1dnl  PowerPC-64 mpn_hamdist.
2
3dnl  Copyright 2012, 2017 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 POWER3/PPC630          -
35C POWER4/PPC970          -
36C POWER5                 -
37C POWER6                 -
38C POWER7                 2.87
39
40define(`up', r3)
41define(`vp', r4)
42define(`n',  r5)
43
44ASM_START()
45PROLOGUE(mpn_hamdist)
46	std	r30, -16(r1)
47	std	r31, -8(r1)
48
49	addi	r0, n, 1
50ifdef(`HAVE_ABI_mode32',
51`	rldicl	r0, r0, 63,33',	C ...branch count
52`	srdi	r0, r0, 1')	C ...for ctr
53	mtctr	r0
54
55	andi.	r0, n, 1
56
57	li	r0, 0
58	li	r12, 0
59
60	beq	L(evn)
61
62L(odd):	ld	r6, 0(up)
63	addi	up, up, 8
64	ld	r8, 0(vp)
65	addi	vp, vp, 8
66	xor	r10, r6, r8
67	popcntd(r0, r10)
68	bdz	L(e1)
69
70L(evn):	ld	r6, 0(up)
71	ld	r8, 0(vp)
72	ld	r7, 8(up)
73	ld	r9, 8(vp)
74	xor	r10, r6, r8
75	addi	up, up, 16
76	addi	vp, vp, 16
77	li	r30, 0
78	li	r31, 0
79	bdz	L(end)
80
81	nop
82	nop
83C	ALIGN(16)
84L(top):	add	r0, r0, r30
85	ld	r6, 0(up)
86	ld	r8, 0(vp)
87	xor	r11, r7, r9
88	popcntd(r30, r10)
89	add	r12, r12, r31
90	ld	r7, 8(up)
91	ld	r9, 8(vp)
92	xor	r10, r6, r8
93	popcntd(r31, r11)
94	addi	up, up, 16
95	addi	vp, vp, 16
96	bdnz	L(top)
97
98L(end):	add	r0, r0, r30
99	xor	r11, r7, r9
100	popcntd(r30, r10)
101	add	r12, r12, r31
102	popcntd(r31, r11)
103
104	add	r0, r0, r30
105	add	r12, r12, r31
106L(e1):	add	r3, r0, r12
107	ld	r30, -16(r1)
108	ld	r31, -8(r1)
109	blr
110EPILOGUE()
111