Multiple for loop in C++ - c++11

I am trying to write current time and some data (distance & amplitude) from a sensor in a text file. Number of data are very large (sensor rotate at 50Hz frequency and number of data could be 5000 per scan). Now I want write current time first and then all the data in a single line like this,
11:23:17 (time)
distance1 amplitude1; distance2 amplitude2; ... distance5000 amplitude5000;
11:23:18
distance1 amplitude1; distance2 amplitude2; ... distance5000 amplitude5000;
.
.
.
11:27:00
distance1 amplitude1; distance2 amplitude2; ... distance5000 amplitude5000;
"So my question is how to get this?"
I am able to write only distance and data like below
for(int t=0; t<distances.size(); t++)
{
pfsave << distances[t] <<"\t" << amplitudes[t]<<";";
}
pfsave<<endl;
Note: data type of distances & amplitudes are
vector<uint32_t> distacnes;
vector<uint32_t> amplitudes;

You can write the current time like this, before the beginning of your for loop:
auto t = std::time(nullptr);
auto tm = *std::localtime(&t);
pfsave << std::put_time(&tm, "%H:%M:%S") << std::endl;
Then your for loop without endl, to write a single line:
for(int t=0; t<distances.size(); t++)
{
pfsave << distances[t] <<"\t" << amplitudes[t]<<";";
}
And finally add endl, to finish the line:
pfsave << endl;
EDIT: following your comment
bool canContinue = true; // Condition used to stop the loop when needded
while(canContinue)
{
// Read data from your scanning device
distances = ... ;
amplitudes = ... ;
// Write output file
auto t = std::time(nullptr);
auto tm = *std::localtime(&t);
pfsave << std::put_time(&tm, "%H:%M:%S") << std::endl;
for(int t=0; t<distances.size(); t++)
pfsave << distances[t] <<"\t" << amplitudes[t]<<";";
pfsave << endl;
// Update of canContinue
canContinue = ... ;
}

Related

how to get the third element (from the beginning and from the end) in a map c++

I want to access the third element(both in forward and reverse order) without iterating the map.
int main() {
map<int,int>mp;
vector<int>v={1,4,2,5,4,2,5,7,6,9,7,9,8};
for(int i=0;i<v.size();++i)
{
mp[v[i]]++;
}
return 0;
}
here the expected output would be 4 2 and 7 2.
You can use the std::advance function, but note that the map container is not designed for random access so even if you avoid iterating yourself for performance reasons, it is what happens under the hood anyway giving linear time complexity.
// forward iterator
auto it_fwd = mp.begin();
std::advance(it_fwd, 2); // go to the 3rd item from the beggining
// backward iterator
auto it_bwd = mp.rbegin();
std::advance(it_bwd, 2); // go to the 3rd item from the end
std::cout << it_fwd->first << ':' << it_fwd->second << '\n';
std::cout << it_bwd->first << ':' << it_bwd->second << '\n';

How can i make this modification on Dijkstra Algorithm more efficient?

