| 1 |
dnl ARM Neon mpn_com optimised for A15. |
| 2 |
|
| 3 |
dnl Copyright 2013 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 cycles/limb |
| 34 |
C StrongARM ? |
| 35 |
C XScale ? |
| 36 |
C Cortex-A8 ? |
| 37 |
C Cortex-A9 2.1 |
| 38 |
C Cortex-A15 0.65 |
| 39 |
|
| 40 |
define(`rp', `r0') |
| 41 |
define(`up', `r1') |
| 42 |
define(`n', `r2') |
| 43 |
|
| 44 |
ASM_START() |
| 45 |
PROLOGUE(mpn_com) |
| 46 |
cmp n, #7 |
| 47 |
ble L(bc) |
| 48 |
|
| 49 |
C Perform a few initial operation until rp is 128-bit aligned |
| 50 |
tst rp, #4 |
| 51 |
beq L(al1) |
| 52 |
vld1.32 {d0[0]}, [up]! |
| 53 |
sub n, n, #1 |
| 54 |
vmvn d0, d0 |
| 55 |
vst1.32 {d0[0]}, [rp]! |
| 56 |
L(al1): tst rp, #8 |
| 57 |
beq L(al2) |
| 58 |
vld1.32 {d0}, [up]! |
| 59 |
sub n, n, #2 |
| 60 |
vmvn d0, d0 |
| 61 |
vst1.32 {d0}, [rp:64]! |
| 62 |
L(al2): vld1.32 {q2}, [up]! |
| 63 |
subs n, n, #12 |
| 64 |
blt L(end) |
| 65 |
|
| 66 |
ALIGN(16) |
| 67 |
L(top): vld1.32 {q0}, [up]! |
| 68 |
vmvn q2, q2 |
| 69 |
subs n, n, #8 |
| 70 |
vst1.32 {q2}, [rp:128]! |
| 71 |
vld1.32 {q2}, [up]! |
| 72 |
vmvn q0, q0 |
| 73 |
vst1.32 {q0}, [rp:128]! |
| 74 |
bge L(top) |
| 75 |
|
| 76 |
L(end): vmvn q2, q2 |
| 77 |
vst1.32 {q2}, [rp:128]! |
| 78 |
|
| 79 |
C Handle last 0-7 limbs. Note that rp is aligned after loop, but not when we |
| 80 |
C arrive here via L(bc) |
| 81 |
L(bc): tst n, #4 |
| 82 |
beq L(tl1) |
| 83 |
vld1.32 {q0}, [up]! |
| 84 |
vmvn q0, q0 |
| 85 |
vst1.32 {q0}, [rp]! |
| 86 |
L(tl1): tst n, #2 |
| 87 |
beq L(tl2) |
| 88 |
vld1.32 {d0}, [up]! |
| 89 |
vmvn d0, d0 |
| 90 |
vst1.32 {d0}, [rp]! |
| 91 |
L(tl2): tst n, #1 |
| 92 |
beq L(tl3) |
| 93 |
vld1.32 {d0[0]}, [up] |
| 94 |
vmvn d0, d0 |
| 95 |
vst1.32 {d0[0]}, [rp] |
| 96 |
L(tl3): bx lr |
| 97 |
EPILOGUE() |