[Gmp-commit] /var/hg/gmp: 3 new changesets
mercurial at gmplib.org
mercurial at gmplib.org
Fri Nov 30 16:35:21 UTC 2018
details: /var/hg/gmp/rev/d0e7d3489be4
changeset: 17718:d0e7d3489be4
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Nov 30 17:33:57 2018 +0100
description:
* mpn/powerpc64/mode64/p9/add_n_sub_n.asm: New file.
details: /var/hg/gmp/rev/ae2950bd55e4
changeset: 17719:ae2950bd55e4
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Nov 30 17:35:04 2018 +0100
description:
New file.
details: /var/hg/gmp/rev/ba8b7b6c9b1a
changeset: 17720:ba8b7b6c9b1a
user: Torbjorn Granlund <tg at gmplib.org>
date: Fri Nov 30 17:35:17 2018 +0100
description:
ChangeLog
diffstat:
ChangeLog | 17 +++
mpn/powerpc64/mode64/p9/add_n_sub_n.asm | 112 +++++++++++++++++++++
mpn/powerpc64/mode64/p9/gmp-mparam.h | 166 ++++++++++++++++++++++++++++++++
3 files changed, 295 insertions(+), 0 deletions(-)
diffs (truncated from 310 to 300 lines):
diff -r 80949837f4c1 -r ba8b7b6c9b1a ChangeLog
--- a/ChangeLog Wed Nov 28 23:18:30 2018 +0100
+++ b/ChangeLog Fri Nov 30 17:35:17 2018 +0100
@@ -1,3 +1,20 @@
+2018-11-30 Torbjörn Granlund <tg at gmplib.org>
+
+ * mpn/powerpc64/mode64/p9/gmp-mparam.h: New file.
+
+ * mpn/powerpc64/mode64/p9/add_n_sub_n.asm: New file.
+
+2018-11-28 Torbjörn Granlund <tg at gmplib.org>
+
+ * mpn/powerpc64/mode64/p9/sqr_basecase.asm: New file.
+ * mpn/powerpc64/mode64/p9/mul_1.asm: New file.
+
+2018-11-18 Torbjörn Granlund <tg at gmplib.org>
+
+ * mpn/powerpc64/mode64/p9/mul_basecase.asm: New file.
+
+ * mpn/powerpc64/mode64/p9/addmul_1.asm: Optimise.
+
2018-11-12 Torbjörn Granlund <tg at gmplib.org>
* mpn/powerpc64/mode64/p9/aorsmul_1.asm: New file, providing fast
diff -r 80949837f4c1 -r ba8b7b6c9b1a mpn/powerpc64/mode64/p9/add_n_sub_n.asm
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/mpn/powerpc64/mode64/p9/add_n_sub_n.asm Fri Nov 30 17:35:17 2018 +0100
@@ -0,0 +1,112 @@
+dnl PowerPC-64 mpn_add_n_sub_n optimised for POWER9.
