Difference between revisions of "Star Exec"
J.waldmann (talk | contribs) (Created page with "[http://www.starexec.org/ StarExec] is a cross-community solver execution and benchmark library service under joint development at the University of Iowa and the University o...") |
|||
| Line 7: | Line 7: | ||
(Manchester). We use this wiki page here to collect and organize some points | (Manchester). We use this wiki page here to collect and organize some points | ||
that we want to make during that workshop. | that we want to make during that workshop. | ||
| + | |||
| + | Open Questions: | ||
| + | * Format of results: The specification does not speak explicitly about the format of the results. Is this going to be handled using Post-Processors? | ||
| + | * Scoring systems: How can a scoring system like we use for the complexity competition be implemented? Post-processors seem to be restricted to working on one solver/result pair, while our scoring needs to see all results in a competition for a given example. | ||
| + | * Will there be "Teams"? StarExec seems to know Communitys and single users, but not about the fact that some tools are produced by a group of people. | ||
Revision as of 14:19, 12 June 2012
StarExec is a cross-community solver execution and benchmark library service under joint development at the University of Iowa and the University of Miami. Its goal is to facilitate the experimental evaluation of logic solvers and other automated reasoning tools by providing a shared storage and computing infrastructure to store, manage and make available benchmark libraries; execute comparative evaluations; and run solver competitions. StarExec should be available for public use in the fall of 2012.
The termination community intends to use this platform for future competitions (and other tasks, like running experiments, archiving benchmarks and results, etc.)
There will be a Starexec Workshop affiliated with IJCAR 2012 (Manchester). We use this wiki page here to collect and organize some points that we want to make during that workshop.
Open Questions:
- Format of results: The specification does not speak explicitly about the format of the results. Is this going to be handled using Post-Processors?
- Scoring systems: How can a scoring system like we use for the complexity competition be implemented? Post-processors seem to be restricted to working on one solver/result pair, while our scoring needs to see all results in a competition for a given example.
- Will there be "Teams"? StarExec seems to know Communitys and single users, but not about the fact that some tools are produced by a group of people.
