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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
|
/*************************************************************************
* *
* 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. *
* *
*************************************************************************/
/*
* L1Transposed Equation Solving Routines
* Copyright (c) 2017-2019 Oleh Derevenko, odar@eleks.com (change all "a" to "e")
*/
#include <ode/common.h>
#include <ode/matrix.h>
#include <ode/matrix_coop.h>
#include "config.h"
#include "threaded_solver_ldlt.h"
#include "threading_base.h"
#include "resource_control.h"
#include "error.h"
#include "fastltsolve_impl.h"
/*static */
void ThreadedEquationSolverLDLT::estimateCooperativeSolvingL1TransposedResourceRequirements(
dxResourceRequirementDescriptor *summaryRequirementsDescriptor,
unsigned allowedThreadCount, unsigned rowCount)
{
dxThreadingBase *threading = summaryRequirementsDescriptor->getrelatedThreading();
unsigned limitedThreadCount = restrictSolvingL1TransposedAllowedThreadCount(threading, allowedThreadCount, rowCount);
if (limitedThreadCount > 1)
{
doEstimateCooperativeSolvingL1TransposedResourceRequirementsValidated(summaryRequirementsDescriptor, allowedThreadCount, rowCount);
}
}
/*static */
void ThreadedEquationSolverLDLT::cooperativelySolveL1Transposed(
dxRequiredResourceContainer *resourceContainer, unsigned allowedThreadCount,
const dReal *L, dReal *b, unsigned rowCount, unsigned rowSkip)
{
dIASSERT(rowCount != 0);
dxThreadingBase *threading = resourceContainer->getThreadingInstance();
unsigned limitedThreadCount = restrictSolvingL1TransposedAllowedThreadCount(threading, allowedThreadCount, rowCount);
if (limitedThreadCount <= 1)
{
solveL1Transposed<SL1T_B_STRIDE>(L, b, rowCount, rowSkip);
}
else
{
doCooperativelySolveL1TransposedValidated(resourceContainer, limitedThreadCount, L, b, rowCount, rowSkip);
}
}
/*static */
unsigned ThreadedEquationSolverLDLT::restrictSolvingL1TransposedAllowedThreadCount(
dxThreadingBase *threading, unsigned allowedThreadCount, unsigned rowCount)
{
unsigned limitedThreadCount = 1;
#if dCOOPERATIVE_ENABLED
const unsigned int blockStep = SL1T_BLOCK_SIZE; // Required by the implementation
unsigned solvingBlockCount = deriveSolvingL1TransposedBlockCount(rowCount, blockStep);
dIASSERT(deriveSolvingL1TransposedThreadCount(SL1T_COOPERATIVE_BLOCK_COUNT_MINIMUM, 2) > 1);
if (solvingBlockCount >= SL1T_COOPERATIVE_BLOCK_COUNT_MINIMUM)
{
limitedThreadCount = threading->calculateThreadingLimitedThreadCount(allowedThreadCount, true);
}
#endif // #if dCOOPERATIVE_ENABLED
return limitedThreadCount;
}
/*static */
void ThreadedEquationSolverLDLT::doEstimateCooperativeSolvingL1TransposedResourceRequirementsValidated(
dxResourceRequirementDescriptor *summaryRequirementsDescriptor,
unsigned allowedThreadCount, unsigned rowCount)
{
const unsigned int blockStep = SL1T_BLOCK_SIZE; // Required by the implementation
unsigned blockCount = deriveSolvingL1TransposedBlockCount(rowCount, blockStep);
dIASSERT(blockCount >= 1);
unsigned threadCountToUse = deriveSolvingL1TransposedThreadCount(blockCount, allowedThreadCount);
dIASSERT(threadCountToUse > 1);
unsigned simultaneousCallCount = 1 + (threadCountToUse - 1);
SolvingL1TransposedMemoryEstimates solvingMemoryEstimates;
sizeint solvingMemoryRequired = estimateCooperativelySolvingL1TransposedMemoryRequirement<blockStep>(rowCount, solvingMemoryEstimates);
const unsigned solvingAlignmentRequired = ALLOCATION_DEFAULT_ALIGNMENT;
unsigned featureRequirement = dxResourceRequirementDescriptor::STOCK_CALLWAIT_REQUIRED;
summaryRequirementsDescriptor->mergeAnotherDescriptorIn(solvingMemoryRequired, solvingAlignmentRequired, simultaneousCallCount, featureRequirement);
}
/*static */
void ThreadedEquationSolverLDLT::doCooperativelySolveL1TransposedValidated(
dxRequiredResourceContainer *resourceContainer, unsigned allowedThreadCount,
const dReal *L, dReal *b, unsigned rowCount, unsigned rowSkip)
{
dIASSERT(allowedThreadCount > 1);
const unsigned int blockStep = SL1T_BLOCK_SIZE; // Required by the implementation
unsigned blockCount = deriveSolvingL1TransposedBlockCount(rowCount, blockStep);
dIASSERT(blockCount >= 1);
unsigned threadCountToUse = deriveSolvingL1TransposedThreadCount(blockCount, allowedThreadCount);
dIASSERT(threadCountToUse > 1);
