CMT-Bone — A Proxy Application for Compressible Multiphase Turbulent Flows

Banerjee, Tania; Hackl, Jason; Shringarpure, Mrugesh; Islam, Tanzima; Balachandar, S.; Jackson, T. L.; Ranka, Sanjay

doi:10.1109/hipc.2016.029

Cited by 7 publications

(9 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is also important to consider imbalances generated in the communication load. CMT‐nek follows a discontinuous Galerkin scheme and hence is light on communication. Communication is required for transferring shared faces of elements residing on different processors.…”

Section: Load Balancing Cmt‐nekmentioning

confidence: 99%

“…The two terms of the right-hand-side of Equation (2) must be evaluated on each element and added together at the surface nodes to provide the right-hand-side to a numerical scheme for integrating U∕ t. Thus, the computations in CMT-nek may be described in terms of three main stages: preliminary computations, evaluating surface integrals, and evaluating volume integrals. Further details about the formulations and the implementation may be found in the work of Banerjee et al 1 We will now describe the support for particles in CMT-nek.…”

Section: Cmt-nekmentioning

confidence: 99%

“…Accurate prediction and simulation of flows with compressible multiphase turbulence (CMT) are challenging because of the complexity of the actual physical processes involved. CMT‐nek is a new scientific application that solves the compressible Navier‐Stokes equations for multiphase flows. Performing high fidelity, predictive simulations of particle‐laden, explosively dispersed turbulent flows under conditions of extreme pressure, and the temperature is the primary goal of the CMT‐nek application.…”

Section: Introductionmentioning

confidence: 99%

“…The three‐dimensional Euler equations modeled using the conservation law are given by Equation

\frac{\partial U_{m}}{\partial t} + \nabla \cdot {boldH}_{m} = R_{m},

where U m is the m th component of U , a vector of the five conserved variables, namely mass, energy, and the three components of momentum. As shown in the work of Banerjee et al, the partial differential equation, given in Equation is solved using discontinuous Galerkin methods by partitioning the domain into non‐overlapping elements. Each element is mapped to a unit cube also known as the reference element, using the Jacobian matrix.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, the computations in CMT‐nek may be described in terms of three main stages: preliminary computations, evaluating surface integrals, and evaluating volume integrals. Further details about the formulations and the implementation may be found in the work of Banerjee et al We will now describe the support for particles in CMT‐nek.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Dynamic load balancing for a mesh‐based scientific application

Zhai

Banerjee

Zwick

et al. 2020

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

Summary CMT‐nek is a new scientific application for performing high fidelity predictive simulations of particle‐laden, explosively dispersed turbulent flows. CMT‐nek is compute‐intensive and targeted for deployment on exascale platforms. The moving particles are the primary source of load imbalance when the application is executed on parallel processors. In a demonstration problem, all the particles are initially in a closed container until a detonation occurs and the particles move apart. If all processors get an equal share of the fluid domain, then only some of the processors get sections of the domain that are initially laden with particles, leading to disparate loads on the processors. To eliminate load imbalance in different processors and to speed up the makespan, we present different load‐balancing algorithms for CMT‐nek on large‐scale multicore platforms. The load on a processor is determined using different techniques. The performance of the different load‐balancing algorithms is compared, and the associated overheads are analyzed. Evaluations of the application with and without load‐balancing are conducted, and these show that with load‐balancing, simulation time becomes faster by a factor of up to 9.97. The performance was further improved by a factor of up to 1.42 using machine‐learning–based algorithms.

show abstract

Section: Load Balancing Cmt‐nekmentioning

confidence: 99%

Section: Cmt-nekmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…The three‐dimensional Euler equations modeled using the conservation law are given by Equation