Multithreaded Global Address Space Communication Techniques for Gyrokinetic Fusion Applications on Ultra-Scale Platforms
作者： Robert Preissl , Nathan Wichmann , Bill Long , John Shalf , Stephane Ethier
关键词:
摘要: We present novel parallel language constructs for the communication intensive part of a magnetic fusion simulation code. The focus this work is shift phase charged particles tokamak code in toroidal geometry. introduce new hybrid PGAS/OpenMP implementations highly optimized MPI/OpenMP based kernels. PGAS use an extension standard programming techniques, enabling distribution high loads underlying kernel among OpenMP threads. Building upon lightweight one-sided CAF (Fortran 2008) we also show benefits spreading out over longer period time, resulting reduction bandwidth requirements and more sustained computation overlap. Experiments on up to 130560 processors are conducted NERSC Hopper system, which currently largest HPC platform with hardware support performance improvements 52% at highest concurrency.
acm.org
lbl.gov
sci-hub.se 参考文章(15)
Dan Bonachea, P. Hargrove, GASNet Specification, v1.1 University of California at Berkeley. ,(2002) , 10.2172/1398512
John Reid, Co-array Fortran for Full and Sparse Matrices parallel computing. pp. 61- 61 ,(2002) , 10.1007/3-540-48051-X_7
Cristian Coarfa, Yuri Dotsenko, Jason Eckhardt, John Mellor-Crummey, Co-array Fortran Performance and Potential: An NPB Experimental Study languages and compilers for parallel computing. pp. 177- 193 ,(2003) , 10.1007/978-3-540-24644-2_12
John Mellor-Crummey, Laksono Adhianto, William N. Scherer, Guohua Jin, A new vision for coarray Fortran Proceedings of the Third Conference on Partitioned Global Address Space Programing Models - PGAS '09. pp. 5- ,(2009) , 10.1145/1809961.1809969
Cristian Coarfa, Yuri Dotsenko, John Mellor-Crummey, A Multi-Platform Co-Array Fortran Compiler international conference on parallel architectures and compilation techniques. pp. 29- 40 ,(2004) , 10.5555/1025127.1025994
Robert W. Numrich, Parallel numerical algorithms based on tensor notation and Co-Array Fortran syntax parallel computing. ,vol. 31, pp. 588- 607 ,(2005) , 10.1016/J.PARCO.2005.03.003
W. X. Wang, Z. Lin, W. M. Tang, W. W. Lee, S. Ethier, J. L. V. Lewandowski, G. Rewoldt, T. S. Hahm, J. Manickam, Gyro-kinetic simulation of global turbulent transport properties in tokamak experiments Physics of Plasmas. ,vol. 13, pp. 092505- ,(2006) , 10.1063/1.2338775
Dan Bonachea, Jason Duell, Problems with using MPI 1.1 and 2.0 as compilation targets for parallel language implementations ieee international conference on high performance computing data and analytics. ,vol. 1, pp. 91- 99 ,(2004) , 10.1504/IJHPCN.2004.007569
Cristian Coarfa, Yuri Dotsenko, John Mellor-Crummey, Experiences with Sweep3D implementations in Co-array Fortran The Journal of Supercomputing. ,vol. 36, pp. 101- 121 ,(2006) , 10.1007/S11227-006-7952-7
J. N. Leboeuf, V. E. Lynch, B. A. Carreras, J. D. Alvarez, L. Garcia, Full torus Landau fluid calculations of ion temperature gradient-driven turbulence in cylindrical geometry Physics of Plasmas. ,vol. 7, pp. 5013- 5022 ,(2000) , 10.1063/1.1323757