PYTHIA-II: A knowledge/database system for managing performance data and recommending scientific software

Often scientists need to locate appropriate software for their problems and then select from among many alternatives. We have previously proposed an a pproach for dealing with this task by processing performance data of the ta rgeted software. This approach has been tested using a customized implement ation referred to as PYTHIA. This experience made us realize the complexity of the algorithmic discovery of knowledge from performance data and of the management of these data together with the discovered knowledge. To addres s this issue, we created PYTHLA-II-a modular framework and system which com bines a general knowledge discovery in databases (KDD) methodology and reco mmender system technologies to provide advice about scientific software/har dware artifacts. The functionality and effectiveness of the system is demon strated for two existing performance studies using sets of software for sol ving partial differential equations. From the end-user perspective, PYTHIA- II allows users to specify the problem to be solved and their computational objectives. In turn, PYTHIA-II (i) selects the software available for the user's problem, (ii) suggests parameter values, and (iii) assesses the reco mmendation provided. PYTHIA-II provides all the necessary facilities to set up database schemas for testing suites and associated performance data in order to test sets of software. Moreover, it allows easy interfacing of alt ernative data mining and recommendation facilities. PYTHIA-II is an open-en ded system implemented on public domain software and has been used for perf ormance evaluation in several different problem domains.