Scalable Implicit Flow Solver for Realistic Wing Simulations with Flow Control

Rasquin, Michel; Smith, Cameron; Chitale, Kedar C.; Seol, E. Seegyoung; Matthews, Benjamin; Martin, Jeffrey L.; Sahni, Onkar; Loy, R. M.; Shephard, Mark S.; Jansen, Kenneth E.

doi:10.1109/mcse.2014.75

Cited by 51 publications

(22 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The largest scale run to date used 256K MPI processes on Argonne National Laboratory's BlueGene/Q Mira machine [RSC*14]. The scaling studies utilized Parallel Hierarchic Adaptive Stabilized Transient Analysis (PHASTA), a highly scalable CFD code, developed by Kenneth Jansen at UC Boulder, for simulating active flow control on complex wing design (see Figure ).…”

Section: In Depth Analysis Of Four In Situ Infrastructuresmentioning

confidence: 99%

In Situ Methods, Infrastructures, and Applications on High Performance Computing Platforms

Bauer

Abbasi

Ahrens

et al. 2016

Computer Graphics Forum

147

View full text Add to dashboard Cite

The considerable interest in the high performance computing (HPC) community regarding analyzing and visualization data without first writing to disk, i.e., in situ processing, is due to several factors. First is an I/O cost savings, where data is analyzed/visualized while being generated, without first storing to a filesystem. Second is the potential for increased accuracy, where fine temporal sampling of transient analysis might expose some complex behavior missed in coarse temporal sampling. Third is the ability to use all available resources, CPU's and accelerators, in the computation of analysis products. This STAR paper brings together researchers, developers and practitioners using in situ methods in extreme-scale HPC with the goal to present existing methods, infrastructures, and a range of computational science and engineering applications using in situ analysis and visualization.

show abstract

Section: In Depth Analysis Of Four In Situ Infrastructuresmentioning

confidence: 99%

In Situ Methods, Infrastructures, and Applications on High Performance Computing Platforms

Bauer

Abbasi

Ahrens

et al. 2016

Computer Graphics Forum

147

View full text Add to dashboard Cite

show abstract

“…Elements may actually not be grouped at all and assembled into an unstructured mesh, some recent references being [27,31,46]. The connectivity of elements can be modeled as a graph, and the partitioning of elements between parallel processes can be translated into partitioning the graph.…”

Section: Introductionmentioning

confidence: 99%

Coarse Mesh Partitioning for Tree-Based AMR

Burstedde¹,

Holke²

2017

SIAM J. Sci. Comput.

View full text Add to dashboard Cite

In tree based adaptive mesh refinement, elements are partitioned between processes using a space filling curve. The curve establishes an ordering between all elements that derive from the same root element, the tree. When representing more complex geometries by patching together several trees, the roots of these trees form an unstructured coarse mesh. We present an algorithm to partition the elements of the coarse mesh such that (a) the fine mesh can be load-balanced to equal element counts per process regardless of the element-to-tree map and (b) each process that holds fine mesh elements has access to the meta data of all relevant trees. As an additional feature, the algorithm partitions the meta data of relevant ghost (halo) trees as well. We develop in detail how each process computes the communication pattern for the partition routine without handshaking and with minimal data movement. We demonstrate the scalability of this approach on up to 917e3 MPI ranks and .37e12 coarse mesh elements, measuring run times of one second or less.

show abstract

“…Common application codes such as FEniCS [33], PLUM [35,36], OpenFOAM [37], or MOAB from the SIGMA toolkit [53] delegate graph partitioning to third-party software like ParMETIS [29] or Scotch [20]. Graph-based algorithms have been advanced to target millions of processes and billions of elements [18,42,49]. Still, increasing the scalability and decreasing the absolute runtime and memory demands of distributed implementations remains a challenge, and the lack of an obvious parent-child structure in many unstructured meshing approaches prevents certain use cases.…”

Section: Introductionmentioning

confidence: 99%

“…[20]. Graph-based algorithms have been advanced to target millions of processes and billions of elements [18,42,49]. Still, increasing the scalability and decreasing the absolute runtime and memory demands of distributed implementations remains a challenge, and the lack of an obvious parent-child structure in many unstructured meshing approaches prevents certain use cases.…”

mentioning

confidence: 99%

A Tetrahedral Space-Filling Curve for Nonconforming Adaptive Meshes

Burstedde¹,

Holke²

2016

SIAM J. Sci. Comput.

View full text Add to dashboard Cite

Abstract. We introduce a space-filling curve for triangular and tetrahedral red-refinement that can be computed using bitwise interleaving operations similar to the well-known Z-order or Morton curve for cubical meshes. To store sufficient information for random access, we define a low-memory encoding using 10 bytes per triangle and 14 bytes per tetrahedron. We present algorithms that compute the parent, children, and face-neighbors of a mesh element in constant time, as well as the next and previous element in the space-filling curve and whether a given element is on the boundary of the root simplex or not. Our presentation concludes with a scalability demonstration that creates and adapts selected meshes on a large distributed-memory system. Key words. Forest of octrees, parallel adaptive mesh refinement, Morton code, high performance computing, nonconforming simplicial mesh, space-filling curve AMS subject classifications. 65M50, 68W10, 65Y05, 65D181. Introduction. Conforming adaptive mesh refinement for simplicial (triangular and tetrahedral) meshes is one of the most successful concepts in numerical mathematics and computational science and engineering; see, e.g., [3,21,50]. Simplices provide high flexibility in meshing to arbitrary domain geometries [46,47] and can be mapped to a reference simplex using an elementwise constant Jacobian in most cases, which allows for an efficient numerical implementation. Discretization and integration methods of various kinds are available for both low and high orders of accuracy.Large-scale scientific computing requires fast and scalable algorithms for (1) adaptive refinement and coarsening (AMR) as well as (2) parallel partitioning. One class of methods for AMR exploits the properties of Delaunay triangulations [23,26,45], while partitioning of unstructured meshes is often approached by formulating the mesh topology as a graph. These triangulations are usually conforming; that is, elements intersect only along whole faces and edges. [20]. Graph-based algorithms have been advanced to target millions of processes and billions of elements [18,42,49]. Still, increasing the scalability and decreasing the absolute runtime and memory demands of distributed implementations remains a challenge, and the lack of an obvious parent-child structure in many unstructured meshing approaches prevents certain use cases.Nonconforming AMR in combination with recursive refinement makes refinement and coarsening nearly trivial operations. The additional mathematical logic to enable hanging faces and edges is well understood for both continuous and discontinuous discretizations [1,22,30,43]. It is local to the loop over the finite elements or volumes and transparent to most of the numerical pipeline, thus offering the possibility to extend existing conforming codes. The resolution may be as coarse as any chosen root mesh (a mesh which is not intended for further coarsening), which poses only a slight limit to geometric flexibility.The challenge of efficient partitioning of meshes may be addre...

show abstract

Scalable Implicit Flow Solver for Realistic Wing Simulations with Flow Control

Cited by 51 publications

References 12 publications

In Situ Methods, Infrastructures, and Applications on High Performance Computing Platforms

In Situ Methods, Infrastructures, and Applications on High Performance Computing Platforms

Coarse Mesh Partitioning for Tree-Based AMR

A Tetrahedral Space-Filling Curve for Nonconforming Adaptive Meshes

Contact Info

Product

Resources

About