The problem is a part of my computer science homework. The homework includes 5 different types of students that travel through a given weighted undirected node graph where each student has a different method. The fifth student is the most difficult one and I haven't been able to implement it efficiently.
The fifth student has a secret power, (s)he can teleport between adjacent nodes, so it takes 0 time to travel between them. However, to recharge that secret power, (s)he needs to pass two edges, and (s)he does not have that secret power at the beginning of his/her journey. Unlike other four students, (s)he can use the same edge multiple times, so in the first move, (s)he may go N_1->N_2 and N_2->N_1 to recharge his/her secret power. (S)he cannot store his/her secret power and must use it right away after after gaining it.
The fifth student wants to know the shortest time to reach the summit. At start, (s)he does not have any power, so (s)he needs to pass two edges to recharge it.
The method i tried was a modification of Dijkstra's Algorithm; where instead of moving node by node, from one node it calculates all three possible jumps, only considering the weights of the first two jumps. It considers all cases such as going to a node and coming back to recharge power and jump a highly weighted node. It does work and i do get all the correct answers for the given test cases, but it is SLOW. We are under a two second limit and right now my algorithm takes around 4 seconds for test cases with 50 000 nodes and 100 000 edges.
I'm guessing the problem is within reaching neighbors since there are 3 nested for loops to reach all possible 3 jump away neighbors (while also being able to use the same edges more than once), which basically makes this O(n^3) (But I'm not great with big-oh notation so I'm not sure if it's actually that.)
Does anyone have any ideas to make this algorithm more efficient, or a different algorithm that isn't so slow?
Any help is appreciated!
Here's the code if it's of any help.
long long int HelpStudents::fifthStudent() {
auto start = std::chrono::system_clock::now();
set< pair<long long int,int> >setds;
vector<long long int> dist(totalNodes+15,std::numeric_limits<long long int>::max());
setds.insert(make_pair(0,1));
dist[1] = 0;
bool change = false;
int counter = 0; //these variables were just for checking some things
int max_counter = 0;
int changed_summit = 0;
int operations_after_last_change = 0;
int w1;
int w2;
int db = 0;
vector<int> neighbors;
vector<int> neighbors2;
vector<int> neighbors3;
int u;
while(!setds.empty()){
pair<long long int,int> tmp = *(setds.begin());
setds.erase(setds.begin());
u = tmp.second; //vertex label
if(dist[u] > dist[summit_no]){
continue;
}
if(!change){
counter++;
}else{
counter = 0; //debugging stuff
}
db++;
//cout << db2 << endl;
operations_after_last_change++;
max_counter = max(counter,max_counter);
//cout << "counter: " << counter <<endl;
change = false;
neighbors = adjacency_map[u]; //adjacency map holds a vector which contains the adjacent nodes for the given key
//cout << "processing: " << "(" << tmp.first << ","<< tmp.second << ") " << endl;
for(int nb : neighbors){
w1 = getWeight(u,nb); //this is one jump,
//nb is neighboor
neighbors2 = adjacency_map[nb];
//cout << "\t->" << nb << endl;
if( nb == summit_no){
if(dist[nb] > dist[u] + (w1)){
auto t = setds.find(make_pair(dist[nb],nb));
if(t != setds.end()){
setds.erase(t);
}
dist[nb] = dist[u] + (w1);
change = true;
changed_summit++;
operations_after_last_change = 0;
//cout << "changed summit to " << (dist[u] + (w1)) << endl;
//continue;
}
}
for(int nb2: neighbors2){ //second jump
w2 = getWeight(nb,nb2);
//cout << "\t\t->" << nb2 << endl;
if( nb2 == summit_no){
if(dist[nb2] > dist[u] + (w1+w2)){
auto t = setds.find(make_pair(dist[nb2],nb2));
if(t != setds.end()){
setds.erase(t);
}
dist[nb2] = dist[u] + (w1+w2);
change=true;
changed_summit++;
operations_after_last_change = 0;
//cout << "changed summit to " << (dist[u] + (w1+w2)) << endl;
//continue;
}
}
neighbors3 = adjacency_map[nb2];
for(int nbf: neighbors3){ //third jump, no weight
//cout << "\t\t\t->" << nbf;
if(dist[nbf] > dist[u] + (w1+w2)){
auto t = setds.find(make_pair(dist[nbf],nbf));
if(t != setds.end()) {
setds.erase(t);
}
change = true;
dist[nbf] = dist[u] + (w1+w2);
if(nbf == summit_no){
changed_summit++;
operations_after_last_change = 0;
//cout << endl;
}else{
setds.insert(make_pair(dist[nbf],nbf));
//cout << "\t\t\t\t inserted ("<<dist[nbf]<<","<<nbf<<")" << endl;
}
//cout << "changed " << nbf << " to " << (dist[u] + (w1+w2)) << "; path: "<< u <<" -> "<<nb<<" -> "<<nb2 << " -> " <<nbf << endl;
//setds.insert(make_pair(dist[nbf],nbf));
}else{
//cout << endl;
}
}
}
}
}
auto end = std::chrono::system_clock::now();
std::chrono::duration<double> elapsed_seconds = end-start;
cout << "time passed: "<< elapsed_seconds.count() <<" total loop: "<<db<< endl;
return dist[summit_no];
You make (or more likely imagine) a new directed graph with a node for each unique situation/state that student 5 can be in -- that is combination of original graph node and charge state (0, 1, or 2). Because there are 3 charge states, this graph will have 3 times as many nodes as the original.
Then you use perfectly ordinary Dijkstra's algorithm on this new graph.

