10.3. Algorithm parallelization¶
This tutorial explains how to use the cooperative multitasking OS for introducing concurrency. The same OS can be used to outsource concurrent tasks to other procesors to achieve parallel processing. Several methods for dispatching tasks are described. Finally an example is given where a large number of optimization runs is distributed among multiple processor that run them in parallel.
- 10.3.1. Context switching in cooperative multitasking
- 10.3.2. Remote tasks (task outsourcing)
- 10.3.3. Dispatching a set of tasks and collecting the results
- 10.3.4. Dispatching an unknown number of tasks, collecting results with a results collector
- 10.3.5. Writing a custom dispatcher
- 10.3.6. Performing many differential evolution runs on a cluster of workstations