1dnl Alpha mpn_sqr_diag_addlsh1. 2 3dnl Copyright 2013 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 EV4: ? 35C EV5: 10.2 36C EV6: 4.5 37 38C Ideally, one-way code could run at 9 c/l (limited by mulq+umulh) on ev5 and 39C about 3.75 c/l on ev6. Two-way code could run at about 3.25 c/l on ev6. 40 41C Algorithm: We allow ourselves to propagate carry to a product high word 42C without worrying for carry out, since (B-1)^2 = B^2-2B+1 has a high word of 43C B-2, i.e, will not spill. We propagate carry similarly to a product low word 44C since the problem value B-1 is a quadratic non-residue mod B, but our 45C products are squares. 46 47define(`rp', `r16') 48define(`tp', `r17') 49define(`up', `r18') 50define(`n', `r19') 51 52ASM_START() 53PROLOGUE(mpn_sqr_diag_addlsh1) 54 ldq r0, 0(up) 55 bis r31, r31, r21 56 bis r31, r31, r3 57 mulq r0, r0, r7 58 stq r7, 0(rp) 59 umulh r0, r0, r6 60 lda n, -1(n) 61 62 ALIGN(16) 63L(top): ldq r0, 8(up) 64 lda up, 8(up) 65 ldq r8, 0(tp) 66 ldq r20, 8(tp) 67 mulq r0, r0, r7 68 lda tp, 16(tp) 69 sll r8, 1, r23 70 srl r8, 63, r22 71 or r21, r23, r23 72 sll r20, 1, r24 73 addq r3, r6, r6 C cannot carry per comment above 74 or r22, r24, r24 75 addq r23, r6, r21 76 umulh r0, r0, r6 77 cmpult r21, r23, r1 78 addq r1, r7, r7 C cannot carry per comment above 79 stq r21, 8(rp) 80 addq r24, r7, r22 81 stq r22, 16(rp) 82 lda n, -1(n) 83 cmpult r22, r7, r3 84 srl r20, 63, r21 85 lda rp, 16(rp) 86 bne n, L(top) 87 88 addq r3, r6, r6 C cannot carry per comment above 89 addq r21, r6, r21 90 stq r21, 8(rp) 91 ret r31, (r26), 1 92EPILOGUE() 93ASM_END() 94