1dnl AMD64 mpn_copyi optimised for AMD bobcat. 2 3dnl Copyright 2003, 2005, 2007, 2011, 2012 Free Software Foundation, Inc. 4 5dnl This file is part of the GNU MP Library. 6dnl 7dnl The GNU MP Library is free software; you can redistribute it and/or modify 8dnl it under the terms of either: 9dnl 10dnl * the GNU Lesser General Public License as published by the Free 11dnl Software Foundation; either version 3 of the License, or (at your 12dnl option) any later version. 13dnl 14dnl or 15dnl 16dnl * the GNU General Public License as published by the Free Software 17dnl Foundation; either version 2 of the License, or (at your option) any 18dnl later version. 19dnl 20dnl or both in parallel, as here. 21dnl 22dnl The GNU MP Library is distributed in the hope that it will be useful, but 23dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY 24dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 25dnl for more details. 26dnl 27dnl You should have received copies of the GNU General Public License and the 28dnl GNU Lesser General Public License along with the GNU MP Library. If not, 29dnl see https://www.gnu.org/licenses/. 30 31include(`../config.m4') 32 33C cycles/limb 34C AMD K8,K9 1 35C AMD K10 1-2 (alignment fluctuations) 36C AMD bd1 ? 37C AMD bobcat 1.5 38C Intel P4 2.8 39C Intel core2 1 40C Intel NHM 1-1.25 41C Intel SBR 1 42C Intel atom 2.87 43C VIA nano 2 44 45C INPUT PARAMETERS 46C rp rdi 47C up rsi 48C n rdx 49 50define(`rp',`%rdi') 51define(`up',`%rsi') 52define(`n',`%rdx') 53 54ABI_SUPPORT(DOS64) 55ABI_SUPPORT(STD64) 56 57ASM_START() 58 TEXT 59 ALIGN(16) 60PROLOGUE(mpn_copyi) 61 FUNC_ENTRY(3) 62 lea -32(up,n,8), up 63 lea -32(rp,n,8), rp 64 neg n 65 add $4, n 66 jg L(end) 67 ALIGN(16) 68L(top): mov (up,n,8), %r8 69 mov %r8, (rp,n,8) 70 mov 8(up,n,8), %r8 71 mov %r8, 8(rp,n,8) 72 mov 16(up,n,8), %r8 73 mov %r8, 16(rp,n,8) 74 mov 24(up,n,8), %r8 75 mov %r8, 24(rp,n,8) 76L(ent): add $4, n 77 jle L(top) 78 79L(end): cmp $4, R32(n) 80 jz L(ret) 81 mov (up,n,8), %r8 82 mov %r8, (rp,n,8) 83 cmp $3, R32(n) 84 jz L(ret) 85 mov 8(up,n,8), %r8 86 mov %r8, 8(rp,n,8) 87 cmp $2, R32(n) 88 jz L(ret) 89 mov 16(up,n,8), %r8 90 mov %r8, 16(rp,n,8) 91 92L(ret): FUNC_EXIT() 93 ret 94EPILOGUE() 95