2010/08/26 by Michael W. Thomas, Erik Schnetter, Thomas, Michael W. +1
Computer Science · Decision Sciences · #Cloud Computing and Resource Management #Computational Physics (physics.comp-ph) #Distributed #Distributed and Parallel Computing Systems #FOS: Computer and information sciences #FOS: Physical sciences #Parallel #Scientific Computing and Data Management #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.1008.4571
openalex publication_date 2010/08/26 · openalex created_date 2025/10/24 · openalex updated_date 2026/07/28
Computational Science on large high performance computing resources is\nhampered by the complexity of these systems. Much of this complexity is due to\nlow-level details on these resources that are exposed to the application and\nthe end user. This includes (but is not limited to) mechanisms for remote\naccess, configuring and building applications from source code, and managing\nsimulations and their output files via batch queue systems. These challenges\nmultiply in a modern research environment, where a research collaboration spans\nmultiple groups, often in loosely defined international collaborations, where\nthere is a constant influx of new students into multi-year projects, and where\nsimulations are performed on several different resources. The Simulation\nFactory addresses these challenges by significantly simplifying remote access,\nbuilding executables, and managing simulations. By abstracting out the\nlow-level differences between different resources, it offers a uniform\ninterface to these resources. At the same time, it can enforce certain\nstandards for performing simulations that encapsulate best practices from\nexperienced users. Furthermore, SimFactory's automation avoids many possible\nuser errors that can in the worst case render month-long simulations worthless.\n