Thread Data Race

Question

Hi

I have a thread pool which has 12 threads.

I have a function which is used as a thread function for all these twelve threads. Last argument is output parameter which gets filled by thread function.

Thread function is something as below:

void myFunc(int a,int b,int c,vector<int>& out);

out vector is filled by this function.

vector<thread> vThreads;
vThreads.resize(12);
vector<vector<int>> resVals(12);

for (int iCurThread = 0; iCurThread < intThreadCnt; ++iCurThread)
{
vThreads.at(iCurThread) = thread(&myFunc,1,2,3, ref(resVals[iCurThread]));
}

for (std::thread& active_thread : vThreads) {
active_thread.join();
}

Is it right that there is no data race in this case even though I have not used mutex?

I think no data race as each thread is modifying specific vector from vector of vector. Right?

Answer 1

I dont think that happen some data race because you are using a distinct vector for thread and the data is an native type as int (no shared obects then)

Answer 2

Yeah vector is different for each thread ... should not be a data race. But I am confused as each vector passed to a thread is a member of same vector of vector.

Main container is same and specified idexed vector is passed to thread function