Tiger Team: tt-stata
Computation of large problems with STATA
Scientists experience difficulties on bwHPC clusters for large datasets or large problem sizes. In one case, an optimization run exceeded the maximal permissible wall clock time. To solve this we investigated ways to save intermediate results by means of STATA tools and reuse them in following jobs. In an other case with large data sets we could identify the node memory limit as cause for job failure. Using nodes with more memory (more than 1TB) we could solve this issue.
Members of the Tiger-Team:
Jonathan Eklund (Uni Tübingen, Prof. Dr. Georg Wamser, Lehrstuhl für Finanzwissenschaft, VWL),
Leo Schmallenbach (Studies on Team Diversity, Uni Mannheim, Prof. Dr. Karin Hoisl, Lehrstuhl für Organisation und Innovation, BWL),
HPC Competence Center MLS&WISO (Uni Mannheim)