1 Performance Analysis and Modeling of the SciDAC GTC Code on Four Large-scale Computer Systems

The Gyrokinetic Toroidal code (GTC) (version 2) is a 3D particle-in-cell application developed at the Princeton Plasma Physics Laboratory to study turbulent transport in magnetic fusion. In this paper, we analyze the performance of the GTC code on four large-scale systems: ORNL Jaguar, UNC RENCI IBM BlueGene/L, SDSC DataStar, and NERSC Seaborg. We discuss the scalability of the GTC code with processor and problem scaling, use processor partitioning to investigate the performance impacts of the application components, and identify possible performance optimizations for further work. Then we use Prophesy to generate performance models for GTC on each platform so that we can use these models to predict the performance on larger number of processors