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
|
/*************************************************************************
* *
* Open Dynamics Engine, Copyright (C) 2001,2002 Russell L. Smith. *
* All rights reserved. Email: russ@q12.org Web: www.q12.org *
* *
* This library is free software; you can redistribute it and/or *
* modify it under the terms of EITHER: *
* (1) The GNU Lesser General Public License as published by the Free *
* Software Foundation; either version 2.1 of the License, or (at *
* your option) any later version. The text of the GNU Lesser *
* General Public License is included with this library in the *
* file LICENSE.TXT. *
* (2) The BSD-style license that is included with this library in *
* the file LICENSE-BSD.TXT. *
* *
* This library is distributed in the hope that it will be useful, *
* but WITHOUT ANY WARRANTY; without even the implied warranty of *
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the files *
* LICENSE.TXT and LICENSE-BSD.TXT for more details. *
* *
*************************************************************************/
/*
* Vector scaling function implementation
* Improvements and cooperative implementation copyright (c) 2017-2019 Oleh Derevenko, odar@eleks.com (change all "a" to "e")
*/
#ifndef _ODE_FASTVECSCALE_IMPL_H_
#define _ODE_FASTVECSCALE_IMPL_H_
template<unsigned int a_stride, unsigned int d_stride>
void scaleLargeVector(dReal *aStart, const dReal *dStart, unsigned elementCount)
{
dAASSERT (aStart && dStart && elementCount >= 0);
const unsigned step = 4;
dReal *ptrA = aStart;
const dReal *ptrD = dStart;
const dReal *const dStepsEnd = dStart + (sizeint)(elementCount & ~(step - 1)) * d_stride;
for (; ptrD != dStepsEnd; ptrA += step * a_stride, ptrD += step * d_stride)
{
dReal a0 = ptrA[0], a1 = ptrA[1 * a_stride], a2 = ptrA[2 * a_stride], a3 = ptrA[3 * a_stride];
dReal d0 = ptrD[0], d1 = ptrD[1 * d_stride], d2 = ptrD[2 * d_stride], d3 = ptrD[3 * d_stride];
a0 *= d0;
a1 *= d1;
a2 *= d2;
a3 *= d3;
ptrA[0] = a0; ptrA[1 * a_stride] = a1; ptrA[2 * a_stride] = a2; ptrA[3 * a_stride] = a3;
dSASSERT(step == 4);
}
switch (elementCount & (step - 1))
{
case 3:
{
dReal a2 = ptrA[2 * a_stride];
dReal d2 = ptrD[2 * d_stride];
ptrA[2 * a_stride] = a2 * d2;
// break; -- proceed to case 2
}
case 2:
{
dReal a1 = ptrA[1 * a_stride];
dReal d1 = ptrD[1 * d_stride];
ptrA[1 * a_stride] = a1 * d1;
// break; -- proceed to case 1
}
case 1:
{
dReal a0 = ptrA[0];
dReal d0 = ptrD[0];
ptrA[0] = a0 * d0;
break;
}
}
dSASSERT(step == 4);
}
template<unsigned int block_step, unsigned int a_stride, unsigned int d_stride>
/*static */
void ThreadedEquationSolverLDLT::participateScalingVector(dReal *ptrAStart, const dReal *ptrDStart, const unsigned elementCount,
volatile atomicord32 &refBlockCompletionProgress/*=0*/)
{
dAASSERT (ptrAStart != NULL);
dAASSERT(ptrDStart != NULL);
dAASSERT(elementCount >= 0);
const unsigned wrapSize = 4;
dSASSERT(block_step % wrapSize == 0);
const unsigned completeBlockCount = elementCount / block_step;
const unsigned trailingBlockElements = elementCount % block_step;
unsigned blockIndex;
while ((blockIndex = ThrsafeIncrementIntUpToLimit(&refBlockCompletionProgress, completeBlockCount)) != completeBlockCount)
{
dReal *ptrAElement = ptrAStart + (sizeint)(blockIndex * block_step) * a_stride;
const dReal *ptrDElement = ptrDStart + (sizeint)(blockIndex * block_step) * d_stride;
const dReal *const ptrDBlockEnd = ptrDElement + block_step * d_stride;
dSASSERT((sizeint)block_step * a_stride < UINT_MAX);
dSASSERT((sizeint)block_step * d_stride < UINT_MAX);
for (; ptrDElement != ptrDBlockEnd; ptrAElement += wrapSize * a_stride, ptrDElement += wrapSize * d_stride)
{
dReal a0 = ptrAElement[0], a1 = ptrAElement[1 * a_stride], a2 = ptrAElement[2 * a_stride], a3 = ptrAElement[3 * a_stride];
dReal d0 = ptrDElement[0], d1 = ptrDElement[1 * d_stride], d2 = ptrDElement[2 * d_stride], d3 = ptrDElement[3 * d_stride];
a0 *= d0;
a1 *= d1;
a2 *= d2;
a3 *= d3;
ptrAElement[0] = a0; ptrAElement[1 * a_stride] = a1; ptrAElement[2 * a_stride] = a2; ptrAElement[3 * a_stride] = a3;
dSASSERT(wrapSize == 4);
}
}
if (trailingBlockElements != 0 && (blockIndex = ThrsafeIncrementIntUpToLimit(&refBlockCompletionProgress, completeBlockCount + 1)) != completeBlockCount + 1)
{
dReal *ptrAElement = ptrAStart + (sizeint)(completeBlockCount * block_step) * a_stride;
const dReal *ptrDElement = ptrDStart + (sizeint)(completeBlockCount * block_step) * d_stride;
const dReal *const ptrDBlockEnd = ptrDElement + (trailingBlockElements & ~(wrapSize - 1)) * d_stride;
for (; ptrDElement != ptrDBlockEnd; ptrAElement += wrapSize * a_stride, ptrDElement += wrapSize * d_stride)
{
dReal a0 = ptrAElement[0], a1 = ptrAElement[1 * a_stride], a2 = ptrAElement[2 * a_stride], a3 = ptrAElement[3 * a_stride];
dReal d0 = ptrDElement[0], d1 = ptrDElement[1 * d_stride], d2 = ptrDElement[2 * d_stride], d3 = ptrDElement[3 * d_stride];
a0 *= d0;
a1 *= d1;
a2 *= d2;
a3 *= d3;
ptrAElement[0] = a0; ptrAElement[1 * a_stride] = a1; ptrAElement[2 * a_stride] = a2; ptrAElement[3 * a_stride] = a3;
dSASSERT(wrapSize == 4);
}
switch (trailingBlockElements & (wrapSize - 1))
{
case 3:
{
dReal a2 = ptrAElement[2 * a_stride];
dReal d2 = ptrDElement[2 * d_stride];
ptrAElement[2 * a_stride] = a2 * d2;
// break; -- proceed to case 2
}
case 2:
{
dReal a1 = ptrAElement[1 * a_stride];
dReal d1 = ptrDElement[1 * d_stride];
ptrAElement[1 * a_stride] = a1 * d1;
// break; -- proceed to case 1
}
case 1:
{
dReal a0 = ptrAElement[0];
dReal d0 = ptrDElement[0];
ptrAElement[0] = a0 * d0;
break;
}
}
dSASSERT(wrapSize == 4);
}
}
#endif
|
