OpenMP Seg Fault

[cpp82]

I have a program which finds similar strings from a database consists of million strings of same length m. The input to my program is a string of length n (>>> m). The program checks for each substring of length m. It is implemented in c++ and uses OpenMP, but I am getting segmentation fault for the following piece of code. The string contains only four characters A, B, C and D. The program finds the similar strings and if the hamming distance is less than given value, then it counts the occurrences of each character (stores it in ARR)

  
std::map<char, int> mapp = {{'A', 0}, {'B', 1}, {'C', 2}, {'D', 3}};
int **ARR = (int**) malloc(sizeof(int*)*n);
for(int z=0; z<n; z++)
  ARR[z] = (int*) malloc(sizeof(int)*4);
for(int i=0; i<n; i++)
  for(int j=0; j<4; j++) ARR[i][j] = 0; 

vector<string> V;
#pragma omp parallel for private(V) shared(ARR) 
for(int i=0; i<str.length()-m-1; i++){  
  std::string substr = str.substr(i, m);
  for(int j=0; j<N; j++){  // N number of databases to search
    V = retrieve(substr);  // find similar strings
    for(int k=0; k<(int)V.size(); ++k) {
    int hd = hamming_distance(substr, V[k]); // no. of mismatches between substr and kth element of V
    if( h < minHD){
      for(int l=0; l<m; l++){
        char c = V[k][l];
        #pragma omp critical
        {
           ARR[j+l][mapp.at(c)]++;
         }
       }
     }
   }
 }
}

I am getting Seg fault for the above code. Any help would be appreciated

[Enoizat]

Hi, cpp82.
Your problem (a lot of strings which can be made only by 4 characters) looks like something which relates with DNA.
I’m sorry I can’t help you, mainly because I haven’t understood what you are trying to do.
I suspect this line could be the one which tries to access beyond your array limits:
ARR[j+l][mapp.at(c)]++;
I think the simplest thing would be to substitute vectors, which you already use, for your C-style arrays, because they throw exceptions, which are a lot more informative.
Plus, with vectors code become much easier to write (and to read).

Since I couldn’t understand your code, even because I don’t know what a “hamming distance” is, I can’t come up with a helpful example; anyway I hope the following one can give you some hints:

#include <iostream>
#include <limits>
#include <map>
#include <vector>

std::size_t makeAllStringsSameLength(std::vector<std::string> phrase);

int main()
{
   std::map<char, int> charintmap;
   std::vector<std::vector<int>> myvec;
   for(char c='A'; c<='Z'; c++) {
      charintmap.emplace(std::make_pair(c, int((c-'A'+1))));
      myvec.emplace_back(std::initializer_list<int>{c, 0});
   }
   for(char c='a'; c<='z'; c++) {
      charintmap.emplace(std::make_pair(c, int((c-'a'+'Z'-'A'+2))));
      myvec.emplace_back(std::initializer_list<int>{c, 0});
   }

   for(const auto& couple : charintmap) {
      std::cout << "charintmap.at(" << couple.first
                << ") == " << couple.second << '\n';
   }

   std::cout << "\nPress ENTER to continue...\n";
   std::cin.ignore(std::numeric_limits<std::streamsize>::max(), '\n');

   for(size_t row{0}; row<myvec.size(); row++) {
      std::cout << "myvec.at(" << row << ").at(0) == "
                << char(myvec.at(row).at(0))
                << "; myvec.at(" << row << ").at(1) == "
                << myvec.at(row).at(1) << '\n';
   }

   std::cout << "\nPress ENTER to continue...\n";
   std::cin.ignore(std::numeric_limits<std::streamsize>::max(), '\n');

   std::vector<std::string> words = { "next", "time", "please",
                                      "provide", "a", "compilable",
                                      "piece", "of", "your",
                                      "code", "along", "with",
                                      "examples", "of", "the",
                                      "data", "you", "are",
                                      "using" };

   // 'wordlen' would inform you about how long strings are, if needed.
   int wordlen = makeAllStringsSameLength(words);
   for(const auto& word : words) {
      for(const auto& ch : word) {
         int index{0};
         if('A' <= ch && ch <= 'Z') index = int('Z' - ch);
         if('a' <= ch && ch <= 'z') index = int('z' - ch);
         myvec.at(index).at(1) += charintmap.at(ch);
      }
   }

   for(size_t row{0}; row<myvec.size(); row++) {
      std::cout << "myvec.at(" << row << ").at(0) == "
                << char(myvec.at(row).at(0))
                << "; myvec.at(" << row << ").at(1) == "
                << myvec.at(row).at(1) << '\n';
   }

   return 0;
}

std::size_t makeAllStringsSameLength(std::vector<std::string> phrase)
{
   std::size_t longest{0};
   for(auto& word : phrase) {
      if(word.length() > longest) {
         longest = word.length();
      }
   }

   for(auto& word : phrase) {
      std::size_t spaces = longest - word.length();
      switch (spaces) {
         case 0:
            break;
         case 1: // 1 / 2 == 0 --> integer division
            word.insert(0, 1, ' ');
            break;
         default:
            word.insert(0, spaces/2, ' ');
            // now word.length != wlen
            word.insert(word.length(), longest-word.length(), ' ');
            break;
      }

   }

   for(auto& word : phrase) {
      std::cout << word << ": " << word.length() << '\n';
   }
   std::cout << "\nPress ENTER to continue...\n";
   std::cin.ignore(std::numeric_limits<std::streamsize>::max(), '\n');

   return longest;
}