aboutsummaryrefslogtreecommitdiff
path: root/vendor/gmp-6.3.0/mpn/powerpc64/lshiftc.asm
blob: 7cf6a83428c59b2e4dc2f673845909be6c2f1886 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
dnl  PowerPC-64 mpn_lshiftc -- rp[] = ~up[] << cnt

dnl  Copyright 2003, 2005, 2010 Free Software Foundation, Inc.

dnl  This file is part of the GNU MP Library.
dnl
dnl  The GNU MP Library is free software; you can redistribute it and/or modify
dnl  it under the terms of either:
dnl
dnl    * the GNU Lesser General Public License as published by the Free
dnl      Software Foundation; either version 3 of the License, or (at your
dnl      option) any later version.
dnl
dnl  or
dnl
dnl    * the GNU General Public License as published by the Free Software
dnl      Foundation; either version 2 of the License, or (at your option) any
dnl      later version.
dnl
dnl  or both in parallel, as here.
dnl
dnl  The GNU MP Library is distributed in the hope that it will be useful, but
dnl  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
dnl  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
dnl  for more details.
dnl
dnl  You should have received copies of the GNU General Public License and the
dnl  GNU Lesser General Public License along with the GNU MP Library.  If not,
dnl  see https://www.gnu.org/licenses/.

include(`../config.m4')

C                   cycles/limb
C POWER3/PPC630          ?
C POWER4/PPC970          ?
C POWER5                 2.25
C POWER6                 9.5
C POWER7                 2.15

C TODO
C  * Try to reduce the number of needed live registers
C  * Micro-optimise header code
C  * Keep in synch with lshift.asm and rshift.asm
C  * Could the long-scheduled std insns be less scheduled?

C INPUT PARAMETERS
define(`rp',  `r3')
define(`up',  `r4')
define(`n',   `r5')
define(`cnt', `r6')

define(`tnc',`r0')
define(`u0',`r30')
define(`u1',`r31')
define(`retval',`r5')

ASM_START()
PROLOGUE(mpn_lshiftc)
	std	r31, -8(r1)
	std	r30, -16(r1)
	subfic	tnc, cnt, 64
	sldi	r7, n, 3	C byte count corresponding to n
	add	up, up, r7	C up = up + n
	add	rp, rp, r7	C rp = rp + n
	rldicl.	r30, n, 0,62	C r30 = n & 3, set cr0
	cmpdi	cr6, r30, 2
	addi	r31, n, 3	C compute count...
	ld	r10, -8(up)	C load 1st limb for b00...b11
	srd	retval, r10, tnc
	srdi	r31, r31, 2	C ...for ctr
	mtctr	r31		C copy count into ctr
	beq	cr0, L(b00)
	blt	cr6, L(b01)
	ld	r11, -16(up)	C load 2nd limb for b10 and b11
	beq	cr6, L(b10)

	ALIGN(16)
L(b11):	sld	r8, r10, cnt
	srd	r9, r11, tnc
	ld	u1, -24(up)
	addi	up, up, -24
	sld	r12, r11, cnt
	srd	r7, u1, tnc
	addi	rp, rp, 16
	bdnz	L(gt3)

	nor	r11, r8, r9
	sld	r8, u1, cnt
	nor	r8, r8, r8
	b	L(cj3)

	ALIGN(16)
L(gt3):	ld	u0, -8(up)
	nor	r11, r8, r9
	sld	r8, u1, cnt
	srd	r9, u0, tnc
	ld	u1, -16(up)
	nor	r10, r12, r7
	b	L(L11)

	ALIGN(32)
L(b10):	sld	r12, r10, cnt
	addi	rp, rp, 24
	srd	r7, r11, tnc
	bdnz	L(gt2)

	sld	r8, r11, cnt
	nor	r10, r12, r7
	nor	r8, r8, r8
	b	L(cj2)

L(gt2):	ld	u0, -24(up)
	sld	r8, r11, cnt
	srd	r9, u0, tnc
	ld	u1, -32(up)
	nor	r10, r12, r7
	sld	r12, u0, cnt
	srd	r7, u1, tnc
	ld	u0, -40(up)
	nor	r11, r8, r9
	addi	up, up, -16
	b	L(L10)

	ALIGN(16)
L(b00):	ld	u1, -16(up)
	sld	r12, r10, cnt
	srd	r7, u1, tnc
	ld	u0, -24(up)
	sld	r8, u1, cnt
	srd	r9, u0, tnc
	ld	u1, -32(up)
	nor	r10, r12, r7
	sld	r12, u0, cnt
	srd	r7, u1, tnc
	addi	rp, rp, 8
	bdz	L(cj4)

L(gt4):	addi	up, up, -32
	ld	u0, -8(up)
	nor	r11, r8, r9
	b	L(L00)

	ALIGN(16)
L(b01):	bdnz	L(gt1)
	sld	r8, r10, cnt
	nor	r8, r8, r8
	std	r8, -8(rp)
	b	L(ret)

L(gt1):	ld	u0, -16(up)
	sld	r8, r10, cnt
	srd	r9, u0, tnc
	ld	u1, -24(up)
	sld	r12, u0, cnt
	srd	r7, u1, tnc
	ld	u0, -32(up)
	nor	r11, r8, r9
	sld	r8, u1, cnt
	srd	r9, u0, tnc
	ld	u1, -40(up)
	addi	up, up, -40
	nor	r10, r12, r7
	bdz	L(end)

	ALIGN(32)
L(top):	sld	r12, u0, cnt
	srd	r7, u1, tnc
	ld	u0, -8(up)
	std	r11, -8(rp)
	nor	r11, r8, r9
L(L00):	sld	r8, u1, cnt
	srd	r9, u0, tnc
	ld	u1, -16(up)
	std	r10, -16(rp)
	nor	r10, r12, r7
L(L11):	sld	r12, u0, cnt
	srd	r7, u1, tnc
	ld	u0, -24(up)
	std	r11, -24(rp)
	nor	r11, r8, r9
L(L10):	sld	r8, u1, cnt
	srd	r9, u0, tnc
	ld	u1, -32(up)
	addi	up, up, -32
	std	r10, -32(rp)
	addi	rp, rp, -32
	nor	r10, r12, r7
	bdnz	L(top)

	ALIGN(32)
L(end):	sld	r12, u0, cnt
	srd	r7, u1, tnc
	std	r11, -8(rp)
L(cj4):	nor	r11, r8, r9
	sld	r8, u1, cnt
	std	r10, -16(rp)
	nor	r8, r8, r8
L(cj3):	nor	r10, r12, r7
	std	r11, -24(rp)
L(cj2):	std	r10, -32(rp)
	std	r8, -40(rp)

L(ret):	ld	r31, -8(r1)
	ld	r30, -16(r1)
ifdef(`HAVE_ABI_mode32',
`	srdi	r3, retval, 32
	mr	r4, retval
',`	mr	r3, retval')
	blr
EPILOGUE()