Abstract
MPI application performance can vary based on the scheduler’s placing of ranks, whether between nodes or on cores in the same multi-core chip. MPI applications, by default, are at the mercy of the application placement software decision that assigns nodes to a job. We describe herein the general approach of node ordering for allocation in a 3D torus, how it improved MPI application performance, even in the face of an anisotropic interconnect. We demonstrate, quantitatively, that our topologically-based ordering results in improved performance for several MPI applications running on a Top10 supercomputer.
This material is based upon work supported by the Defense Advanced Research Projects Agency under its Agreement No. HR0011-07-9-0001. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Defense Advanced Research Projects Agency.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Hilbert curve – from wolfram MathWorld (March 2010), http://mathworld.wolfram.com/HilbertCurve.html
NERSC6 benchmarks (March 2011), http://www.nersc.gov/projects/SDSA/software/?benchmark=NERSC6
Agarwal, T., Sharma, A., Kal, L.V.: Topology-aware task mapping for reducing communication contention on large parallel machines. In: Proceedings of IEEE International Parallel and Distributed Processing Symposium, Rhodes Island, Greece, p. 110 (2006)
Albing, C., Baker, M.: ALPS, topology, and performance: A comparison of linear orderings for application placement in a 3D torus. Cray User Group, Edinburgh, Scotland, UK (May 2010)
Bani-Mohammad, S., Ould-Khaoua, M., Ababneh, I.: An efficient non-contiguous processor allocation strategy for 2D mesh connected multicomputers. Information Sciences 177(14), 2867–2883 (2007)
Bays, C.: A comparison of next-fit, first-fit, and best-fit. Communications of the ACM 20(3), 191–192 (1977)
Bhatele, A., Kale, L.V.: Application-specific topology-aware mapping for three dimensional topologies. In: 2008 IEEE International Symposium on Parallel and Distributed Processing, pp. 1–8. IEEE, Miami (2008)
Bhatele, A., Kal, L.V.: An evaluative study on the effect of contention on message latencies in large supercomputers. In: 2009 IEEE International Symposium on Parallel & Distributed Processing, Rome, Italy, pp. 1–8 (May 2009)
Krevat, E., Castaos, J., Moreira, J.: Job scheduling for the BlueGene/L system. LNCS, pp. 38–54. Springer, Edinburgh (2002)
Leung, V.J., Arkin, E.M., Bender, M.A., Bunde, D., Johnston, J., Lal, A., Mitchell, J.S., Phillips, C., Seiden, S.S.: Processor allocation on Cplant: achieving general processor locality using one-dimensional allocation strategies. In: Proc. 4th IEEE International Conference on Cluster Computing, pp. 296–304 (2002)
Lo, V., Windisch, K., Liu, W., Nitzberg, B.: Noncontiguous processor allocation algorithms for mesh-connected multicomputers. IEEE Transactions on Parallel and Distributed Systems 8(7), 712–726 (1997)
Russell, J.J.: A simulation of first and best fit allocation algorithms in a modern simulation environment. In: Proc. of 6th Annual CCEC Symposium (2008)
Weisser, D., Nystrom, N., Brown, S., Gardner, J., O’Neal, D., Urbanic, J., Lim, J., Reddy, R., Raymond, R., Wang, Y., Welling, J.: Optimizing job placement on the Cray XT3. Lugano, Switzerland (May 2006)
Yu, H., Chung, I., Moreira, J.: Topology mapping for blue Gene/L supercomputer. In: Proceedings of the 2006 ACM/IEEE Conference on Supercomputing, SC 2006, Tampa, Florida, p. 116 (2006)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Albing, C. et al. (2011). Scalable Node Allocation for Improved Performance in Regular and Anisotropic 3D Torus Supercomputers. In: Cotronis, Y., Danalis, A., Nikolopoulos, D.S., Dongarra, J. (eds) Recent Advances in the Message Passing Interface. EuroMPI 2011. Lecture Notes in Computer Science, vol 6960. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-24449-0_9
Download citation
DOI: https://doi.org/10.1007/978-3-642-24449-0_9
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-24448-3
Online ISBN: 978-3-642-24449-0
eBook Packages: Computer ScienceComputer Science (R0)