nach oben

Erschienen in:

2019 | OriginalPaper | Buchkapitel

Automatic Exploration of Reduced Floating-Point Representations in Iterative Methods

verfasst von : Yohan Chatelain, Eric Petit, Pablo de Oliveira Castro, Ghislain Lartigue, David Defour

Erschienen in: Euro-Par 2019: Parallel Processing

Verlag: Springer International Publishing

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config

KI-gestützte Suche

Aus

Abstract

With the ever-increasing need for computation of scientific applications, new application domains, and major energy constraints, the landscape of floating-point computation is changing. New floating-point representation formats are emerging and there is a need for tools to simulate their impact in legacy codes. In this paper, we propose an automatic tool to evaluate the effect of adapting the floating point precision for each operation over time, which is particularly useful in iterative schemes. We present a backend to emulate any IEEE-754 floating-point operation in lower precision. We tested the numerical errors resilience of our solutions thanks to Monte Carlo Arithmetic and demonstrated the effectiveness of this methodology on YALES2, a large Combustion-CFD HPC code, by achieving 28% to 67% reduction in communication volume by lowering precision.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 390 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Maschinenbau + Werkstoffe

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Vorheriges Kapitel Cholesky and Gram-Schmidt Orthogonalization for Tall-and-Skinny QR Factorizations on Graphics Processors

Nächstes Kapitel Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators

Anzt, H., Dongarra, J., et al.: Adaptive precision in block-Jacobi preconditioning for iterative sparse linear system solvers. Concurr. Comput. Pract. Exp. 31, e4460 (2017)CrossRef

Benard, P., Lartigue, G., et al.: Large-eddy simulation of the lean-premixed PRECCINSTA burner with wall heat loss. Proc. Combust. Inst. 37, 5233–5243 (2018)CrossRef

Benard, P., Viré, A., et al.: Large-eddy simulation of wind turbines wakes including geometrical effects. Comput. Fluids 173, 133–139 (2018). https://doi.org/10.1016/j.compfluid.2018.03.015. http://www.sciencedirect.com/science/article/pii/S0045793018301154MathSciNetCrossRefMATH

Benard, P., Balarac, G., et al.: Mesh adaptation for large-eddy simulations in complex geometries. Int. J. Numer. Methods Fluids 81(12), 719–740 (2016). https://doi.org/10.1002/fld.4204. https://onlinelibrary.wiley.com/doi/abs/10.1002/fld.4204MathSciNetCrossRef

Boldo, S., Melquiond, G.: When double rounding is odd. In: 17th IMACS World Congress, Paris, France, p. 11 (2005)

Boulet, L., Bénard, P., et al.: Modeling of conjugate heat transfer in a kerosene/air spray flame used for aeronautical fire resistance tests. Flow Turbul. Combust. 101(2), 579–602 (2018). https://doi.org/10.1007/s10494-018-9965-8CrossRef

Chatelain, Y., de Oliveira Castro, P., et al.: VeriTracer: context-enriched tracer for floating-point arithmetic analysis. In: 25th IEEE Symposium on Computer Arithmetic (ARITH), pp. 61–68 (2018)

Chorin, A.J.: Numerical solution of the Navier-Stokes equations. Math. Comput. 22(104), 745–762 (1968)MathSciNetCrossRef

Cools, S., Yetkin, E.F., et al.: Analysis of rounding error accumulation in Conjugate Gradients to improve the maximal attainable accuracy of pipelined CG. Research Report RR-8849, Inria Bordeaux Sud-Ouest, January 2016. https://hal.inria.fr/hal-01262716

10.

Darulova, E., Horn, E., Sharma, S.: Sound mixed-precision optimization with rewriting. In: Proceedings of the 9th ACM/IEEE International Conference on Cyber-Physical Systems, pp. 208–219. IEEE Press (2018)

11.

Das, D., Mellempudi, N., et al.: Mixed precision training of convolutional neural networks using integer operations. CoRR abs/1802.00930 (2018). http://arxiv.org/abs/1802.00930

12.

Defour, D.: FP-ANR: a representation format to handle floating-point cancellation at run-time. In: 25th IEEE Symposium on Computer Arithmetic (ARITH), pp. 76–83 (2018)

13.

Dendy, J.: Black box multigrid. J. Comput. Phys. 48(3), 366–386 (1982). https://doi.org/10.1016/0021-9991(82)90057-2. http://www.sciencedirect.com/science/article/pii/0021999182900572MathSciNetCrossRefMATH

14.

Denis, C., de Oliveira Castro, P., Petit, E.: Verificarlo: checking floating point accuracy through Monte Carlo arithmetic. In: 23nd IEEE Symposium on Computer Arithmetic (ARITH), pp. 55–62 (2016)

15.

de Dinechin, F., Pasca, B.: Designing custom arithmetic data paths with FloPoCo. IEEE Des. Test Comput. 28, 18–27 (2011)CrossRef

16.

Graillat, S., Jézéquel, F., et al.: PROMISE: floating-point precision tuning with stochastic arithmetic. In: Proceedings of the 17th International Symposium on Scientific Computing, Computer Arithmetics and Verified Numerics (SCAN), pp. 98–99 (2016)

17.

Guedot, L., Lartigue, G., Moureau, V.: Design of implicit high-order filters on unstructured grids for the identification of large-scale features in large-eddy simulation and application to a swirl burner. Phys. Fluids 27(4), 045107 (2015). https://doi.org/10.1063/1.4917280CrossRef

18.

