I have the following simple code. I allocate dynamically memory for 3 doubles, I assign to each double a number and after I deallocate the memory but as one can see if runs the code the only difference before and after the deletion (delete[] x) and the only difference is for the first double of the vector. I can't understand why the content of the first element of the vector changed and the content of x remained the same with the same address of memory.
#include <iostream>
#include <cmath>
int main(int argc, char * argv[])
{
double * x;
x = new double [3];
x[0] = 1; x[1]=3; x[2]=5;
std::cout << x[0] << " " << x[1] << " " << x[2] << "\n";
std::cout << x << "\n";
delete[] x;
std::cout << x[0] << " " << x[1] << " " << x[2] << "\n";
std::cout << x << "\n";
return 0;
}
To my understanding, this is undefined behaviour; x is read after it is deleted.
Related
//How do I count numbers in the input vector that is greater and lesser than input number X?
#include <iostream>
#include
#include
using namespace std;
int main()
{
int i, size;
float verdi = 0;
vector<float> minvektor;
cout << "input vectorsize " << endl;
cin >> size;
cout << "Input vector-numbers, use enten space or enter to space/divide numbers. " << endl;
for (i = 0; i < size; i++)
{
cin >> verdi;
minvektor.push_back(verdi);
}
for (i = 0; i < size; i++)
{
cout << minvektor[i] << " ";
}
// -------------------------------------------------
cout << "\ninput number X to compare to vector: " << endl;
float X;
cin >> X;
//count numbers in a vector that is greater than X
//output ex: 5 numbers in a vector that is greater than number X
}
in an R script I source a cpp file to make some calculations. In that R script, a function defined in the cpp file is called and a matrix and an integer is provided. After a few rounds through the loop it gives the error "Not a matrix" (in line of code resid = (x(_,j) - x(_,i))*(x(_,j) - x(_,i));), even though for the rounds before it worked.
R script:
## all together
# rm(list=ls())
library(RcppArmadillo)
library(Rcpp)
sourceCpp("~/test.cpp",verbose = FALSE)
cat("start loop")
for(n in c(45:46)){
cat("\n", n, "\n")
p_m <- matrix(data=rnorm(n^2,1,1),nrow = n, ncol=n)
print(class(p_m))
print(some_function(p_m,nosamples=10))
}
cat("finished")
I start this R script via the command line. R version R-4.1.0. In R-Studio it crashes with a fatal error.
The cpp file:
// [[Rcpp::depends(RcppArmadillo)]]
#include <RcppArmadillo.h>
using namespace Rcpp;
// [[Rcpp::export]]
NumericVector some_function(NumericMatrix x,int nosamples) {
int ncol = x.ncol();
NumericVector out2(nosamples);
int loops;
int loops2;
double result=0;
NumericVector::iterator it;
double acc = 0;
NumericVector resid(ncol);
NumericVector out(ncol*(ncol-1)/2);
loops2=0;
std::cout << nosamples << std::endl;
std::cout << (ncol-1) << std::endl;
std::cout << ncol*(ncol-1)/2 << std::endl;
while(loops2 < (nosamples)){
std::cout << "loops2:" << std::endl;
std::cout << loops2 << std::endl;
loops=0;
int i;
int j;
for(j=0;j<(ncol-1);++j){
std::cout << " j: " << j << std::endl;
for (i = (j+1); i < (ncol); ++i) {
std::cout << " i: " << i << std::endl;
resid = (x(_,j) - x(_,i))*(x(_,j) - x(_,i)); //here it stops
std::cout << " i: " << i << std::endl;
for(int ii=0; ii<ncol;++ii){
acc += resid[i];
}
result=sqrt(acc);
loops += 1;
out[loops] = result;
std::cout << " i: " << i << std::endl;
}
}
std::cout << "loops:" << std::endl;
std::cout << loops << std::endl;
out = out[out > 0];
it = std::min_element(out.begin(), out.end());
out= *it;
std::cout << out << std::endl;
loops2 += 1;
out2[loops2]=out[0];
}
std::cout << "cpp finished" << std::endl;
return(out2);
}
Can someone explain what the problem is about?
Thanks and kind regards
Edit
I adapted some things in the cpp file (shown below) and the error disappeared. First I thought, everything is fine. But when I increase the number of loops, another problem occurs: the function breaks, but no error is shown. It breaks after loop number 543 ("loop2: 543"). At least it does the same in each round of the while loop with the same data.
I adapted the R-script and the ccp file to make this problem (at least on my machine) reproducible.
I know this code seems to be somehow meaningless, but it is part of a bigger program and I wanted to give here a minimum example.
