Reducing the memory footprint in Large Eddy Simulations of reactive flows

RAM/core is a critical value for coupled CFD chemistry applications.Two implementation options are proposed to share chemistry LUTs on compute nodes.Two test cases show: memory usage is reduced significantly using any implementation.Runtime is larger using MPI-3 based implementation compared to MMAP system call. CFD simulations of reactive flows couple the domains of flame chemistry and computational fluid dynamics. Solving the chemistry domain in-situ is extremely demanding. It is therefore calculated beforehand and stored into a Look-up table (LUT), which is loaded in each MPI-based CFD application process in a parallel simulation. These chemistry database files can become very large when including many variables and solution values, putting a limit on the number of CFD mesh points as well as the solver instances which can be kept in RAM at the same time. In the paper we approach this problem using dynamic memory managing techniques for single core and parallel applications effectively reducing the memory requirements per core, while keeping the same database resolution. Different implementation options for shared memory on compute nodes provided by the MPI-3 standard and MMAP as a POSIX-compliant system call are analyzed and evaluated using two test cases.

[1]  de Lph Philip Goey,et al.  Modeling of complex premixed burner systems by using flamelet-generated manifolds , 2001 .

[2]  Heinz Pitsch,et al.  Prediction of local extinction and re-ignition effects in non-premixed turbulent combustion using a flamelet/progress variable approach , 2005 .

[3]  F. Nicoud,et al.  Using singular values to build a subgrid-scale model for large eddy simulations , 2011 .

[4]  D. Trimis,et al.  Experimental and numerical study of rich inverse diffusion flame structure , 2013 .

[5]  P. Moin,et al.  Progress-variable approach for large-eddy simulation of non-premixed turbulent combustion , 2004, Journal of Fluid Mechanics.

[6]  T. Poinsot,et al.  A two-step chemical scheme for kerosene–air premixed flames , 2010 .

[7]  Torsten Hoefler,et al.  MPI + MPI: a new hybrid approach to parallel programming with MPI plus shared memory , 2013, Computing.

[8]  Peter J. Denning,et al.  The working set model for program behavior , 1968, CACM.

[9]  R. Barlow,et al.  Piloted methane/air jet flames: Transport effects and aspects of scalar structure , 2005 .

[10]  D. Trimis,et al.  Flamelet-Modeling of Inverse Rich Diffusion Flames , 2013 .

[11]  Michael F. Modest,et al.  Spectral Monte Carlo models for nongray radiation analyses in inhomogeneous participating media , 2007 .

[12]  Nasser Darabiha,et al.  Modelling non-adiabatic partially premixed flames using flame-prolongation of ILDM , 2003 .

[13]  Peter J. Denning.  Virtual Memory , 1996, ACM Comput. Surv..

[14]  Mark Bull,et al.  Development of mixed mode MPI / OpenMP applications , 2001, Sci. Program..

[15]  Luc Vervisch,et al.  Multidimensional flamelet-generated manifolds for partially premixed combustion , 2010 .

[16]  B. Meyer,et al.  An abstraction layer for efficient memory management of tabulated chemistry and flamelet solutions , 2013 .

[17]  D. Trimis,et al.  Development of an inverse diffusion partial oxidation flame and model burner contributing to the development of 3rd generation coal gasifiers , 2013 .

[18]  Message P Forum,et al.  MPI: A Message-Passing Interface Standard , 1994 .

[19]  J. Janicka,et al.  Large-Eddy Simulation of a Turbulent Hydrogen Diffusion Flame , 1999, Proceeding of First Symposium on Turbulence and Shear Flow Phenomena.

[20]  Yehuda Salu,et al.  Turbulent diffusion from a quasi-kinematical point of view , 1977 .

[21]  Ning Li,et al.  Incompact3d: A powerful tool to tackle turbulence problems with up to O(105) computational cores , 2011 .

[22]  H. M. Vinkers,et al.  Ab Initio potential grid based docking: From High Performance Computing to In Silico Screening , 2007 .

[23]  R. Barlow,et al.  Effects of turbulence on species mass fractions in methane/air jet flames , 1998 .

[24]  Eduard Ayguadé,et al.  Employing nested OpenMP for the parallelization of multi-zone computational fluid dynamics applications , 2004, 18th International Parallel and Distributed Processing Symposium, 2004. Proceedings..

[25]  Van Oijen,et al.  Modelling of Premixed Laminar Flames using Flamelet-Generated Manifolds , 2000 .

[26]  Heinz Pitsch,et al.  A general flamelet transformation useful for distinguishing between premixed and non-premixed modes of combustion , 2009 .

[27]  Sally A. McKee,et al.  Hitting the memory wall: implications of the obvious , 1995, CARN.

[28]  M. Eiermann,et al.  A Constrained Control Approach for the Automated Choice of an Optimal Progress Variable for Chemistry Tabulation , 2015 .

[29]  J. Janicka,et al.  Flow field measurements of stable and locally extinguishing hydrocarbon-fuelled jet flames , 2003 .

[30]  Alfred V. Aho,et al.  Principles of Optimal Page Replacement , 1971, J. ACM.

[31]  Nasser Darabiha,et al.  Liminar premixed hydrogen/air counterflow flame simulations using flame prolongation of ILDM with differential diffusion , 2000 .