1dnl ARM mpn_mul_2. 2 3dnl Contributed to the GNU project by Torbjörn Granlund. 4 5dnl Copyright 2012 Free Software Foundation, Inc. 6 7dnl This file is part of the GNU MP Library. 8dnl 9dnl The GNU MP Library is free software; you can redistribute it and/or modify 10dnl it under the terms of either: 11dnl 12dnl * the GNU Lesser General Public License as published by the Free 13dnl Software Foundation; either version 3 of the License, or (at your 14dnl option) any later version. 15dnl 16dnl or 17dnl 18dnl * the GNU General Public License as published by the Free Software 19dnl Foundation; either version 2 of the License, or (at your option) any 20dnl later version. 21dnl 22dnl or both in parallel, as here. 23dnl 24dnl The GNU MP Library is distributed in the hope that it will be useful, but 25dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY 26dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 27dnl for more details. 28dnl 29dnl You should have received copies of the GNU General Public License and the 30dnl GNU Lesser General Public License along with the GNU MP Library. If not, 31dnl see https://www.gnu.org/licenses/. 32 33include(`../config.m4') 34 35C cycles/limb 36C StrongARM: - 37C XScale - 38C ARM11 5.25 39C Cortex-A5 3.63 40C Cortex-A7 3.15 41C Cortex-A8 5.0 42C Cortex-A9 2.25 43C Cortex-A15 2.5 44C Cortex-A17 2.13 45C Cortex-A53 3.5 46 47C TODO 48C * This is a trivial edit of the addmul_2 code. Check for simplifications, 49C and possible speedups to 2.0 c/l. 50 51define(`rp',`r0') 52define(`up',`r1') 53define(`n', `r2') 54define(`vp',`r3') 55 56define(`v0',`r6') 57define(`v1',`r7') 58define(`u0',`r3') 59define(`u1',`r9') 60 61define(`cya',`r8') 62define(`cyb',`r12') 63 64 65ASM_START() 66PROLOGUE(mpn_mul_2) 67 push { r4, r5, r6, r7, r8, r9 } 68 69 ldm vp, { v0, v1 } 70 mov cya, #0 71 mov cyb, #0 72 73 tst n, #1 74 beq L(evn) 75L(odd): mov r5, #0 76 ldr u0, [up, #0] 77 mov r4, #0 78 tst n, #2 79 beq L(fi1) 80L(fi3): sub up, up, #12 81 sub rp, rp, #16 82 b L(lo3) 83L(fi1): sub n, n, #1 84 sub up, up, #4 85 sub rp, rp, #8 86 b L(lo1) 87L(evn): mov r4, #0 88 ldr u1, [up, #0] 89 mov r5, #0 90 tst n, #2 91 bne L(fi2) 92L(fi0): sub up, up, #8 93 sub rp, rp, #12 94 b L(lo0) 95L(fi2): subs n, n, #2 96 sub rp, rp, #4 97 bls L(end) 98 99 ALIGN(16) 100L(top): ldr u0, [up, #4] 101 umaal r4, cya, u1, v0 102 str r4, [rp, #4] 103 mov r4, #0 104 umaal r5, cyb, u1, v1 105L(lo1): ldr u1, [up, #8] 106 umaal r5, cya, u0, v0 107 str r5, [rp, #8] 108 mov r5, #0 109 umaal r4, cyb, u0, v1 110L(lo0): ldr u0, [up, #12] 111 umaal r4, cya, u1, v0 112 str r4, [rp, #12] 113 mov r4, #0 114 umaal r5, cyb, u1, v1 115L(lo3): ldr u1, [up, #16]! 116 umaal r5, cya, u0, v0 117 str r5, [rp, #16]! 118 mov r5, #0 119 umaal r4, cyb, u0, v1 120 subs n, n, #4 121 bhi L(top) 122 123L(end): umaal r4, cya, u1, v0 124 ldr u0, [up, #4] 125 umaal r5, cyb, u1, v1 126 str r4, [rp, #4] 127 umaal r5, cya, u0, v0 128 umaal cya, cyb, u0, v1 129 str r5, [rp, #8] 130 str cya, [rp, #12] 131 mov r0, cyb 132 133 pop { r4, r5, r6, r7, r8, r9 } 134 bx r14 135EPILOGUE() 136