| 1 |
dnl x86-64 mpn_lshift optimized for "Core 2". |
| 2 |
|
| 3 |
dnl Copyright 2007, 2009, 2011, 2012 Free Software Foundation, Inc. |
| 4 |
|
| 5 |
dnl This file is part of the GNU MP Library. |
| 6 |
dnl |
| 7 |
dnl The GNU MP Library is free software; you can redistribute it and/or modify |
| 8 |
dnl it under the terms of either: |
| 9 |
dnl |
| 10 |
dnl * the GNU Lesser General Public License as published by the Free |
| 11 |
dnl Software Foundation; either version 3 of the License, or (at your |
| 12 |
dnl option) any later version. |
| 13 |
dnl |
| 14 |
dnl or |
| 15 |
dnl |
| 16 |
dnl * the GNU General Public License as published by the Free Software |
| 17 |
dnl Foundation; either version 2 of the License, or (at your option) any |
| 18 |
dnl later version. |
| 19 |
dnl |
| 20 |
dnl or both in parallel, as here. |
| 21 |
dnl |
| 22 |
dnl The GNU MP Library is distributed in the hope that it will be useful, but |
| 23 |
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 24 |
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
| 25 |
dnl for more details. |
| 26 |
dnl |
| 27 |
dnl You should have received copies of the GNU General Public License and the |
| 28 |
dnl GNU Lesser General Public License along with the GNU MP Library. If not, |
| 29 |
dnl see https://www.gnu.org/licenses/. |
| 30 |
|
| 31 |
include(`../config.m4') |
| 32 |
|
| 33 |
|
| 34 |
C cycles/limb |
| 35 |
C AMD K8,K9 4.25 |
| 36 |
C AMD K10 4.25 |
| 37 |
C Intel P4 14.7 |
| 38 |
C Intel core2 1.27 |
| 39 |
C Intel NHM 1.375 (up to about n = 260, then 1.5) |
| 40 |
C Intel SBR 1.87 |
| 41 |
C Intel atom ? |
| 42 |
C VIA nano ? |
| 43 |
|
| 44 |
|
| 45 |
C INPUT PARAMETERS |
| 46 |
define(`rp', `%rdi') |
| 47 |
define(`up', `%rsi') |
| 48 |
define(`n', `%rdx') |
| 49 |
define(`cnt', `%rcx') |
| 50 |
|
| 51 |
ABI_SUPPORT(DOS64) |
| 52 |
ABI_SUPPORT(STD64) |
| 53 |
|
| 54 |
ASM_START() |
| 55 |
TEXT |
| 56 |
ALIGN(16) |
| 57 |
PROLOGUE(mpn_lshift) |
| 58 |
FUNC_ENTRY(4) |
| 59 |
lea -8(rp,n,8), rp |
| 60 |
lea -8(up,n,8), up |
| 61 |
|
| 62 |
mov R32(%rdx), R32(%rax) |
| 63 |
and $3, R32(%rax) |
| 64 |
jne L(nb00) |
| 65 |
L(b00): C n = 4, 8, 12, ... |
| 66 |
mov (up), %r10 |
| 67 |
mov -8(up), %r11 |
| 68 |
xor R32(%rax), R32(%rax) |
| 69 |
shld R8(cnt), %r10, %rax |
| 70 |
mov -16(up), %r8 |
| 71 |
lea 24(rp), rp |
| 72 |
sub $4, n |
| 73 |
jmp L(00) |
| 74 |
|
| 75 |
L(nb00):C n = 1, 5, 9, ... |
| 76 |
cmp $2, R32(%rax) |
| 77 |
jae L(nb01) |
| 78 |
L(b01): mov (up), %r9 |
| 79 |
xor R32(%rax), R32(%rax) |
| 80 |
shld R8(cnt), %r9, %rax |
| 81 |
sub $2, n |
| 82 |
jb L(le1) |
| 83 |
mov -8(up), %r10 |
| 84 |
mov -16(up), %r11 |
| 85 |
lea -8(up), up |
| 86 |
lea 16(rp), rp |
| 87 |
jmp L(01) |
| 88 |
L(le1): shl R8(cnt), %r9 |
| 89 |
mov %r9, (rp) |
| 90 |
FUNC_EXIT() |
| 91 |
ret |
| 92 |
|
| 93 |
L(nb01):C n = 2, 6, 10, ... |
| 94 |
jne L(b11) |
| 95 |
L(b10): mov (up), %r8 |
| 96 |
mov -8(up), %r9 |
| 97 |
xor R32(%rax), R32(%rax) |
| 98 |
shld R8(cnt), %r8, %rax |
| 99 |
sub $3, n |
| 100 |
jb L(le2) |
| 101 |
mov -16(up), %r10 |
| 102 |
lea -16(up), up |
| 103 |
lea 8(rp), rp |
| 104 |
jmp L(10) |
| 105 |
L(le2): shld R8(cnt), %r9, %r8 |
| 106 |
mov %r8, (rp) |
| 107 |
shl R8(cnt), %r9 |
| 108 |
mov %r9, -8(rp) |
| 109 |
FUNC_EXIT() |
| 110 |
ret |
| 111 |
|
| 112 |
ALIGN(16) C performance critical! |
| 113 |
L(b11): C n = 3, 7, 11, ... |
| 114 |
mov (up), %r11 |
| 115 |
mov -8(up), %r8 |
| 116 |
xor R32(%rax), R32(%rax) |
| 117 |
shld R8(cnt), %r11, %rax |
| 118 |
mov -16(up), %r9 |
| 119 |
lea -24(up), up |
| 120 |
sub $4, n |
| 121 |
jb L(end) |
| 122 |
|
| 123 |
ALIGN(16) |
| 124 |
L(top): shld R8(cnt), %r8, %r11 |
| 125 |
mov (up), %r10 |
| 126 |
mov %r11, (rp) |
| 127 |
L(10): shld R8(cnt), %r9, %r8 |
| 128 |
mov -8(up), %r11 |
| 129 |
mov %r8, -8(rp) |
| 130 |
L(01): shld R8(cnt), %r10, %r9 |
| 131 |
mov -16(up), %r8 |
| 132 |
mov %r9, -16(rp) |
| 133 |
L(00): shld R8(cnt), %r11, %r10 |
| 134 |
mov -24(up), %r9 |
| 135 |
mov %r10, -24(rp) |
| 136 |
add $-32, up |
| 137 |
lea -32(rp), rp |
| 138 |
sub $4, n |
| 139 |
jnc L(top) |
| 140 |
|
| 141 |
L(end): shld R8(cnt), %r8, %r11 |
| 142 |
mov %r11, (rp) |
| 143 |
shld R8(cnt), %r9, %r8 |
| 144 |
mov %r8, -8(rp) |
| 145 |
shl R8(cnt), %r9 |
| 146 |
mov %r9, -16(rp) |
| 147 |
FUNC_EXIT() |
| 148 |
ret |
| 149 |
EPILOGUE() |