+
+dnl Copyright 2018 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 POWER3/PPC630 -
+C POWER4/PPC970 -
+C POWER5 -
+C POWER6 -
+C POWER7 -
+C POWER8 -
+C POWER9 2.25
+
+
+C INPUT PARAMETERS
+define(`arp', `r3')
+define(`srp', `r4')
+define(`up', `r5')
+define(`vp', `r6')
+define(`n', `r7')
+
+ASM_START()
+PROLOGUE(mpn_add_n_sub_n)
+ cmpdi cr7, n, 2
+ subfo r0, r0, r0 C clear OV
+ rldicl. r9, n, 0, 63 C n & 1
+ beq cr0, L(bx0)
+
+L(bx1): ld r10, 0(up)
+ ld r11, 0(vp)
+ ble cr7, L(1)
+ srdi r7, r7, 1
+ mtctr r7
+ ld r8, 8(up)
+ ld r9, 8(vp)
+ addex( r0, r10, r11, 0)
+ subfc r12, r11, r10
+ addi up, up, -8
+ addi vp, vp, -8
+ b L(lo1)
+
+L(bx0): ld r8, 0(up)
+ ld r9, 0(vp)
+ ld r10, 8(up)
+ ld r11, 8(vp)
+ addex( r0, r8, r9, 0)
+ subfc r12, r9, r8
+ addi arp, arp, 8
+ addi srp, srp, 8
+ ble cr7, L(end)
+ addi r7, r7, -1
+ srdi r7, r7, 1
+ mtctr r7
+
+L(top): ld r8, 16(up)
+ ld r9, 16(vp)
+ std r0, -8(arp)
+ std r12, -8(srp)
+ addex( r0, r10, r11, 0)
+ subfe r12, r11, r10
+L(lo1): ld r10, 24(up)
+ ld r11, 24(vp)
+ std r0, 0(arp)
+ std r12, 0(srp)
+ addex( r0, r8, r9, 0)
+ subfe r12, r9, r8
+ addi up, up, 16
+ addi vp, vp, 16
+ addi arp, arp, 16
+ addi srp, srp, 16
+ bdnz L(top)
+
+L(end): std r0, -8(arp)
+ std r12, -8(srp)
+L(1): addex( r0, r10, r11, 0)
+ subfe r12, r11, r10
+ std r0, 0(arp)
+ std r12, 0(srp)
+ subfe r3, r3, r3
+ addex( r3, r3, r3, 0)
+ rldicl r3, r3, 1, 62
+ blr
+EPILOGUE()
+ASM_END()
diff -r 80949837f4c1 -r ba8b7b6c9b1a mpn/powerpc64/mode64/p9/gmp-mparam.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/mpn/powerpc64/mode64/p9/gmp-mparam.h Fri Nov 30 17:35:17 2018 +0100
@@ -0,0 +1,166 @@
+/* POWER9 gmp-mparam.h -- Compiler/machine parameter header file.
+
+Copyright 2018 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
+
+/* 3700MHz POWER9 */
+
+/* Generated by tuneup.c, 2018-11-30, gcc 4.8 */
+
+#define MOD_1_NORM_THRESHOLD 0 /* always */
+#define MOD_1_UNNORM_THRESHOLD 0 /* always */
+#define MOD_1N_TO_MOD_1_1_THRESHOLD 8
+#define MOD_1U_TO_MOD_1_1_THRESHOLD 6
+#define MOD_1_1_TO_MOD_1_2_THRESHOLD 20
+#define MOD_1_2_TO_MOD_1_4_THRESHOLD 0 /* never mpn_mod_1s_2p */
+#define PREINV_MOD_1_TO_MOD_1_THRESHOLD 12
+#define USE_PREINV_DIVREM_1 0
+#define DIV_QR_1N_PI1_METHOD 2
+#define DIV_QR_1_NORM_THRESHOLD 3
+#define DIV_QR_1_UNNORM_THRESHOLD 2
+#define DIV_QR_2_PI2_THRESHOLD 13
+#define DIVEXACT_1_THRESHOLD 0 /* always (native) */
+#define BMOD_1_TO_MOD_1_THRESHOLD 21
+
+#define DIV_1_VS_MUL_1_PERCENT 359
+
+#define MUL_TOOM22_THRESHOLD 34
+#define MUL_TOOM33_THRESHOLD 105
+#define MUL_TOOM44_THRESHOLD 450
+#define MUL_TOOM6H_THRESHOLD 517
+#define MUL_TOOM8H_THRESHOLD 608
+
+#define MUL_TOOM32_TO_TOOM43_THRESHOLD 114
+#define MUL_TOOM32_TO_TOOM53_THRESHOLD 279
