[Gmp-commit] /var/hg/gmp: Add Skylake support.
mercurial at gmplib.org
mercurial at gmplib.org
Mon Aug 17 19:47:25 UTC 2015
details: /var/hg/gmp/rev/7850a6d24896
changeset: 16768:7850a6d24896
user: Torbjorn Granlund <torbjorng at google.com>
date: Mon Aug 17 21:47:23 2015 +0200
description:
Add Skylake support.
diffstat:
ChangeLog | 14 ++-
acinclude.m4 | 2 +-
config.guess | 1 +
config.sub | 2 +-
configure.ac | 7 +
mpn/x86_64/skylake/gmp-mparam.h | 194 ++++++++++++++++++++++++++++++++++++++++
6 files changed, 217 insertions(+), 3 deletions(-)
diffs (278 lines):
diff -r 451c0f77758f -r 7850a6d24896 ChangeLog
--- a/ChangeLog Sat Aug 15 21:08:15 2015 +0200
+++ b/ChangeLog Mon Aug 17 21:47:23 2015 +0200
@@ -1,3 +1,15 @@
+2015-08-17 Torbjörn Granlund <torbjorng at google.com>
+
+ * acinclude.m4 (X86_64_PATTERN): Add skylake.
+ * config.guess: Corresponding changes.
+ * config.sub: Corresponding changes.
+ * configure.ac: Corresponding changes.
+ * mpn/x86_64/skylake/gmp-mparam.h: New file.
+
+2015-08-15 Torbjörn Granlund <torbjorng at google.com>
+
+ * mpn/generic/mullo_basecase.c: Provide alternative code, make default.
+
2015-08-04 Marco Bodrato <bodrato at mail.dm.unipi.it>
* tests/refmpn.c (refmpn_sqrlo): New function.
@@ -12,7 +24,7 @@
* tests/mpn/Makefile.am (check_PROGRAMS): Add new test.
* tests/devel/try.c: Support mpn_sqrlo and mpn_sqrlo_basecase.
- * tune/common.c (speed_mpn_sqrlo{,_basecase}): New functions.
+ * tune/common.c (speed_mpn_sqrlo{,_basecase}): New functions.
* tune/speed.c: Support new functions.
* tune/speed.h (SPEED_ROUTINE_MPN_MULLO_BASECASE): Update.
(SPEED_ROUTINE_MPN_SQRLO): New macro.
diff -r 451c0f77758f -r 7850a6d24896 acinclude.m4
--- a/acinclude.m4 Sat Aug 15 21:08:15 2015 +0200
+++ b/acinclude.m4 Mon Aug 17 21:47:23 2015 +0200
@@ -63,7 +63,7 @@
[[i?86*-*-* | k[5-8]*-*-* | pentium*-*-* | athlon-*-* | viac3*-*-* | geode*-*-* | atom-*-*]])
define(X86_64_PATTERN,
-[[athlon64-*-* | k8-*-* | k10-*-* | bobcat-*-* | jaguar*-*-* | bulldozer*-*-* | piledriver*-*-* | steamroller*-*-* | excavator*-*-* | pentium4-*-* | atom-*-* | silvermont-*-* | core2-*-* | corei*-*-* | x86_64-*-* | nano-*-* | nehalem*-*-* | westmere*-*-* | sandybridge*-*-* | ivybridge*-*-* | haswell*-*-* | broadwell*-*-*]])
+[[athlon64-*-* | k8-*-* | k10-*-* | bobcat-*-* | jaguar*-*-* | bulldozer*-*-* | piledriver*-*-* | steamroller*-*-* | excavator*-*-* | pentium4-*-* | atom-*-* | silvermont-*-* | core2-*-* | corei*-*-* | x86_64-*-* | nano-*-* | nehalem*-*-* | westmere*-*-* | sandybridge*-*-* | ivybridge*-*-* | haswell*-*-* | broadwell*-*-* | skylake*-*-*]])
dnl GMP_FAT_SUFFIX(DSTVAR, DIRECTORY)
dnl ---------------------------------