The R script:
## all together
# rm(list=ls())
library(RcppArmadillo)
library(Rcpp)
sourceCpp("~/test.cpp",verbose = FALSE)
cat("start loop")
for(n in c(100:101)){
cat("\n", n, "\n")
p_m <- matrix(data=rnorm(n^2,1,1),nrow = n, ncol=n)
print(class(p_m))
print(some_function(p_m,nosamples=800))
}
cat("finished")
The cpp file:
// [[Rcpp::depends(RcppArmadillo)]]
// [[Rcpp::depends(RcppEigen)]]
#include <RcppArmadillo.h>
#include <RcppEigen.h>
using namespace Rcpp;
using Eigen::Map;
using Eigen::VectorXd;
typedef Map<VectorXd> MapVecd;
// [[Rcpp::export]]
NumericVector some_function(NumericMatrix x,int nosamples) {
int ncol = x.ncol();
NumericVector out(ncol*(ncol-1)/2);
NumericVector out2(nosamples);
NumericVector out3(ncol*(ncol-1)/2);
NumericVector resid(ncol);
int loops;
int loops2;
double result=0;
double acc = 0;
int show_cout=0;
loops2=0;
std::cout << nosamples << std::endl;
std::cout << (ncol-1) << std::endl;
std::cout << ncol*(ncol-1)/2 << std::endl;
while(loops2 < (nosamples)){
std::cout << "loops2:" << loops2 << std::endl;
loops=0;
int i;
int j;
for(j=0;j<(ncol-1);++j){
// std::cout << " j: " << j << std::endl;
for (i = (j+1); i < (ncol); ++i) {
if(show_cout==1){
std::cout << " i: " << i << std::endl;
}
resid = (x(_,j) - x(_,i))*(x(_,j) - x(_,i));
if(show_cout==1){
std::cout << " i: " << i << std::endl;
}
for(int ii=0; ii<ncol;++ii){
acc += resid[ii];
}
result=sqrt(acc);
loops += 1;
out[loops] = result;
if(show_cout==1){
std::cout << " i: " << i << std::endl;
}
}
}
// std::cout << "loops:" << loops << std::endl;
//
out = out[out > 0];
const MapVecd xy(as<MapVecd>(out));
out3=xy.minCoeff();
out2[loops2]=out3[0];
loops2 += 1;
}
std::cout << "cpp finished" << std::endl;
return(out2);
}
Two things here:
Use out[loops++] = result; instead of loops += 1; out[loops] = result; because you were starting at 1, and probably accessing the last element outside of the range of this vector.
Use
for(int ii=0; ii<ncol;++ii){ double eps = x(ii, j) - x(ii, i); acc += eps * eps; }
instead of relying on this resid vector.
I need to send array pieces to all processes using MPI_Scatter then to get sum of all elements. Where should I initialize array then to scatter it? In root rank?
If I initialize array on root rank then other ranks dont get their data. Otherway I can initialize array for everyone (out of if(rank == root)...else), but it means, that I create array several times.
#include <stdio.h>
#include <stdlib.h>
#include <mpi.h>
#include <iostream>
#include <time.h>
using namespace std;
int main(int argc, char* argv[])
{
int size;
int rank;
srand(time(NULL));
MPI_Init(&argc, &argv);
MPI_Comm_size(MPI_COMM_WORLD, &size);
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
int arr_size = size * 2;
int block = arr_size / (size);
int* B = new int[block];
if (rank == 0)
{
int* A = new int[arr_size];
cout << "generated array: " << endl;
for (int i = 0; i < arr_size; i++)
{
A[i] = rand() % 100;
cout << A[i] << " ";
}
cout << endl;
MPI_Scatter(A, block, MPI_INT, B, block, MPI_INT, 0, MPI_COMM_WORLD);
}
cout << "process " << rank << " received: " << endl;
for (int i = 0; i < block; i++)
{
cout << B[i] << " ";
}
cout << endl;
int local_sum = 0;
for (int i = 0; i < block; i++)
{
local_sum += B[i];
}
cout << "sum in process " << rank << " = " << local_sum << endl;
cout << endl;
int global_sum;
MPI_Reduce(&local_sum, &global_sum, 1, MPI_INT, MPI_SUM, 0, MPI_COMM_WORLD);
if (rank == 0)
{
cout << "sum = " << global_sum << endl;
}
MPI_Finalize();
return 0;
}
I get something like this (only root rank got its data):
process 1 received:
process 3 received:
-842150451 -842150451
-842150451 -842150451
sum in process 1 = -1684300902
sum in process 3 = -1684300902
process 2 received:
-842150451 -842150451
sum in process 2 = -1684300902
process 0 received:
4 9
sum in process 0 = 13
sum = -757935397
MPI_Scatter() is a collective operation and must hence be invoked by all the ranks.
