Result checking in global computing systems
ICS '03 Proceedings of the 17th annual international conference on Supercomputing
Software—Practice & Experience
AP2PC'04 Proceedings of the Third international conference on Agents and Peer-to-Peer Computing
| Hi-index | 0.00 |
Individual-based models in evolutionary biology easilylead to multi-parameter applications that need globalcomputing power to exploit their full potential.Mainlydue to varying population size parameters, they easilygenerate computational complexities from less than asecond to more than 100 years in case of theSimulator005 of evolution@home.The poorlyunderstood biology of the system leads to automatedpredictions that may be way off.This report describesfirst experiences of a global computing system, whereusers can choose between tasks of different complexity.Besides theoretical complexity limits of tasks that fitglobal computing, choices of users are analyzed.Potential of incomplete results to increase predictionaccuracy is discussed as well as benchmarking computersystems that vary nearly 2 orders of magnitude in theiridle processing power.Finally, prediction accuracy isanalyzed with the help of a newly defined parameter:error of magnitude.It is concluded, that globalcomputing has great potential for projects with poorlypredictable single-run-complexities, if frameworks aredesigned to allow users to choose their commitment,and if they make use of incomplete results to improvepredictions.