Embedded Files
In this page, you can find all experiments run to answer research question 2: How do the proposed optimizations impact the performance of bounded exhaustive generation from an API? by running and comparing different configurations of BEAPI.
As in the article, BEAPI(Default) refers to BEAPI with state matching (M) and builder identification (B) enabled, BEAPI(SM) refers to BEAPI with state matching but no builder identification, and BEAPI(BLD) only activate builder identification . BEAPI(NoOPT) has both techniques disabled.
As in the article, BEAPI(Default) refers to BEAPI with state matching (M) and builder identification (B) enabled, BEAPI(SM) refers to BEAPI with state matching but no builder identification, and BEAPI(BLD) only activate builder identification . BEAPI(NoOPT) has both techniques disabled.
To replicate these experiments you can run the script ./run-testgen-beapi-optimizations.sh <benchmark> <case study> <scope> <config> located in folder scripts (see home page). '<benchmark>' is one of '0_korat', '1_kiasan', '2_roops', '3_fajita', '4_real_world'; `<case study>` is one of the case studies for `<benchmark>` ,`<scope>` is the maximum number of nodes and the number of integers (from 0 to scope-1) available for generation, and `<config>` is one of the four aformentioned 'BEAPI' configurations: 'DEFAULT', 'SM', 'BLD', 'NoOpt'.
Log files after runs for each case study, technique and scope are located in folder scripts/results-optimizations/ in a file with the name results_optimizations.csv
Page updated
Google Sites
Report abuse