dCallWaitID completionWait = resourceContainer->getStockCallWait();
dAASSERT(completionWait != NULL);
atomicord32 blockCompletionProgress;
cellindexint *blockProgressDescriptors;
SolveL1TransposedCellContext *cellContexts;
SolvingL1TransposedMemoryEstimates solvingMemoryEstimates;
sizeint solvingMemoryRequired = estimateCooperativelySolvingL1TransposedMemoryRequirement<blockStep>(rowCount, solvingMemoryEstimates);
dIASSERT(solvingMemoryRequired <= resourceContainer->getMemoryBufferSize());
void *bufferAllocated = resourceContainer->getMemoryBufferPointer();
dIASSERT(bufferAllocated != NULL);
dIASSERT(dALIGN_PTR(bufferAllocated, ALLOCATION_DEFAULT_ALIGNMENT) == bufferAllocated);
void *bufferCurrentLocation = bufferAllocated;
bufferCurrentLocation = markCooperativelySolvingL1TransposedMemoryStructuresOut(bufferCurrentLocation, solvingMemoryEstimates, blockProgressDescriptors, cellContexts);
dIVERIFY(bufferCurrentLocation <= (uint8 *)bufferAllocated + solvingMemoryRequired);
initializeCooperativelySolveL1TransposedMemoryStructures<blockStep>(rowCount, blockCompletionProgress, blockProgressDescriptors, cellContexts);
dCallReleaseeID calculationFinishReleasee;
SolveL1TransposedWorkerContext workerContext; // The variable must exist in the outer scope
workerContext.init(L, b, rowCount, rowSkip, blockCompletionProgress, blockProgressDescriptors, cellContexts);
dxThreadingBase *threading = resourceContainer->getThreadingInstance();
threading->PostThreadedCall(NULL, &calculationFinishReleasee, threadCountToUse - 1, NULL, completionWait, &solveL1Transposed_completion_callback, NULL, 0, "SolveL1Transposed Completion");
threading->PostThreadedCallsGroup(NULL, threadCountToUse - 1, calculationFinishReleasee, &solveL1Transposed_worker_callback, &workerContext, "SolveL1Transposed Work");
participateSolvingL1Transposed<blockStep, SL1T_B_STRIDE>(L, b, rowCount, rowSkip, blockCompletionProgress, blockProgressDescriptors, cellContexts, threadCountToUse - 1);
threading->WaitThreadedCallExclusively(NULL, completionWait, NULL, "SolveL1Transposed End Wait");
}
/*static */
int ThreadedEquationSolverLDLT::solveL1Transposed_worker_callback(void *callContext, dcallindex_t callInstanceIndex, dCallReleaseeID dUNUSED(callThisReleasee))
{
SolveL1TransposedWorkerContext *ptrContext = (SolveL1TransposedWorkerContext *)callContext;
solveL1Transposed_worker(*ptrContext, dCAST_TO_SMALLER(unsigned, callInstanceIndex));
return 1;
}
/*static */
void ThreadedEquationSolverLDLT::solveL1Transposed_worker(SolveL1TransposedWorkerContext &ref_context, unsigned ownThreadIndex)
{
const unsigned blockStep = SL1T_BLOCK_SIZE;
participateSolvingL1Transposed<blockStep, SL1T_B_STRIDE>(ref_context.m_L, ref_context.m_b, ref_context.m_rowCount, ref_context.m_rowSkip,
*ref_context.m_ptrBlockCompletionProgress, ref_context.m_blockProgressDescriptors, ref_context.m_cellContexts, ownThreadIndex);
}
/*static */
int ThreadedEquationSolverLDLT::solveL1Transposed_completion_callback(void *dUNUSED(callContext), dcallindex_t dUNUSED(callInstanceIndex), dCallReleaseeID dUNUSED(callThisReleasee))
{
return 1;
}
//////////////////////////////////////////////////////////////////////////
// Public interface functions
/*extern ODE_API */
void dSolveL1T(const dReal *L, dReal *B, int rowCount, int rowSkip)
{
dAASSERT(rowCount != 0);
if (rowCount != 0)
{
dAASSERT(L != NULL);
dAASSERT(B != NULL);
solveL1Transposed<1>(L, B, rowCount, rowSkip);
}
}
/*extern ODE_API */
void dEstimateCooperativelySolveL1TransposedResourceRequirements(dResourceRequirementsID requirements,
unsigned maximalAllowedThreadCount, unsigned maximalRowCount)
{
dAASSERT(requirements != NULL);
dxResourceRequirementDescriptor *requirementsDescriptor = (dxResourceRequirementDescriptor *)requirements;
ThreadedEquationSolverLDLT::estimateCooperativeSolvingL1TransposedResourceRequirements(requirementsDescriptor, maximalAllowedThreadCount, maximalRowCount);
}
/*extern ODE_API */
void dCooperativelySolveL1Transposed(dResourceContainerID resources, unsigned allowedThreadCount,
const dReal *L, dReal *b, unsigned rowCount, unsigned rowSkip)
{
dAASSERT(resources != NULL);
dxRequiredResourceContainer *resourceContainer = (dxRequiredResourceContainer *)resources;
ThreadedEquationSolverLDLT::cooperativelySolveL1Transposed(resourceContainer, allowedThreadCount, L, b, rowCount, rowSkip);
}
|