Gustafson, Y.: Beating floating point at its own game: posit arithmetic. Supercomput. Front. Innov. Int. J. 4(2), 71–86 (2017)

19.

Haidar, A., Tomov, S., et al.: Harnessing GPU tensor cores for fast FP16 arithmetic to speed up mixed-precision iterative refinement solvers. In: Proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, SC 2018, Piscataway, NJ, USA, pp. 47:1–47:11. IEEE Press (2018)

20.

Higham, N.J.: Accuracy and Stability of Numerical Algorithms. SIAM, Philadelphia (2002)CrossRef

21.

Ho, N.M., Manogaran, E., et al.: Efficient floating point precision tuning for approximate computing. In: 22nd Asia and South Pacific Design Automation Conference (ASP-DAC), pp. 63–68. IEEE (2017)

22.

Intel Corp.: Intel VML (2018). https://software.intel.com/en-us/mkl-developer-reference-c-vector-mathematical-functions

23.

Lam, M.O., Hollingsworth, J.K., et al.: Automatically adapting programs for mixed-precision floating-point computation. In: Proceedings of the 27th International conference on supercomputing, pp. 369–378. ACM (2013)

24.

Lartigue, G., Meier, U., Bérat, C.: Experimental and numerical investigation of self-excited combustion oscillations in a scaled gas turbine combustor. Appl. Therm. Eng. 24(11–12), 1583–1592 (2004)CrossRef

25.

Legrand, N., Lartigue, G., Moureau, V.: A multi-grid framework for the extraction of large-scale vortices in large-eddy simulation. J. Comput. Phys. 349, 528–560 (2017). https://doi.org/10.1016/j.jcp.2017.08.030. http://www.sciencedirect.com/science/article/pii/S0021999117306010MathSciNetCrossRefMATH

26.

Malandain, M.: Massively parallel simulation of low-Mach number turbulent flows. Theses, INSA de Rouen, January 2013. https://tel.archives-ouvertes.fr/tel-00801502

27.

Malandain, M., Maheu, N., Moureau, V.: Optimization of the deflated conjugate gradient algorithm for the solving of elliptic equations on massively parallel machines. J. Comput. Phys. 238, 32–47 (2013). https://doi.org/10.1016/j.jcp.2012.11.046. http://www.sciencedirect.com/science/article/pii/S0021999112007280MathSciNetCrossRef

28.

Moureau, V., Domingo, P., Vervisch, L.: Design of a massively parallel CFD code for complex geometries. Comptes Rendus Mécanique 339, 141–148 (2011)CrossRef

29.

Nicolaides, R.A.: Deflation of conjugate gradients with applications to boundary value problems. SIAM J. Numer. Anal. 24(2), 355–365 (1987)MathSciNetCrossRef

30.

Panchekha, P., Sanchez-Stern, A., et al.: Automatically improving accuracy for floating point expressions. In: Proceedings of the 36th ACM SIGPLAN Conference on Programming Language Design and Implementation, pp. 1–11. ACM (2015)

31.

Parker, S.: Monte carlo arithmetic: exploiting randomness in floating-point arithmetic. Technical report CSD-970002, UCLA Computer Science Department (1997)

32.

Pierce, C.D., Moin, P.: Progress-variable approach for large-eddy simulation of non-premixed turbulent combustion. J. Fluid Mech. 504, 73–97 (2004). https://doi.org/10.1017/S0022112004008213MathSciNetCrossRefMATH

33.

Rubio-González, C., Nguyen, C., et al.: Floating-point precision tuning using blame analysis. In: Proceedings of the 38th International Conference on Software Engineering, pp. 1074–1085. ACM (2016)

34.

Rubio-González, C., Nguyen, C., et al.: Precimonious: tuning assistant for floating-point precision. In: International Conference for High Performance Computing, Networking, Storage and Analysis (SC), pp. 1–12. IEEE (2013)

35.

Schkufza, E., Sharma, R., Aiken, A.: Stochastic optimization of floating-point programs with tunable precision. ACM SIGPLAN Not. 49(6), 53–64 (2014)CrossRef

36.

Serebryany, K., Bruening, D., et al.: AddressSanitizer: a fast address sanity checker. In: USENIX ATC 2012 (2012)

37.

Sohier, D., De Oliveira Castro, P., et al.: Confidence Intervals for Stochastic Arithmetic (2018). https://hal.archives-ouvertes.fr/hal-01827319, preprint

38.

Tagliavini, G., Mach, S., et al.: A transprecision floating-point platform for ultra-low power computing. In: Design, Automation & Test in Europe Conference & Exhibition (DATE), pp. 1051–1056. IEEE (2018)

Titel: Automatic Exploration of Reduced Floating-Point Representations in Iterative Methods
verfasst von: Yohan Chatelain
Eric Petit
Pablo de Oliveira Castro
Ghislain Lartigue
David Defour
Verlag: Springer International Publishing
Buch: Euro-Par 2019: Parallel Processing
Print ISBN: 978-3-030-29399-4

Electronic ISBN: 978-3-030-29400-7

Copyright-Jahr: 2019
DOI: https://doi.org/10.1007/978-3-030-29400-7_34

Springer Professional

Abstract

Bitte loggen Sie sich ein, um Zugang zu Ihrer Lizenz zu erhalten.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Springer Professional "Technik"

Springer Professional "Wirtschaft"