Parallel implementation of matrix multiplication using Microsoft MPI
Intelligent Systems Design / 6. Semester
The multiplication of two matrices is a task that can be effectively parallelized. As processors will not become much faster in frequency but instead will grow into width with more cores and processes running at the same time, this becomes more and more important. Practically this is very interesting especially for the computations inside neural networks. This implementation creates two matrices filled with pseudorandom integers and multiplies them in n different processes on your CPU.
mpiexex -n n = Count of wanted processes
- Parameter: Seed
- Parameter: LogLevel
- Parameter: CommonSize
- Parameter: ResultWidth
- Parameter: ResultHeight
Microsoft MPI