1dnl PowerPC-64 mpn_copyd 2 3dnl Copyright 2004, 2005 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 POWER3/PPC630 1 35C POWER4/PPC970 1 36C POWER5 ? 37C POWER6 ? 38C POWER7 1.4 39 40C INPUT PARAMETERS 41C rp r3 42C up r4 43C n r5 44 45ASM_START() 46PROLOGUE(mpn_copyd) 47 rldic. r0, r5, 3, 59 C r0 = (r5 & 3) << 3; cr0 = (n == 4t)? 48 cmpldi cr6, r0, 16 C cr6 = (n cmp 4t + 2)? 49 50ifdef(`HAVE_ABI_mode32', 51` rldic r6, r5, 3, 32', C byte count corresponding to n 52` rldicr r6, r5, 3, 60') C byte count corresponding to n 53 54 addi r5, r5, 4 C compute... 55ifdef(`HAVE_ABI_mode32', 56` rldicl r5, r5, 62,34', C ...branch count 57` rldicl r5, r5, 62, 2') C ...branch count 58 mtctr r5 59 60 add r4, r4, r6 61 add r3, r3, r6 62 sub r4, r4, r0 C offset up 63 sub r3, r3, r0 C offset rp 64 65 beq cr0, L(L00) 66 blt cr6, L(L01) 67 beq cr6, L(L10) 68 b L(L11) 69 70 ALIGN(16) 71L(oop): ld r6, 24(r4) 72 std r6, 24(r3) 73L(L11): ld r6, 16(r4) 74 std r6, 16(r3) 75L(L10): ld r6, 8(r4) 76 std r6, 8(r3) 77L(L01): ld r6, 0(r4) 78 std r6, 0(r3) 79L(L00): addi r4, r4, -32 80 addi r3, r3, -32 81 bdnz L(oop) 82 83 blr 84EPILOGUE() 85