multithreading - Parallel computing in Java for big data -

there 2 sets of matrices i.e. a(1 500 matrices) , b(150 000 matrices).

for each matrix set a, iterate through elements b b checking criteria on , b (if it's true matrix b, stop iterating , return true. if no match found after iterating entire set b, return false).

my question how make program parallel using java?

i understand matrices sets quite big (the numbers of elements posted above instance, in real-life case bigger). how correctly computations? create 1 5000 threads , computation not idea. optimal amount of threads, take computational work?

as said, may need check pair , there no heuristic order of checking if checking not need i/o or sth cause wast of cpu, if have 4 example 4 core, create 4 thread(0-3) , thread responsible checking a(4k+i) , every b. , before each checking in each thread, must check whether find true pair or not, can accomplished static variable in classes extends thread

public class worker extends thread{     static int found=0;     /**      * @param args command line arguments      */     list<matrix> a;     list<matrix> b;     int myid;     int corenumber;     worker(list<matrix> a, list<matrix> b, int myid,int corenumber){         this.a=a;         this.b=b;         this.myid=myid;         this.corenumber=corenumber;      }      @override     public void run() {         for(int i=myid;i<a.size();i+=corenumber){             for(int j=0;j<b.size();j++){                 if(found==1){                     return;                 }                 if(check(a.get(i),b.get(j))){                     found=1;                     return;                 }             }         }     }  } 

you need wait until threads stop working.