Papers by Phillip Colella
Astrophysical Journal Supplement Series, Dec 22, 2011
We present a description of the adaptive mesh refinement (AMR) implementation of the PLUTO code f... more We present a description of the adaptive mesh refinement (AMR) implementation of the PLUTO code for solving the equations of classical and special relativistic magnetohydrodynamics (MHD and RMHD). The current release exploits, in addition to the static grid version of the code, the distributed infrastructure of the CHOMBO library for multidimensional parallel computations over block-structured, adaptively refined grids. We employ a conservative finite-volume approach where primary flow quantities are discretized at the cell-center in a dimensionally unsplit fashion using the Corner Transport Upwind (CTU) method. Time stepping relies on a characteristic tracing step where piecewise parabolic method (PPM), weighted essentially non-oscillatory (WENO) or slope-limited linear interpolation schemes can be handily adopted. A characteristic decomposition-free version of the scheme is also illustrated. The solenoidal condition of the magnetic field is enforced by augmenting the equations with a generalized Lagrange multiplier (GLM) providing propagation and damping of divergence errors through a mixed hyperbolic/parabolic explicit cleaning step. Among the novel features, we describe an extension of the scheme to include non-ideal dissipative processes such as viscosity, resistivity and anisotropic thermal conduction without operator splitting. Finally, we illustrate an efficient treatment of point-local, potentially stiff source terms over hierarchical nested grids by taking advantage of the adaptivity in time. Several multidimensional benchmarks and applications to problems of astrophysical relevance assess the potentiality of the AMR version of PLUTO in resolving flow features separated by large spatial and temporal disparities.
arXiv (Cornell University), Oct 4, 2011
We present a description of the adaptive mesh refinement (AMR) implementation of the PLUTO code f... more We present a description of the adaptive mesh refinement (AMR) implementation of the PLUTO code for solving the equations of classical and special relativistic magnetohydrodynamics (MHD and RMHD). The current release exploits, in addition to the static grid version of the code, the distributed infrastructure of the CHOMBO library for multidimensional parallel computations over block-structured, adaptively refined grids. We employ a conservative finite-volume approach where primary flow quantities are discretized at the cell-center in a dimensionally unsplit fashion using the Corner Transport Upwind (CTU) method. Time stepping relies on a characteristic tracing step where piecewise parabolic method (PPM), weighted essentially non-oscillatory (WENO) or slope-limited linear interpolation schemes can be handily adopted. A characteristic decomposition-free version of the scheme is also illustrated. The solenoidal condition of the magnetic field is enforced by augmenting the equations with a generalized Lagrange multiplier (GLM) providing propagation and damping of divergence errors through a mixed hyperbolic/parabolic explicit cleaning step. Among the novel features, we describe an extension of the scheme to include non-ideal dissipative processes such as viscosity, resistivity and anisotropic thermal conduction without operator splitting. Finally, we illustrate an efficient treatment of point-local, potentially stiff source terms over hierarchical nested grids by taking advantage of the adaptivity in time. Several multidimensional benchmarks and applications to problems of astrophysical relevance assess the potentiality of the AMR version of PLUTO in resolving flow features separated by large spatial and temporal disparities.
Communications in applied mathematics and computational science, Sep 4, 2015
We present a high-order finite-volume approach for solving the shallow-water equations on the sph... more We present a high-order finite-volume approach for solving the shallow-water equations on the sphere, using multiblock grids on the cubed sphere. This approach combines a Runge-Kutta time discretization with a fourth-order-accurate spatial discretization and includes adaptive mesh refinement and refinement in time. Results of tests show fourth-order convergence for the shallow-water equations as well as for advection in a highly deformational flow. Hierarchical adaptive mesh refinement allows solution error to be achieved that is comparable to that obtained with uniform resolution of the most refined level of the hierarchy but with many fewer operations.
Discrete and Continuous Dynamical Systems, Mar 1, 2016
Computers & Fluids, Dec 1, 2015
A fourth-order accurate finite-volume method is presented for solving timedependent hyperbolic sy... more A fourth-order accurate finite-volume method is presented for solving timedependent hyperbolic systems of conservation laws on mapped grids that are adaptively refined in space and time. Novel considerations for formulating the semi-discrete system of equations in computational space are combined with detailed mechanisms for accommodating the adapting grids. These considerations ensure that conservation is maintained and that the divergence of a constant vector field is always zero (freestream-preservation property). The solution in time is advanced with a fourth-order Runge-Kutta method. A series of tests verifies that the expected accuracy is achieved in smooth flows and the solution of a Mach reflection problem demonstrates the effectiveness of the algorithm in resolving strong discontinuities.
