Gyrokinetic particle-in-cell optimization on emerging multi- and manycore platforms

  • Authors:

  • Kamesh Madduri;Eun-Jin Im;Khaled Z. Ibrahim;Samuel Williams;StéPhane Ethier;Leonid Oliker

  • Affiliations:

  • Computational Research Division, Lawrence Berkeley National Laboratory, CA 94720, United States;School of Computer Science, Kookmin University, Seoul 136-702, Republic of Korea;Computational Research Division, Lawrence Berkeley National Laboratory, CA 94720, United States;Computational Research Division, Lawrence Berkeley National Laboratory, CA 94720, United States;Princeton Plasma Physics Laboratory, Princeton, NJ 08543, United States;Computational Research Division, Lawrence Berkeley National Laboratory, CA 94720, United States

  • Venue:
  • Parallel Computing
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

The next decade of high-performance computing (HPC) systems will see a rapid evolution and divergence of multi- and manycore architectures as power and cooling constraints limit increases in microprocessor clock speeds. Understanding efficient optimization methodologies on diverse multicore designs in the context of demanding numerical methods is one of the greatest challenges faced today by the HPC community. In this work, we examine the efficient multicore optimization of GTC, a petascale gyrokinetic toroidal fusion code for studying plasma microturbulence in tokamak devices. For GTC's key computational components (charge deposition and particle push), we explore efficient parallelization strategies across a broad range of emerging multicore designs, including the recently-released Intel Nehalem-EX, the AMD Opteron Istanbul, and the highly multithreaded Sun UltraSparc T2+. We also present the first study on tuning gyrokinetic particle-in-cell (PIC) algorithms for graphics processors, using the NVIDIA C2050 (Fermi). Our work discusses several novel optimization approaches for gyrokinetic PIC, including mixed-precision computation, particle binning and decomposition strategies, grid replication, SIMDized atomic floating-point operations, and effective GPU texture memory utilization. Overall, we achieve significant performance improvements of 1.3-4.7x on these complex PIC kernels, despite the inherent challenges of data dependency and locality. Our work also points to several architectural and programming features that could significantly enhance PIC performance and productivity on next-generation architectures.