| 1 |
dnl Alpha ev6 nails mpn_addmul_2. |
| 2 |
|
| 3 |
dnl Copyright 2002, 2005, 2006 Free Software Foundation, Inc. |
| 4 |
|
| 5 |
dnl This file is part of the GNU MP Library. |
| 6 |
dnl |
| 7 |
dnl The GNU MP Library is free software; you can redistribute it and/or modify |
| 8 |
dnl it under the terms of either: |
| 9 |
dnl |
| 10 |
dnl * the GNU Lesser General Public License as published by the Free |
| 11 |
dnl Software Foundation; either version 3 of the License, or (at your |
| 12 |
dnl option) any later version. |
| 13 |
dnl |
| 14 |
dnl or |
| 15 |
dnl |
| 16 |
dnl * the GNU General Public License as published by the Free Software |
| 17 |
dnl Foundation; either version 2 of the License, or (at your option) any |
| 18 |
dnl later version. |
| 19 |
dnl |
| 20 |
dnl or both in parallel, as here. |
| 21 |
dnl |
| 22 |
dnl The GNU MP Library is distributed in the hope that it will be useful, but |
| 23 |
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 24 |
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
| 25 |
dnl for more details. |
| 26 |
dnl |
| 27 |
dnl You should have received copies of the GNU General Public License and the |
| 28 |
dnl GNU Lesser General Public License along with the GNU MP Library. If not, |
| 29 |
dnl see https://www.gnu.org/licenses/. |
| 30 |
|
| 31 |
include(`../config.m4') |
| 32 |
|
| 33 |
C Runs at 4.0 cycles/limb. |
| 34 |
|
| 35 |
C We could either go for 2-way unrolling over 11 cycles, or 2.75 c/l, |
| 36 |
C or 4-way unrolling over 20 cycles, for 2.5 c/l. |
| 37 |
|
| 38 |
|
| 39 |
C INPUT PARAMETERS |
| 40 |
define(`rp',`r16') |
| 41 |
define(`up',`r17') |
| 42 |
define(`n',`r18') |
| 43 |
define(`vp',`r19') |
| 44 |
|
| 45 |
C Useful register aliases |
| 46 |
define(`numb_mask',`r24') |
| 47 |
define(`ulimb',`r25') |
| 48 |
define(`rlimb',`r27') |
| 49 |
|
| 50 |
define(`m0a',`r0') |
| 51 |
define(`m0b',`r1') |
| 52 |
define(`m1a',`r2') |
| 53 |
define(`m1b',`r3') |
| 54 |
|
| 55 |
define(`acc0',`r4') |
| 56 |
define(`acc1',`r5') |
| 57 |
|
| 58 |
define(`v0',`r6') |
| 59 |
define(`v1',`r7') |
| 60 |
|
| 61 |
C Used for temps: r8 r19 r28 |
| 62 |
|
| 63 |
define(`NAIL_BITS',`GMP_NAIL_BITS') |
| 64 |
define(`NUMB_BITS',`GMP_NUMB_BITS') |
| 65 |
|
| 66 |
C This declaration is munged by configure |
| 67 |
NAILS_SUPPORT(3-63) |
| 68 |
|
| 69 |
ASM_START() |
| 70 |
PROLOGUE(mpn_addmul_2) |
| 71 |
lda numb_mask,-1(r31) |
| 72 |
srl numb_mask,NAIL_BITS,numb_mask |
| 73 |
|
| 74 |
ldq v0, 0(vp) |
| 75 |
ldq v1, 8(vp) |
| 76 |
|
| 77 |
bis r31, r31, acc0 C zero acc0 |
| 78 |
sll v0,NAIL_BITS, v0 |
| 79 |
bis r31, r31, acc1 C zero acc1 |
| 80 |
sll v1,NAIL_BITS, v1 |
| 81 |
bis r31, r31, r19 |
| 82 |
|
| 83 |
ldq ulimb, 0(up) |
| 84 |
lda up, 8(up) |
| 85 |
mulq v0, ulimb, m0a C U1 |
| 86 |
umulh v0, ulimb, m0b C U1 |
| 87 |
mulq v1, ulimb, m1a C U1 |
| 88 |
umulh v1, ulimb, m1b C U1 |
| 89 |
lda n, -1(n) |
| 90 |
beq n, L(end) C U0 |
| 91 |
|
| 92 |
ALIGN(16) |
| 93 |
L(top): bis r31, r31, r31 C U1 nop |
| 94 |
addq r19, acc0, acc0 C U0 propagate nail |
| 95 |
ldq rlimb, 0(rp) C L0 |
| 96 |
ldq ulimb, 0(up) C L1 |
| 97 |
|
| 98 |
lda rp, 8(rp) C L1 |
| 99 |
srl m0a,NAIL_BITS, r8 C U0 |
| 100 |
lda up, 8(up) C L0 |
| 101 |
mulq v0, ulimb, m0a C U1 |
| 102 |
|
| 103 |
addq r8, acc0, r19 C U0 |
| 104 |
addq m0b, acc1, acc0 C L1 |
| 105 |
umulh v0, ulimb, m0b C U1 |
| 106 |
bis r31, r31, r31 C L0 nop |
| 107 |
|
| 108 |
addq rlimb, r19, r19 C L1 FINAL PROD-SUM |
| 109 |
srl m1a,NAIL_BITS, r8 C U0 |
| 110 |
lda n, -1(n) C L0 |
| 111 |
mulq v1, ulimb, m1a C U1 |
| 112 |
|
| 113 |
addq r8, acc0, acc0 C U0 |
| 114 |
bis r31, m1b, acc1 C L1 |
| 115 |
umulh v1, ulimb, m1b C U1 |
| 116 |
and r19,numb_mask, r28 C L0 extract numb part |
| 117 |
|
| 118 |
unop |
| 119 |
srl r19,NUMB_BITS, r19 C U1 extract nail part |
| 120 |
stq r28, -8(rp) C L1 |
| 121 |
bne n, L(top) C U0 |
| 122 |
|
| 123 |
L(end): ldq rlimb, 0(rp) |
| 124 |
addq r19, acc0, acc0 C propagate nail |
| 125 |
lda rp, 8(rp) |
| 126 |
srl m0a,NAIL_BITS, r8 C U0 |
| 127 |
addq r8, acc0, r19 |
| 128 |
addq m0b, acc1, acc0 |
| 129 |
addq rlimb, r19, r19 |
| 130 |
srl m1a,NAIL_BITS, r8 C U0 |
| 131 |
addq r8, acc0, acc0 |
| 132 |
bis r31, m1b, acc1 |
| 133 |
and r19,numb_mask, r28 C extract limb |
| 134 |
|
| 135 |
srl r19,NUMB_BITS, r19 C extract nail |
| 136 |
stq r28, -8(rp) |
| 137 |
|
| 138 |
addq r19, acc0, acc0 C propagate nail |
| 139 |
and acc0,numb_mask, r28 |
| 140 |
stq r28, 0(rp) |
| 141 |
srl acc0,NUMB_BITS, r19 |
| 142 |
addq r19, acc1, r0 |
| 143 |
|
| 144 |
ret r31, (r26), 1 |
| 145 |
EPILOGUE() |
| 146 |
ASM_END() |