Declare int *A = NULL; on all ranks and only allocate and populate on rank zero.
int* A = NULL;
int* B = new int[block];
if (rank == 0)
{
A = new int[arr_size];
cout << "generated array: " << endl;
for (int i = 0; i < arr_size; i++)
{
A[i] = rand() % 100;
cout << A[i] << " ";
}
cout << endl;
}
MPI_Scatter(A, block, MPI_INT, B, block, MPI_INT, 0, MPI_COMM_WORLD);
Memory address shows two different values.
We have const variable (a) and we put the address of the variable into two pointers (b and c). After changing the value at the address in one of the pointers (c), we had a situation where the same memory address has two different values.
Is there any explanation for this behavior?
#include <iostream>
int main(void)
{
const int a = 99;
const int *b = &a;
int *c = (int *)b;
std::cout << &a << " - " << a << '\n';
std::cout << b << " - " << *b << '\n';
std::cout << c << " - " << *c << "\n\n";
*c = 61;
std::cout << &a << " - " << a << '\n';
std::cout << b << " - " << *b << '\n';
std::cout << c << " - " << *c << '\n';
return 0;
}
//here are the result(output)
003CFAA4 - 99
003CFAA4 - 99
003CFAA4 - 99
003CFAA4 - 99
003CFAA4 - 61
003CFAA4 - 61
I made a small reproducible example:
#include <Rcpp.h>
using namespace Rcpp;
class Index {
public:
Index(int i_) : i(i_) {}
int getI() { return i; }
private:
int i;
};
// [[Rcpp::export]]
SEXP getXPtrIndex(int i) {
Rcout << "getXPtrIndex: i = " << i << std::endl;
Index ind(i);
Rcout << "getXPtrIndex: ind.i = " << ind.getI() << std::endl;
return XPtr<Index>(&ind, true);
}
// [[Rcpp::export]]
void getXPtrIndexValue(SEXP ptr) {
XPtr<Index> ind_ptr(ptr);
Rcout << "getXPtrIndexValue: ind_ptr->i = " << ind_ptr->getI() << std::endl;
Index ind = *ind_ptr;
Rcout << "getXPtrIndexValue: ind.i = " << ind.getI() << std::endl;
}
Basically, I define a small class, along with a function to get an external pointer of an element of this class. The last function is used to print the weird accessor when returning the class element back to C++.
Results in R:
> (extptr <- getXPtrIndex(10))
getXPtrIndex: i = 10
getXPtrIndex: ind.i = 10
<pointer: 0x7ffeeec31b00>
> getXPtrIndexValue(extptr)
getXPtrIndexValue: ind_ptr->i = 33696400
getXPtrIndexValue: ind.i = 0
Why can't I access 10?
I'm using Rcpp version 0.12.12 (the latest I think).
It seems to have something to do with the temporary object---by the time your second function runs the "content" of the first is already gone.
So either just make
Index ind(10);
a global, and comment out the line in your first function. Then all is peachy (I changed the R invocation slightly):
R> extptr <- getXPtrIndex(10)
getXPtrIndex: i = 10
getXPtrIndex: ind.i = 10
R> getXPtrIndexValue(extptr)
getXPtrIndexValue: ind_ptr->i = 10
getXPtrIndexValue: ind.i = 10
R>
Or it also works the same way when you make you Index object static to ensure persistence. Corrected example below.
#include <Rcpp.h>
using namespace Rcpp;
class Index {
public:
Index(int i_) : i(i_) {}
int getI() { return i; }
private:
int i;
};
// [[Rcpp::export]]
SEXP getXPtrIndex(int i) {
Rcout << "getXPtrIndex: i = " << i << std::endl;
static Index ind(i);
Rcout << "getXPtrIndex: ind.i = " << ind.getI() << std::endl;
return XPtr<Index>(&ind, true);
}
// [[Rcpp::export]]
void getXPtrIndexValue(SEXP ptr) {
XPtr<Index> ind_ptr(ptr);
Rcout << "getXPtrIndexValue: ind_ptr->i = " << ind_ptr->getI() << std::endl;
Index ind = *ind_ptr;
Rcout << "getXPtrIndexValue: ind.i = " << ind.getI() << std::endl;
}
/*** R
extptr <- getXPtrIndex(10)
getXPtrIndexValue(extptr)
*/