%r" (al), "rI" (bl) __CLOBBER_CC)
+#endif
#endif /* defined (__GNUC__) */
#ifndef add_mssaaaa
More information about the gmp-commit
mailing list