-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathtime_CPU_pinnochio.cpp
195 lines (178 loc) · 8.61 KB
/
time_CPU_pinnochio.cpp
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
/***
clang++-10 -std=c++11 -o pinnochio_timing.exe time_CPU_pinnochio.cpp -O3 -DPINOCCHIO_URDFDOM_TYPEDEF_SHARED_PTR -DPINOCCHIO_WITH_URDFDOM -lboost_system -L/opt/openrobots/lib -lpinocchio -lurdfdom_model -lpthread
***/
#include "utils/experiment_helpers.h" // include constants and other experiment consistency helpers
#define BOOST_BIND_GLOBAL_PLACEHOLDERS // fixes a warning I get
#include "pinocchio/spatial/fwd.hpp"
#include "pinocchio/spatial/se3.hpp"
#include "pinocchio/multibody/visitor.hpp"
#include "pinocchio/multibody/model.hpp"
#include "pinocchio/multibody/data.hpp"
#include "pinocchio/algorithm/joint-configuration.hpp"
#include "pinocchio/algorithm/kinematics.hpp"
#include "pinocchio/algorithm/kinematics-derivatives.hpp"
#include "pinocchio/algorithm/rnea-derivatives.hpp"
#include "pinocchio/algorithm/aba-derivatives.hpp"
#include "pinocchio/algorithm/crba.hpp"
#include "pinocchio/algorithm/centroidal.hpp"
#include "pinocchio/algorithm/aba.hpp"
#include "pinocchio/algorithm/rnea.hpp"
#include "pinocchio/algorithm/cholesky.hpp"
#include "pinocchio/algorithm/jacobian.hpp"
#include "pinocchio/algorithm/center-of-mass.hpp"
#include "pinocchio/algorithm/compute-all-terms.hpp"
#include "pinocchio/algorithm/kinematics.hpp"
#include "pinocchio/parsers/urdf.hpp"
#include "pinocchio/parsers/sample-models.hpp"
#include "pinocchio/container/aligned-vector.hpp"
#include <stdlib.h>
#include <iostream>
#include <fstream>
#include <sstream>
#include <sys/time.h>
#define EIGEN_DEFAULT_TO_ROW_MAJOR
#include <Eigen/Dense>
#include <Eigen/StdVector>
EIGEN_DEFINE_STL_VECTOR_SPECIALIZATION(Eigen::VectorXd)
using namespace Eigen;
using namespace pinocchio;
void dynamicsGradientThreaded_inner(Model *model, Data *datas, MatrixXd *dqdd_dqs, MatrixXd *dqdd_dqds, \
VectorXd *qs, VectorXd *qds, VectorXd *qdds, MatrixXd *Minvs, int tid, int kStart, int kMax){
MatrixXd drnea_dq = MatrixXd::Zero(model->nq,model->nq);
MatrixXd drnea_dv = MatrixXd::Zero(model->nv,model->nv);
MatrixXd drnea_da = MatrixXd::Zero(model->nv,model->nv);
for(int k = kStart; k < kMax; k++){
computeRNEADerivatives(*model,datas[tid],qs[k],qds[k],qdds[k],drnea_dq,drnea_dv,drnea_da);
dqdd_dqs[k] = -Minvs[k]*drnea_dq;
dqdd_dqds[k] = -Minvs[k]*drnea_dv;
}
}
template<int NUM_THREADS, int NUM_TIME_STEPS>
void dynamicsGradientReusableThreaded(Model *model, Data *datas, MatrixXd *dqdd_dqs, MatrixXd *dqdd_dqds, \
VectorXd *qs, VectorXd *qds, VectorXd *qdds, MatrixXd *Minvs, ReusableThreads<NUM_THREADS> *threads){
for (int tid = 0; tid < NUM_THREADS; tid++){
int kStart = NUM_TIME_STEPS/NUM_THREADS*tid; int kMax = NUM_TIME_STEPS/NUM_THREADS*(tid+1);
if(tid == NUM_THREADS-1){kMax = NUM_TIME_STEPS;}
threads->addTask(tid, dynamicsGradientThreaded_inner, std::ref(model), std::ref(datas), std::ref(dqdd_dqs), std::ref(dqdd_dqds),
std::ref(qs), std::ref(qds), std::ref(qdds), std::ref(Minvs), tid, kStart, kMax);
}
threads->sync();
}
template<int TEST_ITERS, int NUM_THREADS_TEST, int NUM_TIME_STEPS_TEST>
void test(){
// Setup timer
struct timespec start, end;
// Import URDF model and prepare pinnochio
Model model;
std::string urdf_filename = "utils/iiwa_14.urdf";
pinocchio::urdf::buildModel(urdf_filename,model);
// model.gravity.setZero();
model.gravity.linear(Eigen::Vector3d(0,0,-9.81));
Data datas[NUM_THREADS_TEST];
for(int i = 0; i < NUM_THREADS_TEST; i++){datas[i] = Data(model);}
// allocate and load on CPU
VectorXd qs[NUM_TIME_STEPS_TEST];
VectorXd qds[NUM_TIME_STEPS_TEST];
VectorXd qdds[NUM_TIME_STEPS_TEST];
VectorXd us[NUM_TIME_STEPS_TEST];
MatrixXd Minvs[NUM_TIME_STEPS_TEST];
MatrixXd dqdd_dqs[NUM_TIME_STEPS_TEST];
MatrixXd dqdd_dqds[NUM_TIME_STEPS_TEST];
#pragma unroll
for(int i = 0; i < NUM_TIME_STEPS_TEST; i++){
qs[i] = VectorXd::Zero(model.nq);
qds[i] = VectorXd::Zero(model.nv);
qdds[i] = VectorXd::Zero(model.nv);
us[i] = VectorXd::Zero(model.nv);