+#define MUL_TOOM42_TO_TOOM53_THRESHOLD 204
+#define MUL_TOOM42_TO_TOOM63_THRESHOLD 211
+#define MUL_TOOM43_TO_TOOM54_THRESHOLD 185
+
+#define SQR_BASECASE_THRESHOLD 0 /* always (native) */
+#define SQR_TOOM2_THRESHOLD 46
+#define SQR_TOOM3_THRESHOLD 158
+#define SQR_TOOM4_THRESHOLD 686
+#define SQR_TOOM6_THRESHOLD 0 /* always */
+#define SQR_TOOM8_THRESHOLD 866
+
+#define MULMID_TOOM42_THRESHOLD 66
+
+#define MULMOD_BNM1_THRESHOLD 19
+#define SQRMOD_BNM1_THRESHOLD 26
+
+#define MUL_FFT_MODF_THRESHOLD 530 /* k = 5 */
+#define MUL_FFT_TABLE3 \
+ { { 530, 5}, { 25, 6}, { 13, 5}, { 27, 6}, \
+ { 27, 7}, { 14, 6}, { 29, 7}, { 15, 6}, \
+ { 31, 7}, { 25, 8}, { 13, 7}, { 28, 8}, \
+ { 15, 7}, { 32, 8}, { 17, 7}, { 36, 8}, \
+ { 19, 7}, { 39, 8}, { 27, 9}, { 15, 8}, \
+ { 35, 9}, { 19, 8}, { 41, 9}, { 23, 8}, \
+ { 49, 9}, { 27,10}, { 15, 9}, { 31, 8}, \
+ { 63, 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}, \
+ { 143,10}, { 287,11}, { 159,12}, { 95,11}, \
+ { 191,13}, { 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 68
+#define MUL_FFT_THRESHOLD 5248
+
+#define SQR_FFT_MODF_THRESHOLD 496 /* k = 5 */
+#define SQR_FFT_TABLE3 \
+ { { 496, 5}, { 23, 6}, { 12, 5}, { 25, 6}, \
+ { 13, 5}, { 27, 6}, { 29, 7}, { 15, 6}, \
+ { 31, 7}, { 16, 6}, { 33, 7}, { 28, 8}, \
+ { 15, 7}, { 32, 8}, { 17, 7}, { 36, 8}, \
+ { 19, 7}, { 39, 8}, { 29, 9}, { 15, 8}, \
+ { 35, 9}, { 19, 8}, { 41, 9}, { 23, 8}, \
+ { 47, 9}, { 27,10}, { 15, 9}, { 31, 8}, \
+ { 63, 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}, { 95,12}, { 63,11}, { 127,10}, \
+ { 255,11}, { 143,10}, { 287,11}, { 159,12}, \
+ { 95,11}, { 191,13}, { 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 69
+#define SQR_FFT_THRESHOLD 4224
+
+#define MULLO_BASECASE_THRESHOLD 3
+#define MULLO_DC_THRESHOLD 40
+#define MULLO_MUL_N_THRESHOLD 10323
+#define SQRLO_BASECASE_THRESHOLD 6
+#define SQRLO_DC_THRESHOLD 40
+#define SQRLO_SQR_THRESHOLD 8397
+
+#define DC_DIV_QR_THRESHOLD 30
+#define DC_DIVAPPR_Q_THRESHOLD 91
+#define DC_BDIV_QR_THRESHOLD 36
+#define DC_BDIV_Q_THRESHOLD 63
+
+#define INV_MULMOD_BNM1_THRESHOLD 78
+#define INV_NEWTON_THRESHOLD 11
+#define INV_APPR_THRESHOLD 13
+
+#define BINV_NEWTON_THRESHOLD 272
+#define REDC_1_TO_REDC_2_THRESHOLD 8
+#define REDC_2_TO_REDC_N_THRESHOLD 79
+
+#define MU_DIV_QR_THRESHOLD 1589
+#define MU_DIVAPPR_Q_THRESHOLD 1589
+#define MUPI_DIV_QR_THRESHOLD 0 /* always */
+#define MU_BDIV_QR_THRESHOLD 1652
+#define MU_BDIV_Q_THRESHOLD 2089
+
+#define POWM_SEC_TABLE 3,16,139,852
+
+#define GET_STR_DC_THRESHOLD 9
+#define GET_STR_PRECOMPUTE_THRESHOLD 17
+#define SET_STR_DC_THRESHOLD 406
+#define SET_STR_PRECOMPUTE_THRESHOLD 956
+
More information about the gmp-commit
mailing list