High-performance Computation and Visualization of Plasma Turbulence on Graphics Processors

Direct numerical simulation (DNS) of turbulence is computationally very intensive and typically relies on some form of parallel processing. Spectral kernels used for spat ial discretization are a common computational bottleneck on di stributed memory architectures. One way to increase the effic iency of DNS algorithms is to parallelize spectral kernels using t ightlycoupled Single-Program-Multiple-Data (SPMD) multiprocessor units with minimal inter-processor communication latency. We present techniques to map DNS computations to modern Graphics Processing Units (GPUs), which are characterized by a ve ry high memory bandwidth and hundreds of SPMD processors. We compare and contrast the performance of our parallel algori thm running on a GPU versus the associated CPU implementation of a solver for one of the fundamental nonlinear models of turbul ence theory. We also demonstrate a prototype of a scalable comput ational steering framework based on turbulence simulation and visualization coupling on the GPU.

[1]  Richard D Hazeltine,et al.  Hamiltonian formulation of reduced magnetohydrodynamics , 1983 .

[2]  F. Jenko,et al.  Electron temperature gradient turbulence. , 2000, Physical review letters.

[3]  J. Krommes,et al.  Nonlinear gyrokinetic equations , 1983 .

[4]  Ian Buck.  GPU computation strategies & tricks , 2005, SIGGRAPH Courses.

[5]  Mitsuo Yokokawa,et al.  16.4-Tflops Direct Numerical Simulation of Turbulence by a Fourier Spectral Method on the Earth Simulator , 2002, ACM/IEEE SC 2002 Conference (SC'02).

[6]  A. Hasegawa,et al.  Stationary spectrum of strong turbulence in magnetized nonuniform plasma , 1977 .

[7]  S. Parker,et al.  A fully nonlinear characteristic method for gyrokinetic simulation , 1993 .

[8]  Jens H. Krüger,et al.  A Survey of General‐Purpose Computation on Graphics Hardware , 2007, Eurographics.

[9]  Ramani Duraiswami,et al.  Middleware for programming NVIDIA GPUs from Fortran 9X , 2007 .

[10]  Klaus Mueller,et al.  The lattice-Boltzmann method for simulating gaseous phenomena , 2004, IEEE Transactions on Visualization and Computer Graphics.

[11]  W. Dorland,et al.  Kinetic and fluid turbulent cascades in magnetized weakly collisional astrophysical plasmas , 2007 .

[12]  William Dorland,et al.  Quantitative predictions of tokamak energy confinement from first‐principles simulations with kinetic effects , 1995 .

[13]  Kenneth Moreland,et al.  The FFT on a GPU , 2003, HWWS '03.

[14]  Steven G. Johnson,et al.  The Design and Implementation of FFTW3 , 2005, Proceedings of the IEEE.

[15]  William Dorland,et al.  Astrophysical Gyrokinetics: Basic Equations and Linear Theory , 2005, astro-ph/0511812.

[16]  W. Dorland,et al.  Discrete Particle Noise in Particle-in-Cell Simulations of Plasma Microturbulence , 2005 .