1dnl PowerPC-64 mpn_modexact_1_odd -- mpn by limb exact remainder. 2 3dnl Copyright 2006 Free Software Foundation, Inc. 4 5dnl This file is part of the GNU MP Library. 6 7dnl The GNU MP Library is free software; you can redistribute it and/or modify 8dnl it under the terms of the GNU Lesser General Public License as published 9dnl by the Free Software Foundation; either version 3 of the License, or (at 10dnl your option) any later version. 11 12dnl The GNU MP Library is distributed in the hope that it will be useful, but 13dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY 14dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public 15dnl License for more details. 16 17dnl You should have received a copy of the GNU Lesser General Public License 18dnl along with the GNU MP Library. If not, see http://www.gnu.org/licenses/. 19 20include(`../config.m4') 21 22C cycles/limb 23C POWER3/PPC630 13-19 24C POWER4/PPC970 16 25C POWER5 16 26C POWER6 ? 27C POWER7 12 28 29C TODO 30C * Check if n=1 code is really an improvement. It probably isn't. 31C * Make more similar to dive_1.asm. 32 33C INPUT PARAMETERS 34define(`up', `r3') 35define(`n', `r4') 36define(`d', `r5') 37define(`cy', `r6') 38 39 40ASM_START() 41 42EXTERN(binvert_limb_table) 43 44PROLOGUE(mpn_modexact_1c_odd) 45 addic. n, n, -1 C set carry as side effect 46 ld r8, 0(up) 47 bne cr0, L(2) 48 cmpld cr7, r6, r8 49 bge cr7, L(4) 50 subf r8, r6, r8 51 divdu r3, r8, d 52 mulld r3, r3, d 53 subf. r3, r3, r8 54 beqlr cr0 55 subf r3, r3, d 56 blr 57 58L(4): subf r3, r8, r6 59 divdu r8, r3, d 60 mulld r8, r8, d 61 subf r3, r8, r3 62 blr 63 64L(2): LEA( r7, binvert_limb_table) 65 rldicl r9, d, 63, 57 66 mtctr n 67 lbzx r0, r7, r9 68 mulld r7, r0, r0 69 sldi r0, r0, 1 70 mulld r7, d, r7 71 subf r0, r7, r0 72 mulld r9, r0, r0 73 sldi r0, r0, 1 74 mulld r9, d, r9 75 subf r0, r9, r0 76 mulld r7, r0, r0 77 sldi r0, r0, 1 78 mulld r7, d, r7 79 subf r9, r7, r0 80 81 ALIGN(16) 82L(loop): 83 subfe r0, r6, r8 84 ld r8, 8(up) 85 addi up, up, 8 86 mulld r0, r9, r0 87 mulhdu r6, r0, d 88 bdnz L(loop) 89 90 cmpld cr7, d, r8 91 blt cr7, L(10) 92 93 subfe r0, r0, r0 94 subf r6, r0, r6 95 cmpld cr7, r6, r8 96 subf r3, r8, r6 97 bgelr cr7 98 add r3, d, r3 99 blr 100 101L(10): subfe r0, r6, r8 102 mulld r0, r9, r0 103 mulhdu r3, r0, d 104 blr 105EPILOGUE() 106ASM_END() 107