Benchmarking of communication techniques for GPUs

Journal of Parallel and Distributed Computing

Volumn 73, Issue 2, 2013, Pages 250-255

  Bernaschi, M ^a   Bisson, M ^a   Rossetti, D ^b  

^a CNR   (Italy)

^b INFN SEZIONE DI ROMA   (Italy)

Author keywords

3D Torus; Asynchronous communication; GPU; MPI

Indexed keywords

3D TORUS; APPLICATION LEVEL; ASYNCHRONOUS COMMUNICATION; COMMUNICATION TECHNIQUES; GPU; GRAPHIC PROCESSING UNITS; HEISENBERG SPIN GLASS; HIGH-PERFORMANCE INTERCONNECT; INFINIBAND; MPI; NVIDIA CUDA; SINGLE SPIN;

ALGORITHMS; APPLICATION PROGRAMMING INTERFACES (API); PROGRAM PROCESSORS; THREE DIMENSIONAL COMPUTER GRAPHICS;

COMMUNICATION;

EID: 84870952042     PISSN: 07437315     EISSN: None     Source Type: Journal    
DOI: 10.1016/j.jpdc.2012.09.006     Document Type: Article

References (13)

1. S. Adler Over-relaxation method for the Monte Carlo evaluation of the partition function for multiquadratic actions Phys. Rev. D 23 1981 2901 2904
2. R. Ammendola, et al. APEnet: a 3D toroidal network enabling petaFLOPS scale Lattice QCD simulations on commodity clusters, in: Proceedings of Lattice 2010 conference, PoS LAT2010, 2010, p. 022.
3. R. Ammendola, et al. APEnet: LQCD clusters à la APE, in: Proceedings of Lattice 2004 conference, Nucl. Phys. B - Proc. Suppl., vol. 140, pp. 826-828, March 2005. arXiv:hep-lat/0409071v1.
4. M. Bernaschi, M. Fatica, G. Parisi, and L. Parisi Multi-GPU codes for spin systems simulations Comput. Phys. Comm. 183 2012 1416
5. M. Bernaschi, G. Parisi, and L. Parisi Benchmarking GPU and CPU codes for Heisenberg spin glass overrelaxation Comput. Phys. Comm. 182 2011 6
6. NVIDIA CUDA Compute Unified Device Architecture Programming Guide. http://www.nvidia.com/cuda.
7. Source code for the Heisenberg spin glass model used for the tests. http://www.iac.rm.cnr.it/~massimo/hsgfiles.tgz.
8. CUDA-Aware MVAPICH2 distribution. http://mvapich.cse.ohio-state.edu/ overview/mvapich2.
9. CUDA-Aware OpenMPI distribution. http://www.open-mpi.org/faq/?category= buildingbuild-cuda.
10. D.P. Playne, K.A. Hawick, Asynchronous communication for finite-difference simulations on GPU clusters using CUDA and MPI, in: Proc. 2011 Int. Conf on Parallel and Distributed Processing Techniques and Applications, PDPTA'11, Las Vegas, 2011.
11. M. Weigel Performance potential for simulating spin models on GPU J. Comput. Phys. 231 2012 3064
12. J.B. White III, J. Dongarra, Overlapping computation and communication for advection on hybrid parallel computers, in: Proc. of 27th IEEE International Parallel & Distributed Processing Symposium, IPDPS'11, Anchorage, 2011.
13. T. Yavors'kii, and M. Weigel Optimized GPU simulation of continuous-spin glass models Eur. Phys. J. Spec. Top. 210 2012 159