Journal of Parallel and Distributed Computing, Dec 1, 2014
Journal of Computational Physics, 1999
In Sussman, Smereka and Osher (1994), a numerical method using the level set approach was formula... more In Sussman, Smereka and Osher (1994), a numerical method using the level set approach was formulated for solving incompressible two-phase flow with surface tension. In the level set approach, the interface is represented as the zero level set of a smooth function; this has the effect of replacing the advection of density, which has steep gradients at the interface, with the advection of the level set function, which is smooth. In addition, the interface can merge or break up with no special treatment. We maintain the level set function as the signed distance from the interface in order to robustly compute flows with high density ratios and stiff surface tension effects. In this work, we couple the level set scheme to an adaptive projection method for the incompressible Navier-Stokes equations, in order to achieve higher resolution of the interface with a minimum of additional expense. We present two-dimensional axisymmetric and fully three-dimensional results of air bubble and water drop computations.
Reactive melt infiltration is viewed as a promising means of achieving near-net shape manufacturi... more Reactive melt infiltration is viewed as a promising means of achieving near-net shape manufacturing with quick processing time and at low cost. Since the reactants and products are, in general, of varying density, overall conservation of mass dictates that there is a force related to chemical conversion which can directly influence infiltration behavior. In effect, the driving pressure forces may compete with the forces from chemical conversion, affecting the advancement of the front. We have developed a twodimensional numerical code to examine these effects, using reaction-formed silicon carbide as a model system for this process. We have examined a range of initial porosities, pore radii, and reaction rates in order to investigate their effects on infiltration dynamics.
50th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition, Jan 9, 2012
A fourth-order accurate finite-volume method is presented for solving time-dependent hyperbolic s... more A fourth-order accurate finite-volume method is presented for solving time-dependent hyperbolic systems of conservation laws on mapped grids that are adaptively refined in space and time. Novel considerations for formulating the semi-discrete system of equations in computational space combined with detailed mechanisms for accommodating the adapting grids ensure that conservation is maintained and that the divergence of a constant vector field is always zero (freestream-preservation property). Advancement in time is achieved with a fourth-order Runge-Kutta method.
Bulletin of the American Physical Society, Nov 13, 2013
Laboratory-Progress on the development of the continuum gyrokinetic code COGENT for edge plasma s... more Laboratory-Progress on the development of the continuum gyrokinetic code COGENT for edge plasma simulations is reported. The COGENT code models an axisymmetric gyrokinetic equation coupled to the long-wavelength limit of the gyro-Poisson equation. COGENT is distinguished by application of fourthorder conservative discretization, and mapped multiblock grid technology to handle the geometric complexity of the tokamak edge. The code has also a number of model collision operator options, which have been successfully verified in neoclassical simulations. Our recent development work has focused on incorporation of the full (nonlinear) Fokker-Planck collision model. The implementation of the Fokker-Plank operator is discussed in detail, and the results of the initial verification studies are presented. In addition, we report on progress and status of the newly available divertor version of the COGENT code that includes both closed and open magnetic field line regions and a model for recycled neutral gas.
OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information), Oct 18, 1995
This is a preprint of a paper intended for publication in a journal or proceedings. Since changes... more This is a preprint of a paper intended for publication in a journal or proceedings. Since changes may be made before publication, this preprint is made available with the understanding that it will not be cited orreproducedwithout the permission of the author.
The Energy Citations Database (ECD) provides access to historical and current research (1948 to t... more The Energy Citations Database (ECD) provides access to historical and current research (1948 to the present) from the Department of Energy (DOE) and predecessor agencies.
Bulletin of the American Physical Society, Nov 20, 2017
embedded boundaries. The embedded boundary is allowed to freely cut the locally-refined backgroun... more embedded boundaries. The embedded boundary is allowed to freely cut the locally-refined background Cartesian grid. Implicit-function representation is used for the embedded boundary, which allows us to convert the required geometric moments in the Taylor series expansion (upto arbitrary order) of polynomials into an algebraic problem in lower dimensions. The computed geometric moments are then used to construct stencils for various operators like the Laplacian, divergence, gradient, etc., by solving a least-squares system locally. We also construct the inter-level data-transfer operators like prolongation and restriction for multi grid solvers using the same least-squares system approach. This allows us to retain high-order of accuracy near coarse-fine interface and near embedded boundaries. Canonical problems like Taylor-Green vortex flow and flow past bluff bodies will be presented to demonstrate the proposed method.
