Benchmark Actions

The page describes actions that the benchmark software can perfrom and which phases (in file: test.properties) need to be enabled. Each benchmark action consists of one or several phases described here: https://github.com/ldbc/ldbc_spb_bm/wiki/Operational-Phases

• Actions:
  • Load Reference Knowledge (Data) and Ontologies - Loads the reference data and ontologies distributed with the benchmark software. That action is required to be started in combination with other actions e.g. when generating data, loading ontologies and reference data should also be enabled.
    Enable phases : loadOntologies, loadReferenceDatasets
  • Generate Data - Generates synthetic data used to populate the database and measures performance.
    Enable phases : loadOntologies, loadReferenceDatasets, generateCreativeWorks
  • Load Generated Data - The benchmark driver can optionally load generated synthetic data into the database. That can also be achieved by loading generated data manually. Generated data is saved in a folder defined by parameter 'creativeWorksPath'. The benchmark driver will attempt to start all executable script files (files with extension .sh or .bat) saved in folder '/data/scripts/postLoad'. It is not necessary to provide such scripts. Requires action : Generate Data. Enable phase : loadCreativeWorks
  • Generate Query Substitution Parameters - Generates substitution parameters used in queries during the benchmark run. Requires actions : Generate Data and Load Generated Data. Enable phase : generateQuerySubstitutionParameters
  • Validate Query Results - validates results of all editorial operations and each query; to be executed on an empty database.
    Enable phases : loadOntologies, loadDatasets, validateQueryResults
  • Warm-up - executes mixes of queries without measuring performance. This is an optional action. Requires actions : Generate Data, Load Generated Data, Generate Query Substitution Parameters. Enable phases : warmUp
  • Run The Benchmark - measures performance of RDF database. Requires actions : Generate Data, Load Generated Data, Generate Query Substitution Parameters. Enable phases : runBenchmark
  • Run Online Replication and Backup Benchmark - Experimental action which measures performance of RDF databases during ongoing replication and backup. For more details consult the reference documentation. Requires actions : Generate Data, Load Generated Data, Generate Query Substitution Parameters. Enable phase : runBenchmarkOnlineReplicationAndBackup.
  • Check Conformance to OWL2-RL Rule-Set - Tests the inference engine for conformance to the OWL2-RL rule-set. To be started on empty databases supporting the OWL2-RL rule-set, no data generation or loading is required. Enable phase : loadOntologies.

Following diagram depicts the flow of actions and corresponding properties used to control it.