Minvs[i] = MatrixXd::Zero(model.nq,model.nq);
dqdd_dqs[i] = MatrixXd::Zero(model.nv,model.nq);
dqdd_dqds[i] = MatrixXd::Zero(model.nv,model.nv);
#if TEST_FOR_EQUIVALENCE
for(int j = 0; j < model.nq; j++){qs[i][j] = 0.1; qds[i][j] = 0.1; us[i][j] = 0.1;}
#else
for(int j = 0; j < model.nq; j++){qs[i][j] = getRand<double>(); qds[i][j] = getRand<double>(); us[i][j] = getRand<double>();}
#endif
aba(model,datas[0],qs[i],qds[i],us[i]);
computeMinverse(model,datas[0],qs[i]);
datas[0].Minv.template triangularView<Eigen::StrictlyLower>() = datas[0].Minv.transpose().template triangularView<Eigen::StrictlyLower>();
qdds[i] = datas[0].ddq;
Minvs[i] = datas[0].Minv;
}
// Single call
if(NUM_TIME_STEPS_TEST == 1){
#if TEST_FOR_EQUIVALENCE
int SINGLE_TEST_ITERS = TEST_ITERS*1;
#else
int SINGLE_TEST_ITERS = TEST_ITERS*10;
#endif
MatrixXd drnea_dq = MatrixXd::Zero(model.nq,model.nq);
MatrixXd drnea_dv = MatrixXd::Zero(model.nv,model.nv);
MatrixXd drnea_da = MatrixXd::Zero(model.nv,model.nv);
clock_gettime(CLOCK_MONOTONIC,&start);
for(int i = 0; i < SINGLE_TEST_ITERS; i++){int ind = 0;
computeRNEADerivatives(model,datas[0],qs[ind],qds[ind],qdds[ind],drnea_dq,drnea_dv,drnea_da);
dqdd_dqs[ind] = -Minvs[ind]*drnea_dq;
dqdd_dqds[ind] = -Minvs[ind]*drnea_dv;
#if TEST_FOR_EQUIVALENCE
std::cout << "q,qd,qdd,u" << std::endl;
std::cout << qs[0].transpose() << std::endl;
std::cout << qds[0].transpose() << std::endl;
std::cout << qdds[0].transpose() << std::endl;
std::cout << us[0].transpose() << std::endl;
std::cout << "Minv" << std::endl << Minvs[0] << std::endl;
std::cout << "dRNEA_dq" << std::endl << drnea_dq << std::endl;
std::cout << "dRNEA_dqd" << std::endl << drnea_dv << std::endl;
std::cout << "dqdd_dq" << std::endl << dqdd_dqs[0] << std::endl;
std::cout << "dqdd_dqd" << std::endl << dqdd_dqds[0] << std::endl;
#endif
}
clock_gettime(CLOCK_MONOTONIC,&end);
printf("Single Call vaf+dc/du+dqdd/du %fus\n",time_delta_us_timespec(start,end)/static_cast<double>(SINGLE_TEST_ITERS));
clock_gettime(CLOCK_MONOTONIC,&start);
for(int i = 0; i < SINGLE_TEST_ITERS; i++){int ind = 0;
computeRNEADerivatives(model,datas[0],qs[ind],qds[ind],qdds[ind],drnea_dq,drnea_dv,drnea_da);
}
clock_gettime(CLOCK_MONOTONIC,&end);
printf("Single Call vaf+dc/du %fus\n",time_delta_us_timespec(start,end)/static_cast<double>(SINGLE_TEST_ITERS));
clock_gettime(CLOCK_MONOTONIC,&start);
for(int i = 0; i < SINGLE_TEST_ITERS; i++){int ind = 0;
rnea(model,datas[0],qs[ind],qds[ind],qdds[ind]);
}
clock_gettime(CLOCK_MONOTONIC,&end);
printf("Single Call vaf %fus\n",time_delta_us_timespec(start,end)/static_cast<double>(SINGLE_TEST_ITERS));
clock_gettime(CLOCK_MONOTONIC,&start);
for(int i = 0; i < SINGLE_TEST_ITERS; i++){int ind = 0;
dqdd_dqs[ind] = -Minvs[ind]*drnea_dq;
dqdd_dqds[ind] = -Minvs[ind]*drnea_dv;
}
clock_gettime(CLOCK_MONOTONIC,&end);
printf("Single Call dqdd/du %fus\n",time_delta_us_timespec(start,end)/static_cast<double>(SINGLE_TEST_ITERS));
printf("----------------------------------------\n");
}
// multi call with threadPools
else{
ReusableThreads<NUM_THREADS_TEST> threads;
std::vector<double> times = {};
for(int iter = 0; iter < TEST_ITERS; iter++){
clock_gettime(CLOCK_MONOTONIC,&start);
dynamicsGradientReusableThreaded<NUM_THREADS_TEST,NUM_TIME_STEPS_TEST>(&model,datas,dqdd_dqs,dqdd_dqds,qs,qds,qdds,Minvs,&threads);
clock_gettime(CLOCK_MONOTONIC,&end);
times.push_back(time_delta_us_timespec(start,end));
}
printf("[N:%d]: CPU ReusableThreaded: ",NUM_TIME_STEPS_TEST); printStats<PRINT_DISTRIBUTIONS_GLOBAL>(×);
printf("----------------------------------------\n");
}
}
int main(int argc, const char ** argv)
{
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,1>();
#if !TEST_FOR_EQUIVALENCE
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,10>();
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,16>();
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,32>();
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,64>();
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,128>();
test<TEST_ITERS_GLOBAL,CPU_THREADS_GLOBAL,256>();
#endif
return 0;
}