AGU Fall Meeting Abstracts, Dec 15, 2014
Nucleation and Atmospheric Aerosols, 1990
We present an example of our computations on a shock wave refracting from CO2 into CH4. The resul... more We present an example of our computations on a shock wave refracting from CO2 into CH4. The results are compared with the experiments of Abdel‐Fattah and Henderson. Our input data was the same as the parameters they measured and we used the same boundaries. We also took into account the mutal gas contamination present in the experiments and the effects of a separating membrane. All the principle properties of a free‐precursor refraction were detected.
This document was prepared as an account of work sponsored by PD agency of the Unitedstates Gover... more This document was prepared as an account of work sponsored by PD agency of the Unitedstates Government NeithertheUnitedStates GovernmentnortheUniversity of Cllifornianor any ofthar employees, makes any wurplty, express or implied, or lamnesanyIegal~ilityorwponribiIitpforthePccnracy,compietenes ofanyinformation, apparatwi,pmduct,orprocecs disdoscd,orrrprrstntsthatitsnse wonldnotinfringeprip~~yo~~rights. Referenceherciuto payspedficcommercial p d u d s , process, or service by trade name, trademark, maaufactnrer, or otherwise, doesnotnecessarily constitute orimply itsendorsement, recommendation, or favoring by the United States Government or the Univusity of California. The views and opinions of authors expressed herein do not necessarily state or d e c t those of the United States Government or the Ziniversity of California, and shall not be used for advertising or product endorsement purposes.
Proceedings of SPIE, Jun 7, 1996
DISCLAIMER 'thisd~tw~~A*ana~t& wti~~magq& *Utiti*t~Gv~t. Ndther the United Statea Government nor ... more DISCLAIMER 'thisd~tw~~A*ana~t& wti~~magq& *Utiti*t~Gv~t. Ndther the United Statea Government nor the UNVeraityof California nor qny of their employees,makes any warranty, expressor implied, or assumesany legalliabilityor responsibilityfor the accurq, co@etemsa, or WeMnesaofany inforrnatim apparatus, pmdu@ or~dkloae4 or~ts that its use would not Wrings privatelyown dghts. Referenceherein to qy spaific commerd products, process,or serviceby trade nanw trad~manufacturer,or otherwke, doea not necesaady mnatitute or imply its endmemenL mmmmmdatiol&orfavodIig bytheunitedstatea Covernrnent or the Universityof " CaMomia. The views and opinionsof authors eqmased herein do not mmaarily state or reflect those of the United States COVemM ent or the Univemityof Califoti and shall not be used for q dvedaing or productendomemen tpqmaea. Ihfsreporthaabeenreproduced directly hum the beat qvailablecopy.
AIAA SCITECH 2022 Forum, 2022
Author(s): Friedman, Alex; Davidson, Ronald C.; Welch, Dale R.; Cohen, Ronald H.; Grote, David P.... more Author(s): Friedman, Alex; Davidson, Ronald C.; Welch, Dale R.; Cohen, Ronald H.; Grote, David P.; Sharp, William M.; Colella, Phillip; Vay, Jean-Luc; Yu, Simon S.; Lee, W. Wei-li; Qin, Hong; Rose, David V.; O'Shea, Patrick G.; Kishek, Rami A.; Haber, Irving
2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW), 2018
We propose FFTX, a new framework for building high-performance FFT-based applications on exascale... more We propose FFTX, a new framework for building high-performance FFT-based applications on exascale machines. Complex node architectures lead to multiple levels of parallelism and demand efficient ways of data communication. The current FFTW interface falls short in maximizing performance in such scenarios. FFTX is designed to enable application developers to leverage expert-level, automatic optimizations while navigating a familiar interface. FFTX is backwards compatible to FFTW and extends the FFTW Interface into an embedded Domain Specific Language (DSL) expressed as a library interface. By means of a SPIRAL-based back end, this enables build-time source-to-source translation and advanced performance optimizations, such as cross-library calls optimizations, targeting of accelerators through offload-ing, and inlining of user-provided kernels. We demonstrate the use of FFTX with the prototypical example of 1D and 3D pruned convolutions and discuss future extensions.
Uploads
Papers by Phillip Colella