Accelerating Time-to-Results with Parallel I O
As organizations rapidly expand their Linux based compute cluster deployments, many loose engineering productivity when their CFD users must wait for serial read/write operations, and large file transfers to complete. The challenge is to provide a scalable, cost-effective storage solution that can easily deliver high performance concurrent read and write access generated by CFD simulations. In addition, the system needs to scale simply with the increasing volumes of data and continue to scale performance as the storage and clusters grow. Panasas and ANSYS have worked together to optimize ANSYS Fluent to take full advantage of the performance and scalability features of Panasas ActiveStor parallel storage system. Panasas will discuss how high-performance parallel storage architectures and cost-effective, scale-out designs can meet design requirements, along with the results of performance benchmark tests and a description of how ANSYS customers have accelerated their time-to-results while simplifying storage operations and management.