The optimum number of groups for artificial groups for parallel processing should match the maximum number of processes allowable considering your hardware and license, assuming your workspace can process that many groups at once.

• A. False
• B. Not sure
• C. It depends
• D. True

Answer: (D)

Maximizing parallelism by using all available processing slots is the guiding principle for tuning FME workflows for parallel execution. If your hardware and license allow a certain number of processes and your workspace can actually run that many groups at once, then configuring the artificial groups to match that ceiling typically yields the highest throughput because all cores and licensed capacity are utilized. Real-world performance can be affected by memory usage, I/O, or per-group overhead, but under the given assumption that the workspace can process that many groups concurrently, matching the maximum allowable parallelism is the best approach.