source/libs/gmp/gmp-src/mpn/x86_64/bobcat/mul_1.asm

   1 dnl  AMD64 mpn_mul_1 optimised for AMD bobcat.
   2
   3 dnl  Copyright 2003-2005, 2007, 2008, 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 C            cycles/limb
  34 C AMD K8,K9      4.5
  35 C AMD K10        4.5
  36 C AMD bd1        4.62
  37 C AMD bobcat     5
  38 C Intel P4      14
  39 C Intel core2    4.5
  40 C Intel NHM      4.23
  41 C Intel SBR      3.0
  42 C Intel atom    21
  43 C VIA nano       4.94
  44
  45 C The loop of this code is the result of running a code generation and
  46 C optimisation tool suite written by David Harvey and Torbjorn Granlund.
  47
  48 ABI_SUPPORT(DOS64)
  49 ABI_SUPPORT(STD64)
  50
  51 C Standard parameters
  52 define(`rp',              `%rdi')
  53 define(`up',              `%rsi')
  54 define(`n_param',         `%rdx')
  55 define(`v0',              `%rcx')
  56 define(`cy',              `%r8')
  57 C Standard allocations
  58 define(`n',               `%rbx')
  59 define(`w0',              `%r8')
  60 define(`w1',              `%r9')
  61 define(`w2',              `%r10')
  62 define(`w3',              `%r11')
  63
  64 C DOS64 parameters
  65 IFDOS(` define(`rp',      `%rcx')    ') dnl
  66 IFDOS(` define(`up',      `%rsi')    ') dnl
  67 IFDOS(` define(`n_param', `%r8')     ') dnl
  68 IFDOS(` define(`v0',      `%r9')     ') dnl
  69 IFDOS(` define(`cy',      `56(%rsp)')') dnl
  70 C DOS64 allocations
  71 IFDOS(` define(`n',       `%rbx')    ') dnl
  72 IFDOS(` define(`w0',      `%r8')     ') dnl
  73 IFDOS(` define(`w1',      `%rdi')    ') dnl
  74 IFDOS(` define(`w2',      `%r10')    ') dnl
  75 IFDOS(` define(`w3',      `%r11')    ') dnl
  76
  77 ASM_START()
  78         TEXT
  79         ALIGN(16)
  80 PROLOGUE(mpn_mul_1c)
  81 IFDOS(` push    %rsi            ')
  82 IFDOS(` push    %rdi            ')
  83 IFDOS(` mov     %rdx, %rsi      ')
  84         mov     cy, w2
  85         jmp     L(com)
  86 EPILOGUE()
  87
  88 PROLOGUE(mpn_mul_1)
  89 IFDOS(` push    %rsi            ')
  90 IFDOS(` push    %rdi            ')
  91 IFDOS(` mov     %rdx, %rsi      ')
  92         xor     w2, w2
  93 L(com): push    %rbx
  94         mov     (up), %rax
  95
  96         lea     -16(rp,n_param,8), rp
  97         lea     -16(up,n_param,8), up
  98
  99         mov     n_param, n
 100         and     $3, R32(n_param)
 101         jz      L(b0)
 102         cmp     $2, R32(n_param)
 103         ja      L(b3)
 104         jz      L(b2)
 105
 106 L(b1):  mul     v0
 107         cmp     $1, n
 108         jz      L(n1)
 109         neg     n
 110         add     $3, n
 111         add     %rax, w2
 112         mov     %rdx, w3
 113         jmp     L(L1)
 114 L(n1):  add     %rax, w2
 115         mov     %rdx, %rax
 116         mov     w2, 8(rp)
 117         adc     $0, %rax
 118         pop     %rbx
 119 IFDOS(` pop     %rdi            ')
 120 IFDOS(` pop     %rsi            ')
 121         ret
 122
 123 L(b3):  mul     v0
 124         neg     n
 125         inc     n
 126         add     %rax, w2
 127         mov     %rdx, w3
 128         jmp     L(L3)
 129
 130 L(b0):  mul     v0
 131         mov     %rax, w0
 132         mov     %rdx, w1
 133         neg     n
 134         add     $2, n
 135         add     w2, w0
 136         jmp     L(L0)
 137
 138 L(b2):  mul     v0
 139         mov     %rax, w0
 140         mov     %rdx, w1
 141         neg     n
 142         add     w2, w0
 143         jmp     L(L2)
 144
 145         ALIGN(16)
 146 L(top): mov     w0, -16(rp,n,8)
 147         add     w1, w2
 148 L(L1):  adc     $0, w3
 149         mov     0(up,n,8), %rax
 150         mul     v0
 151         mov     %rax, w0
 152         mov     %rdx, w1
 153         mov     w2, -8(rp,n,8)
 154         add     w3, w0
 155 L(L0):  adc     $0, w1
 156         mov     8(up,n,8), %rax
 157         mul     v0
 158         mov     %rax, w2
 159         mov     %rdx, w3
 160         mov     w0, 0(rp,n,8)
 161         add     w1, w2
 162 L(L3):  adc     $0, w3
 163         mov     16(up,n,8), %rax
 164         mul     v0
 165         mov     %rax, w0
 166         mov     %rdx, w1
 167         mov     w2, 8(rp,n,8)
 168         add     w3, w0
 169 L(L2):  adc     $0, w1
 170         mov     24(up,n,8), %rax
 171         mul     v0
 172         mov     %rax, w2
 173         mov     %rdx, w3
 174         add     $4, n
 175         js      L(top)
 176
 177 L(end): mov     w0, (rp)
 178         add     w1, w2
 179         adc     $0, w3
 180         mov     w2, 8(rp)
 181         mov     w3, %rax
 182
 183         pop     %rbx
 184 IFDOS(` pop     %rdi            ')
 185 IFDOS(` pop     %rsi            ')
 186         ret
 187 EPILOGUE()