MPI is an external C library for multiprocessing. It can be used with
Python via mpi4py package. See README.md about
installation and requirements.
Usage:
mpiexec -n 10 python chapter3/scatter_communication.py
Examples:
- Basic usage
- Avoiding deadlocks
- Point to Point communication between two processes
- Broadcast communication between one and many processes (send same data to multiple processes)
- Scatter communication between one and many processes (send different data to multiple processes)
- Gather communication between many and one (send data from many processes to one)
- Alltoall collective communication between all processes
- Reduce communication between many and one (plus reduce on the result)
- Communication optimisation and virtual topology
- Concurrent Futures Pooling - easy way to obtain pool of threads or processes
- Asyncio basics - asynchronous programming with asyncio
- Asyncio coroutines - coroutines using in asyncio
- Asyncio task manipulation - parallel tasks
- Asyncio and Futures - using futures with asyncio