diff -r 451c0f77758f -r 7850a6d24896 config.guess
--- a/config.guess Sat Aug 15 21:08:15 2015 +0200
+++ b/config.guess Mon Aug 17 21:47:23 2015 +0200
@@ -832,6 +832,7 @@
else if (model == 0x4d) cpu_64bit = 1, modelstr = "silvermont"; /* Silvermont/Avoton */
else if (model == 0x4f) cpu_64bit = 1, cpu_avx=1, modelstr = "broadwell"; /* Broadwell server */
else if (model == 0x56) cpu_64bit = 1, cpu_avx=1, modelstr = "broadwell"; /* Broadwell microserver */
+ else if (model == 0x5e) cpu_64bit = 1, cpu_avx=1, modelstr = "skylake"; /* Skylake */
else cpu_64bit = 1, modelstr = "nehalem"; /* default */
if (strcmp (modelstr, "haswell") == 0)
diff -r 451c0f77758f -r 7850a6d24896 config.sub
--- a/config.sub Sat Aug 15 21:08:15 2015 +0200
+++ b/config.sub Mon Aug 17 21:47:23 2015 +0200
@@ -102,7 +102,7 @@
test_cpu=ia64 ;;
pentium | pentiummmx | pentiumpro | pentium[234m] | k[567] | k6[23] | geode | athlon | viac3*)
test_cpu=i386 ;;
-athlon64 | atom | silvermont | core2 | corei* | opteron | k[89] | k10 | bobcat | jaguar* | bulldozer* | piledriver* | steamroller* | excavator* | nano | nehalem* | westmere* | sandybridge* | ivybridge* | haswell* | broadwell*)
+athlon64 | atom | silvermont | core2 | corei* | opteron | k[89] | k10 | bobcat | jaguar* | bulldozer* | piledriver* | steamroller* | excavator* | nano | nehalem* | westmere* | sandybridge* | ivybridge* | haswell* | broadwell* | skylake*)
test_cpu=x86_64 ;;
power[2-9] | power2sc)
test_cpu=power ;;
diff -r 451c0f77758f -r 7850a6d24896 configure.ac
--- a/configure.ac Sat Aug 15 21:08:15 2015 +0200
+++ b/configure.ac Mon Aug 17 21:47:23 2015 +0200
@@ -1800,6 +1800,13 @@
path_64="x86_64/coreibwl x86_64/coreihwl x86_64/coreisbr x86_64/coreinhm x86_64/core2 x86_64"
# extra_functions_64="missing" # enable for bmi2/adx simulation
;;
+ skylake | skylakenoavx)
+ gcc_cflags_cpu="-mtune=skylake -mtune=broadwell -mtune=corei7 -mtune=core2 -mtune=k8"
+ gcc_cflags_arch="-march=skylake -march=broadwell -march=corei7 -march=core2 -march=core2~-mno-sse2 -march=k8 -march=k8~-mno-sse2"
+ path="x86/coreisbr x86/p6/sse2 x86/p6/p3mmx x86/p6/mmx x86/p6 x86/mmx x86"
+ path_64="x86_64/skylake x86_64/coreibwl x86_64/coreihwl x86_64/coreisbr x86_64/coreinhm x86_64/core2 x86_64"
+ # extra_functions_64="missing" # enable for bmi2/adx simulation
+ ;;
atom)
gcc_cflags_cpu="-mtune=atom -mtune=pentium3"
gcc_cflags_arch="-march=atom -march=pentium3"
diff -r 451c0f77758f -r 7850a6d24896 mpn/x86_64/skylake/gmp-mparam.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/mpn/x86_64/skylake/gmp-mparam.h Mon Aug 17 21:47:23 2015 +0200
@@ -0,0 +1,194 @@
+/* Skylake gmp-mparam.h -- Compiler/machine parameter header file.
