1 dnl ARM mpn_mul_1 optimised for A15.
3 dnl Copyright
2012, 2013 Free Software Foundation
, Inc.
5 dnl
This file is part of the GNU MP Library.
7 dnl The GNU MP Library is free software
; you can redistribute it and/or modify
8 dnl it under the terms of
either:
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.
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
20 dnl
or both
in parallel
, as here.
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
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
/.
31 include(`..
/config.m4
')
39 C Cortex-A15 2.25 this
42 C This runs well on A15 but very poorly on A9. By scheduling loads and adds
43 C it is possible to get good A9 performance as well, but at the cost of using
44 C many more (callee-saves) registers.
46 C This is armv5 code, optimized for the armv7a cpu A15. Its location in the
47 C GMP file structure might be misleading.
68 L(bx1): umull r4, r7, r6, v0
74 L(bx0): umull r4, r5, r6, v0
80 L(top): ldr r6, [up], #4
84 L(lo0): ldr r6, [up], #4
88 L(lo3): ldr r6, [up], #4
92 L(lo2): ldr r6, [up], #4
96 L(lo1): adc r7, r7, #0