performing N independent 1D FFT on a 2D matrix with FFTW

I have a 2 dimensional matrix with each column corresponding to one independent signal. I am going to perform N 1D fft on each column. In matlab, apply a fft to a 2D matrix will do the trick. But I am porting my code to c++ with fftw. I wonder if there is a way to do so. I try the following code by setting the column size to 1 and row size to 4 (total row number), but it does not help.
#include <iostream>
#include <complex>
#include "fftw3.h"
using namespace std;
int main(int argc, char** argv)
{
complex<double> data[4][2];
data[0][0] = complex<double>(1,1);
data[1][0] = complex<double>(2,1);
data[2][0] = complex<double>(3,1);
data[3][0] = complex<double>(4,1);
data[0][1] = complex<double>(1,1);
data[1][1] = complex<double>(1,2);
data[2][1] = complex<double>(1,3);
data[3][1] = complex<double>(1,4);
cout << "original data ..." << endl;
cout << data[0][0] << '\t' << data[0][1] << endl;
cout << data[1][0] << '\t' << data[1][1] << endl;
cout << data[2][0] << '\t' << data[2][1] << endl;
cout << data[3][0] << '\t' << data[3][1] << endl;
cout << endl << endl;
fftw_plan plan=fftw_plan_dft_2d(4, 1,(fftw_complex*)&data[0][0], (fftw_complex*)&data[0][0], FFTW_FORWARD, FFTW_ESTIMATE);
fftw_execute(plan);
cout << "after fftw ..." << endl;
cout << data[0][0] << '\t' << data[0][1] << endl;
cout << data[1][0] << '\t' << data[1][1] << endl;
cout << data[2][0] << '\t' << data[2][1] << endl;
cout << data[3][0] << '\t' << data[3][1] << endl;
return 0;
}
Above code takes the first and second row and reshape them to 2x2 matrix then perform a 2D fft.
Up to now, the only way that comes to my mind is as follow. Let's say I have NxM (N rows, M columns), I create M fftw plans for M 1D fftw. I execute M fftw in serial to get the result. But in practical application, the matrix is very big, M is so large. It is very inefficient to do this way. Any better idea? Thanks.
For those stumbling across this nowadays, the FFTW devs have implemented routines for this operation, which is faster than looping through each column and taking a separate transform. You certainly don't want to take a 2D transform (as is shown in the question), which is mathematically different than row-wise 1D transforms.
The key to you question is in fftw_plan_many_dft. Here is a link to the full documentation.
Here is an example (modifed from the above link) that illustrates what you're looking for.
#include "fftw3.h"
int main() {
fftw_complex *A; // array of data
A = (fftw_complex*) fftw_malloc(sizeof(fftw_complex)*10*3);
// ...
/* Transform each column of a 2d array with 10 rows and 3 columns */
int rank = 1; /* not 2: we are computing 1d transforms */
int n[] = {10}; /* 1d transforms of length 10 */
int howmany = 3;
int idist = 1;
int odist = 1;
/* distance between two elements in the same column */
int istride = 3;
int ostride = 3;
int *inembed = n, *onembed = n;
/* forward, in-place, 1D transform of each column */
fftw_plan p;
p = fftw_plan_many_dft(rank, n, howmany, A, inembed, istride, idist, A, onembed, ostride, odist, FFTW_FORWARD, FFTW_ESTIMATE);
// ...
/* run transform */
fftw_execute_dft(p, A, A);
// ...
/* we don't want memory leaks */
fftw_destroy_plan(p);
fftw_free(A);
}

Why is problemsPathCoordinate vector of a vector of a pair empty?