+
+Copyright 1991, 1993, 1994, 2000-2015 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
+
+/* Disable use of slow functions. FIXME: We should disable lib inclusion. */
+#undef HAVE_NATIVE_mpn_mul_2
+#undef HAVE_NATIVE_mpn_addmul_2
+
+/* 3500 MHz i5-6600K Skylake */
+/* Generated by tuneup.c, 2015-04-23, gcc 4.5 */
+
+#define MOD_1_NORM_THRESHOLD 0 /* always */
+#define MOD_1_UNNORM_THRESHOLD 0 /* always */
+#define MOD_1N_TO_MOD_1_1_THRESHOLD 4
+#define MOD_1U_TO_MOD_1_1_THRESHOLD 3
+#define MOD_1_1_TO_MOD_1_2_THRESHOLD 13
+#define MOD_1_2_TO_MOD_1_4_THRESHOLD 34
+#define PREINV_MOD_1_TO_MOD_1_THRESHOLD 7
+#define USE_PREINV_DIVREM_1 1 /* native */
+#define DIV_QR_1_NORM_THRESHOLD 1
+#define DIV_QR_1_UNNORM_THRESHOLD MP_SIZE_T_MAX /* never */
+#define DIV_QR_2_PI2_THRESHOLD MP_SIZE_T_MAX /* never */
+#define DIVEXACT_1_THRESHOLD 0 /* always (native) */
+#define BMOD_1_TO_MOD_1_THRESHOLD 22
+
+#define MUL_TOOM22_THRESHOLD 26
+#define MUL_TOOM33_THRESHOLD 73
+#define MUL_TOOM44_THRESHOLD 208
+#define MUL_TOOM6H_THRESHOLD 418
+#define MUL_TOOM8H_THRESHOLD 0 /* always */
+
+#define MUL_TOOM32_TO_TOOM43_THRESHOLD 73
+#define MUL_TOOM32_TO_TOOM53_THRESHOLD 152
+#define MUL_TOOM42_TO_TOOM53_THRESHOLD 137
+#define MUL_TOOM42_TO_TOOM63_THRESHOLD 163
+#define MUL_TOOM43_TO_TOOM54_THRESHOLD 203
+
+#define SQR_BASECASE_THRESHOLD 0 /* always (native) */
+#define SQR_TOOM2_THRESHOLD 33
+#define SQR_TOOM3_THRESHOLD 105
+#define SQR_TOOM4_THRESHOLD 450
+#define SQR_TOOM6_THRESHOLD 552
+#define SQR_TOOM8_THRESHOLD 597
+
+#define MULMID_TOOM42_THRESHOLD 48
+
+#define MULMOD_BNM1_THRESHOLD 15
+#define SQRMOD_BNM1_THRESHOLD 18
+
+#define MUL_FFT_MODF_THRESHOLD 400 /* k = 5 */
+#define MUL_FFT_TABLE3 \
+ { { 400, 5}, { 21, 6}, { 11, 5}, { 23, 6}, \
+ { 23, 7}, { 12, 6}, { 25, 7}, { 16, 6}, \
+ { 33, 7}, { 25, 8}, { 13, 7}, { 28, 8}, \
+ { 15, 7}, { 32, 8}, { 17, 7}, { 35, 8}, \
+ { 19, 7}, { 39, 8}, { 21, 9}, { 11, 8}, \
+ { 27, 9}, { 15, 8}, { 35, 9}, { 19, 8}, \
+ { 41, 9}, { 23, 8}, { 49, 9}, { 27,10}, \
+ { 15, 9}, { 43,10}, { 23, 9}, { 51,11}, \
+ { 15,10}, { 31, 9}, { 67,10}, { 39, 9}, \
+ { 83,10}, { 47, 9}, { 99,10}, { 55,11}, \
+ { 31,10}, { 63, 9}, { 127,10}, { 79,11}, \
+ { 47,10}, { 95,12}, { 31,11}, { 63,10}, \
+ { 135,11}, { 79,10}, { 167,11}, { 95,10}, \
+ { 191, 9}, { 383,11}, { 111,12}, { 63,11}, \
+ { 127,10}, { 255, 9}, { 511,10}, { 271, 9}, \
+ { 543,11}, { 143,10}, { 287, 9}, { 575,11}, \
+ { 159,10}, { 319,12}, { 95,11}, { 191,10}, \
+ { 383,11}, { 207,10}, { 415,13}, { 63, 9}, \
+ { 1023,11}, { 271,10}, { 543, 9}, { 1087,10}, \
+ { 575,11}, { 303,10}, { 607,12}, { 159,11}, \
+ { 319,10}, { 639,11}, { 335,10}, { 671,11}, \
+ { 351,10}, { 703,12}, { 191,11}, { 383,10}, \
+ { 767,11}, { 415,10}, { 831,11}, { 447,13}, \
+ { 127,11}, { 511,10}, { 1023,11}, { 543,10}, \
+ { 1087,12}, { 287,11}, { 607,12}, { 319,11}, \
+ { 671,12}, { 351,11}, { 703,13}, { 191,12}, \
+ { 383,11}, { 767,12}, { 415,11}, { 831,12}, \
