ORNL has a long history of performing in-depth assessments of HPC architectures using important mission-based applications. This team performs ongoing evaluation and assessment of technology for high performance computing. The team has expertise in benchmarking, system performance evaluations, workload monitoring, use of application modeling tools, and future algorithm scaling and technology assessment. Using scientific methods, the team develops methods for analyzing possible technical alternatives and creates a clear understanding of current and future HPC workloads. Our methodology ranges from using low level microbenchmarks to isolate component performance characteristics up to full applications, executing across some of the world’s largest HPC systems. Our evaluations include commodity systems, large-scale HPC systems, and novel architectures, such as graphics processors and FPGAs.
Our most recent processor evaluation looked at the tradeoffs in AMD's Llano APU.