MPI Gathering arrays of strings - c++

Question

I'm trying to merge a collection of dictionaries into the root process. Here's a short example:
#define MAX_CF_LENGTH 55
map<string, int> dict;
if (rank == 0)
{
dict = {
{"Accelerator Defective", 33},
{"Aggressive Driving/Road Rage", 27},
{"Alcohol Involvement", 19},
{"Animals Action", 30}};
}
if (rank == 1)
{
dict = {
{"Driver Inexperience", 6},
{"Driverless/Runaway Vehicle", 46},
{"Drugs (Illegal)", 38},
{"Failure to Keep Right", 24}};
}
if (rank == 2)
{
dict = {
{"Lost Consciousness", 1},
{"Obstruction/Debris", 8},
{"Other Electronic Device", 25},
{"Other Lighting Defects", 43},
{"Other Vehicular", 7}};
}
Scatterer scatterer(rank, MPI_COMM_WORLD, num_workers);
scatterer.gatherDictionary(dict, MAX_CF_LENGTH);
The idea inside gatherDictionary() is to put every key in a char array at each process (duplicates are allowed). After that, gathering all keys into the root and creating the final (merged) dictionary before broadcasting it. Here's the code:
void Scatterer::gatherDictionary(map<string,int> &dict, int maxKeyLength)
{
// Calculate destination dictionary size
int numKeys = dict.size();
int totalLength = numKeys * maxKeyLength;
int finalNumKeys = 0;
MPI_Reduce(&numKeys, &finalNumKeys, 1, MPI_INT, MPI_SUM, 0, comm);
// Computing number of elements that are received from each process
int *recvcounts = NULL;
if (rank == 0)
recvcounts = new int[num_workers];
MPI_Gather(&totalLength, 1, MPI_INT, recvcounts, 1, MPI_INT, 0, comm);
// Computing displacement relative to recvbuf at which to place the incoming data from each process
int *displs = NULL;
if (rank == 0)
{
displs = new int[num_workers];
displs[0] = 0;
for (int i = 1; i < num_workers; i++)
displs[i] = displs[i - 1] + recvcounts[i - 1] + 1;
}
char(*dictKeys)[maxKeyLength];
char(*finalDictKeys)[maxKeyLength];
dictKeys = (char(*)[maxKeyLength])malloc(numKeys * sizeof(*dictKeys));
if (rank == 0)
finalDictKeys = (char(*)[maxKeyLength])malloc(finalNumKeys * sizeof(*finalDictKeys));
// Collect keys for each process
int i = 0;
for (auto pair : dict)
{
strncpy(dictKeys[i], pair.first.c_str(), maxKeyLength);
i++;
}
MPI_Gatherv(dictKeys, totalLength, MPI_CHAR, finalDictKeys, recvcounts, displs, MPI_CHAR, 0, comm);
// Create new dictionary and distribute it to all processes
dict.clear();
if (rank == 0)
{
for (int i = 0; i < finalNumKeys; i++)
dict[finalDictKeys[i]] = dict.size();
}
delete[] dictKeys;
if (rank == 0)
{
delete[] finalDictKeys;
delete[] recvcounts;
delete[] displs;
}
broadcastDictionary(dict, maxKeyLength);
}
I'm sure of broadcastDicitonary() correctness as I've already tested it. Debugging into the gathering function I'm getting the following partial results:
Recvcounts:
220
220
275
Displacements:
0
221
442
FinalDictKeys:
Rank:0 Accelerator Defective
Rank:0 Aggressive Driving/Road Rage
Rank:0 Alcohol Involvement
Rank:0 Animals Action
Rank:0
Rank:0
Rank:0
Rank:0
Rank:0
Rank:0
Rank:0
Rank:0
Rank:0
Since only root data is being collected I'm wondering if this has something to do with the characters allocation even if it should be contiguous. I don't think this is related to a missing null character at the end since there's already a lot of padding for each string/key.
Thanks in advance for pointing out any missings or improvements and please comment if you need any extra infos.
If you wish to test it yourself I've put in a one-file only the code all together, it is compile&run ready (of course this works with 3 mpi processes). Code Here

Answer 1

displs[i] = displs[i - 1] + recvcounts[i - 1] + 1;
That + 1 at the end is superfluous. Change it to:
displs[i] = displs[i - 1] + recvcounts[i - 1];