A data structure for concurrent data management task execution

A data repository for the world fasted tomography device is currently implemented at the synchrotron facility ANKA to speed up and improve investigations with data management services. A data workflow is executed to migrate the measurement data from the measurement PCs to the data repository. Each workflow step like computeChecksum() or copyFiles() is encapsulated into a task class. A relational database persist the current state of execution. In order to execute tasks with long processing time in parallel, a generic but slim parallelization interface has to be developed and implemented.

Your task
Investigate the existing mechanisms for concurrent shared memory data access. Based on the literature search a parallel date structure for concurrent and thread safe task execution has to be implemented. The programming pattern should be based on the Master-Worker pattern. One master is generating work items and workers are waiting for work items, which can be processed in parallel. The final step is an evaluation of the concurrent performance.

We offer
• Modern workstations and large scale data facilities
• A productive and dynamic environment in a team of scientists
• Cooperation with national and international research teams



Weitere Informationen

Unternehmen
Helmholtz Gemeinschaft
Bereich/Abteilung
Institute for Data Processing and Electronics (IPE)
Abschlussart
Bachelorarbeit / Masterarbeit / Diplomarbeit
Branche
Informationstechnologie und -dienste
Schlagwörter
C++ Master-Worker shared memory data access data management services data processing Computer Science
Anforderungen
• Good C++ programming skills
• Interested in shared memory parallelization
• Communication in international and interdisciplinary research teams
Zusatzinformationen
Karlsruher Institut für Technologie

Vertragsdauer: befristet limited according to the study regulations

Eintrittstermin: On appointment

Bewerbung bis: 31.05.2015

Ansprechpartner/in für fachliche Fragen:

Fachliche Auskünfte erteilt Ihnen gerne Alexander Vondrous, IPE, vondrous@kit.edu, Phone: +49721/608-25676





TIPP: Dein Profil wird dem Unternehmen übermittelt. Erziele einen besseren Eindruck, indem Du es vollständig ausfüllst.