Post-Processing Results

Stage 3 transforms Raw Output Data files from completed experiments into Processed Output Data ready for graph generation in stage 4. Processing is driven by the --proc plugin chain: one or more plugin names passed in order, each executed sequentially. The default chain is proc.statistics proc.collate. For what each built-in plugin does, see New Experiment Results Processing Plugin (--proc). For the data quality and resource options that apply across all plugins, see Core Command Line Reference.

Composing the Chain

Plugins are run in the order given, and order matters. The constraints between the built-in plugins are documented in their respective pages:

• Ordering Considerations

• Ordering Considerations

• Ordering Considerations

• Ordering Considerations

• Ordering Considerations

• Ordering Considerations

• Ordering Considerations

Common compositions:

# Default: statistics then collation
sierra ... --proc proc.statistics proc.collate

# Add imagizing for later video rendering via prod.render
sierra ... --proc proc.statistics proc.collate proc.imagize

# Compress raw outputs after processing to recover disk space
sierra ... --proc proc.statistics proc.collate proc.compress

# Decompress, reprocess, recompress after fixing a stage 3 issue
sierra ... --proc proc.decompress proc.statistics proc.collate proc.compress