Waseda University is currently conducting important research at their Yamaji Laboratory on nuclear reactor design and safety analysis cited in this recent case study provided by HPC Tech. Their research encompasses four major themes including:
- Research on core design of 4th generation reactors
- Design and behavior analysis of accident-tolerant fuel using calculation codes similar to those used for safety review of nuclear power plants
- Large simulation of the Fukushima power plant accident by large-scale system analysis
- Elucidation of molten core behavior mechanism by new numerical analysis method
As you can imagine, a lot of computational resources are required by the lab to undergo this type of research and analysis. Yamaji Laboratory typically utilizes large-scale computation via clusters for this type of research, and they have cited some common issues with this process. Version control of the OS, managing applications, and management of the scheduler is complex, and in many cases, the time spent on administering these items inhibits important time that should otherwise have been dedicated to the research itself. This time, however, Bright Cluster Manager was introduced to simplify all aspects of the cumbersome cluster management process.
Bright Cluster Manager was used in tandem with a 264-core cluster to perform efficient parallel computation and to carry our large-scale simulations. Prior to the introduction of Bright Cluster Manager, these computations were challenging to carry out, limiting the scale in which these analyses could be done. The Yamaji lab states in their case study that in order to perform efficient parallel computation, it is necessary to distribute the computational load efficiently between a total of six workstations. Thanks to Bright Cluster Manager, complicated parallel computations, such as those used in the case study, can be easily performed without complex configurations.
As experienced by the Yamaji Laboratory, Bright Cluster Manager lets you administer clusters as a single entity, making it easier to build and manage reliable clusters from the ground up. Once your cluster is up and running, the intuitive Bright management console lets you see and respond to what’s going on in your cluster at any time.
To read more about the Waseda University case study published by HPC Tech, click here. If you have any questions about Bright Cluster Manager, you can visit our website here, or drop us an email at firstname.lastname@example.org.