Summary: Optimized Floe Design

Bullet points of discussed optimization steps:

• Concurrency.

  • Independent processes.

  • Rate limiting factors.

  • The bottleneck problem.

  • The bookkeeping problem.

• Computational resources.

  • Transportation cost.

  • Cube groups and limitations.

• Parallelization.

  • Ensemble calculations.

  • When not to parallelize.

  • Parallel cube groups.

  • end() method.

• Data traffic.

  • Ports and packaging.

  • Buffering vs. batching.

  • Optimizing batch size.

• Collection I/O.

  • Reading collections.

  • Writing collections.