+ { 479,14}, { 127,12}, { 543,11}, { 1087,12}, \
+ { 607,13}, { 319,12}, { 735,13}, { 383,12}, \
+ { 831,13}, { 447,12}, { 895,11}, { 1791,12}, \
+ { 959,13}, { 511,12}, { 1023,11}, { 2047,12}, \
+ { 1087,13}, { 575,12}, { 1215,13}, { 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 138
+#define MUL_FFT_THRESHOLD 4736
+
+#define SQR_FFT_MODF_THRESHOLD 376 /* k = 5 */
+#define SQR_FFT_TABLE3 \
+ { { 376, 5}, { 23, 6}, { 25, 7}, { 13, 6}, \
+ { 27, 7}, { 25, 8}, { 13, 7}, { 28, 8}, \
+ { 15, 7}, { 31, 8}, { 21, 9}, { 11, 8}, \
+ { 27, 9}, { 15, 8}, { 35, 9}, { 19, 8}, \
+ { 41, 9}, { 23, 8}, { 47, 9}, { 27,10}, \
+ { 15, 9}, { 39,10}, { 23, 9}, { 51,11}, \
+ { 15,10}, { 31, 9}, { 67,10}, { 39, 9}, \
+ { 79,10}, { 47, 9}, { 95,10}, { 55,11}, \
+ { 31,10}, { 79,11}, { 47,10}, { 95,12}, \
+ { 31,11}, { 63,10}, { 135,11}, { 79,10}, \
+ { 159,11}, { 95,10}, { 191,12}, { 63,11}, \
+ { 127,10}, { 255, 9}, { 511,10}, { 271,11}, \
+ { 143,10}, { 287, 9}, { 575,10}, { 303,11}, \
+ { 159,10}, { 319,12}, { 95,11}, { 191,10}, \
+ { 383, 9}, { 767,13}, { 63,12}, { 127,11}, \
+ { 255,10}, { 511,11}, { 271,10}, { 575, 6}, \
+ { 9727, 7}, { 4991, 5}, { 19967, 7}, { 5375, 9}, \
+ { 1407,10}, { 735,11}, { 383,10}, { 767, 9}, \
+ { 1535,10}, { 799,11}, { 415,10}, { 831,11}, \
+ { 447,10}, { 895,11}, { 479,13}, { 127,11}, \
+ { 511,10}, { 1023,11}, { 575,10}, { 1151,11}, \
+ { 607,12}, { 319,11}, { 671,12}, { 351,11}, \
+ { 703,12}, { 383,11}, { 799,12}, { 415,11}, \
+ { 831,12}, { 479,14}, { 127,13}, { 255,12}, \
+ { 607,13}, { 319,12}, { 671,11}, { 1343,12}, \
+ { 703,13}, { 383,12}, { 831,13}, { 447,12}, \
+ { 895,11}, { 1791,12}, { 959,13}, { 511,12}, \
+ { 1023,13}, { 575,12}, { 1215,13}, { 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 122
+#define SQR_FFT_THRESHOLD 3712
+
+#define MULLO_BASECASE_THRESHOLD 0 /* always */
+#define MULLO_DC_THRESHOLD 25
+#define MULLO_MUL_N_THRESHOLD 9174
+
+#define DC_DIV_QR_THRESHOLD 52
+#define DC_DIVAPPR_Q_THRESHOLD 167
+#define DC_BDIV_QR_THRESHOLD 86
+#define DC_BDIV_Q_THRESHOLD 152
+
+#define INV_MULMOD_BNM1_THRESHOLD 50
+#define INV_NEWTON_THRESHOLD 157
+#define INV_APPR_THRESHOLD 157
+
+#define BINV_NEWTON_THRESHOLD 294
+#define REDC_1_TO_REDC_2_THRESHOLD 30
+#define REDC_2_TO_REDC_N_THRESHOLD 75
+
+#define MU_DIV_QR_THRESHOLD 1620
+#define MU_DIVAPPR_Q_THRESHOLD 1652
+#define MUPI_DIV_QR_THRESHOLD 69
+#define MU_BDIV_QR_THRESHOLD 1620
+#define MU_BDIV_Q_THRESHOLD 1620
+
+#define POWM_SEC_TABLE 1,10,361,960,1099,2577
+
+#define GET_STR_DC_THRESHOLD 12
+#define GET_STR_PRECOMPUTE_THRESHOLD 19
+#define SET_STR_DC_THRESHOLD 987
+#define SET_STR_PRECOMPUTE_THRESHOLD 1948
+
+#define FAC_DSC_THRESHOLD 781
+#define FAC_ODD_THRESHOLD 0 /* always */
+
+#define MATRIX22_STRASSEN_THRESHOLD 15
+#define HGCD_THRESHOLD 71
+#define HGCD_APPR_THRESHOLD 55
+#define HGCD_REDUCE_THRESHOLD 3014
+#define GCD_DC_THRESHOLD 706
+#define GCDEXT_DC_THRESHOLD 393
+#define JACOBI_BASE_METHOD 3
More information about the gmp-commit
mailing list