View Source

h2. Scenario
- 3x random autocomplete (most of which with 3-letters), according to a predefined list
- 6x FR Search - either Keyword, Place (thesaurus) or DateRange (with 33% probability each)
- for each FR Search with 30% probability loadCompleteMO() for the first result
- if loadCompleteMO is executed, also get JSON/RDF
- if loadCompleteMO is executed, add an annotation to it

h2. Test setup
- performed on small (8k) and big (115k) BM data sets
- executed with 1, 2, 3 and 5 threads

h2. Results
In the attached bm-perf-stats.xlsx