I have tried to fill a smaller vector of a vector of pairs with some contents from a bigger vector of a vector of pairs without success. Below is the relevant code with couts and their output. Hopefully this is detailed enough.
/*******************Problems Occur*****************/
int iFirst=problemsStartAt;//first index to copy
int iLast=problemsEndAt-1;//last index -1, 11th stays
int iLen=iLast-iFirst;//10-8=2
//if(problemsStartAt!=0)//I.a
if(problemsStartAt!=0&&problemsEndAt!=0)//I.b
{
v_problem_temp=allPathCoordinates[problemsStartAt];
cout<<"266:"<<v_problem_temp.size()<<endl;
cout<<"267:"<<allPathCoordinates.at(1).size()<<endl;
for(vector<pair<int,int>>::iterator it2=v_problem_temp.begin();
it2!=v_problem_temp.end();
++it2)
{
apair=*it2;
point[apair.first][apair.second]=Yellow;
cout<<apair.first<<","<<apair.second<<endl;
}
problemsPathCoordinate.resize(iLen);
cout<<"iLen*sizeof(problemsPathCoordinate):" <<iLen*sizeof(problemsPathCoordinate)<<endl;
memcpy(&problemsPathCoordinate[0],&allPathCoordinates[iFirst],iLen*sizeof(problemsPathCoordinate));
cout<<"279:problemsPathCoordinate.size():"<<problemsPathCoordinate.size()<<endl;
problemsPathCoordinate.resize(iLen);
memcpy(&problemsPathCoordinate[0],&allPathCoordinates[iFirst],iLen*sizeof(problemsPathCoordinate));
cout<<"283:problemsPathCoordinate.size():"<<problemsPathCoordinate[0].size()<<endl;
cout<<"284:problemsPathCoordinate.size():"<<problemsPathCoordinate[1].size()<<endl;
cout<<"286:allPathCoordinates.size():"<<allPathCoordinates.size()<<endl;
cout<<"287:allPathCoordinates.size():"<<allPathCoordinates.size()<<endl;
//from http://stackoverflow.com/questions/35265577/c-reverse-a-smaller-range-in-a-vector
}
Output:
759: path NOT full-filled, number: 8
755: Problems START here at:8
759: path NOT full-filled, number: 9
700: Problems END here at: 11
266:0
267:0
iLen*sizeof(problemsPathCoordinate):72
279:problemsPathCoordinate.size():3
283:problemsPathCoordinate.size():0
284:problemsPathCoordinate.size():0
286:allPathCoordinates.size():79512
287:allPathCoordinates.size():79512
time:39 seconds
Why are the three problemsPathCoordinate elements empty. How to fix it?
Bo
for (vector< vector > >::iterator it = allPathCoordinates.begin(); it != allPathCoordinates.end(); ++it)
{
allPathCoordinates.erase(allPathCoordinates.begin()+5,allPathCoordinates.end()-2);
v_temp = *it;
//cout<<"v_temp.size():"<
for (vector<pair<int,int> >::iterator it2 = v_temp.begin(); it2 != v_temp.end(); ++it2) {
//v_temp.erase(v_temp.begin()+2);
apair = *it2;
//cout << "(" << apair.first << "," << apair.second << ") ; ";
openPoints[apair.first][apair.second]=0;
closedPoints[apair.first][apair.second]=1;
allObstacles[apair.first][apair.second]=Wall;
point[apair.first][apair.second]=Yellow;
}
/

Read the position Actor in PhysX

I have a question about PhysX SDK 2.8.1
I'm an actor:
NxActorDesc actorDesc;
NxBodyDesc bodyDesc;
NxSphereShapeDesc sphereDesc;
sphereDesc.radius = 1.5f;
actorDesc.shapes.pushBack(&sphereDesc);
actorDesc.body = &bodyDesc;
actorDesc.density = 10;
actorDesc.globalPose.t = NxVec3(0.0f, 25.0f, 0.0f);
NxActor *dynamicActor = gsc->createActor(actorDesc);
I want the console print out the current position of the actor. How to do it? This below doesn't work:
for (int i = 0; i <= 10; i++) {
//Step PhysX simulation
if (gsc)
StepPhysX();
NxMat34 pose = dynamicActor->getGlobalPose();
cout <<pose.t << endl;
}
Specifically depends on my reading the position Y.
std::cout can't take an NxVec3 (which your post.t is).
If you want to print out the global position of your dynamicActor,
you need to print out X, Y, Z components of your NxVec3 variable separately.
NxVec3 trans = dynamicActor->getGlobalPose().t; // or you could use "dynamicActor->getGlobalPosition()"
std::cout << trans.x << ", " << trans.y << ", " << trans.z << std::endl;

Resources