H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchunk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....we leave the iinspectorj task to the mesh partitioner, and may select the overlapping pattern freely. On the other hand, Koppler [7] and Ujald#n et al. 15] propose a usergiven description of the relationship between the application data. In these works, as well as in the DIME [16] or ARCHIMEDES [1] environments, this knowledge spares the inspector phase. Koppler has the same target: meshbased applications. His approach shares similarities with our tool, with two main dioeerences. One is the target language, which is HPF. The other is how the user species how mesh elements are related. He ....

BAO H., BIELAK J., GHATTAS O., KALLIVOKAS L., O'HALLARON D., SHEWCHUK J., XU J., Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers., Computer Methods in Applied Mechanics and Engineering 152(1-2):85-102, (1998).

....is an earthquake modelling program. It simulates the propagation of elastic seismic waves in large, heterogeneous valleys in order to recover the time history of the ground motion everywhere in the valley due to a specific seismic event. It uses a finite element method on an unstructured mesh [11]. Its C source code is 1500 lines long. WUPWISE (Wuppertal Wilson Fermion Solver) is a program in the field of lattice gauge theory. Lattice gauge theory is a discretization of quantum chromodynamics. Quark propagators are computed within a chromodynamic background field. The inhomogeneous ....

Hesheng Bao, Jacobo Bielak, Omar Ghattas, Loukas F. Kallivokas, David R. O'Hallaron, Jonathan R. Shewchuk, and Jifeng Xu, Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering 152(1-2):85-102, 22 January 1998.

....and v and w are dense vectors. The code is a modified version of the Spark98 kernels [38] which are written for symmetric matrices. The sparse matrix in our experiments is generated from a finite element mesh used to simulate the motion of the ground after an earthquake in the San Fernando valley [3, 4]; it has 30,169 rows and 151,239 non zeroes. In the coarsegrained version, one thread is created for each processor at the beginning of the simulation, and the threads execute a barrier at the end of each iteration. Each processor (thread) is assigned a disjoint and contiguous set of rows of M , ....

Hesheng Bao, Jacobo Bielak, Omar Ghattas, Loukas F. Kallivokas, David R. O'Hallaron, Jonathan R. Shewchuk, and Jifeng Xu. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering, 152(1--2):85--102, January 1998.

....or can view the frequency response characteristics of the room computed from them. In this mode, the acoustic room modeling application shares many characteristics with the design optimization applications (for example, shape optimization [50] currently being discussed in the Quake project [5, 4]. In both acoustic room modeling and design optimization, the user repeatedly adjusts the model parameters (furniture position and composition; soil and rock composition) simulates the physical system, and views the results. In the room modeling application, the user iteratively adjusts the ....

BAO, H., BIELAK, J., GHATTAS, O., KALLIVOKAS, L. F., O'HALLARON, D. R., SHEWCHUK, J. R., AND XU, J. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering 152, 1--2 (Jan. 1998), 85--102.

....matrix and v and w are dense vectors. The code is a modified version of the Spark98 kernels [34] written for symmetric matrices. The sparse matrix in our experiments is generated from a finite element mesh used to simulate the motion of the ground after an earthquake in the San Fernando valley [4, 3]; it has 30,169 rows and 151,239 non zeroes 14 . In the coarse grained version, one thread is created for each processor at the beginning of the simulation, and the threads execute a barrier at the end of each iteration. Each processor (thread) is assigned a disjoint and contiguous set of rows ....

Hesheng Bao, Jacobo Bielak, Omar Ghattas, Loukas F. Kallivokas, David R. O'Hallaron, Jonathan R. Shewchuk, and Jifeng Xu. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering, 152(1-- 2):85--102, January 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchunk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchuk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 152:85--102, Jan. 1998.

No context found.

Bao H., Bielak J., Ghattas O., Kallivokas L., O'Hallaron D., Shewchunk J., Xu J. \Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers." Computer Methods in Applied Mechanics and Engineering, 1998

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D.R. O'Hallaron, J. Shewchunk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchuk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 152:85--102, Jan. 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchuk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 152:85--102, Jan. 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D.R. O'Hallaron, J. Shewchunk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 1998.

No context found.

Bao, H., Bielak, J., Ghattas, O., Kallivokas, L. F., O'Hallaron, D. R., Shewchuk, J. R., and Xu, J. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering 152, 1--2 (Jan. 1998), 85--102.

....the estimated frame rate with the measured frame rate. The input dataset was produced by the 183.equake program from the SPEC CPU2000 benchmark suite [8, 20] The dataset characterizes the simulated motion of the ground in a 50km x 50km x 10km chunk of the San Fernando Basin during an earthquake [3]. The volume of earth is discretized into an unstructured three dimensional mesh with 30K nodes and 151K tetrahedral elements. The dataset contains 101 frames, one frame per output time step. Each frame is 120KB, and consists of scalar values that correspond to the horizontal displacements of the ....

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchuk, and J. Xu. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering, 152:85--102, Jan. 1998.

.... getHostNow( 6: return null; 4 Motivating remote viz service We have the active frame system to implement a general remote visualization service called Dv [1] The specific viz application (Quakeviz) produces an animation of the ground motion during the 1994 Northridge earthquake [2]. Each Quakeviz input dataset consists of a 3D tetrahedral mesh and a sequence of frames, where each frame describes the magnitude of the earth s displacement at each node in the mesh at a particular point in time. Ranging in size from hundreds of GB to TBs, the datasets are generally too massive ....

BAO, H., BIELAK, J., GHATTAS, O., KALLIVOKAS, L., O'HALLARON, D., SHEWCHUK, J., AND XU, J. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering 152 (Jan. 1998), 85--102.

....geology and source parameters are only indirectly observable, and thus introduce uncertainty into the modeling process. Current Capabilities The Quake group at CMU has been working since 1993 on modeling earthquakes in large basins on parallel supercomputers (see www.cs. cmu.edu quake and Bao el al, 1998). Over the past five years we have used such distributed memory machines as the Thinking Machines CM 2, Intel Paragon XPS, TMC CM 5, and Cray T3D and T3E for the most computationally demanding portion of the modeling. The basic computational steps include: 1) generating a mesh that resolves the ....

H. Bao, J. Bielak, O. Ghattas, L.F. Kallivokas, D.R. O'Hallaron, J.R. Shewchuk, and J. Xu (1998), "Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers," Comput. Methods Appl. Mech. Engrg. 152, 85-102.

....simulations. For example, we are part of a group of seismologists and engineers (called the Quake project) that is developing the capability for predicting, by computer simulation, the ground motion of large basins during strong earthquakes in the Greater Los Angeles and Kobe, Japan basins [1]. The Quake ground motion simulations, which must be run remotely on supercomputers because of their complexity, can produce datasets on the order of hundreds of gigabytes to terabytes of floating point numbers that represent displacements of points in the earth during an earthquake. In order to ....

....An earthquake visualization application. 2 Motivating application The initial motivatingapplication for the Dv framework is the visualization of massive datasets that model the motion of the ground during strong earthquakes. The datasets are produced by simulations developed by the Quake project [1]. Quake simulations manipulate large unstructured finite element meshes. For a moderately sized basin such as the San Fernando Valley of Southern California, the meshes consist of 15M nodes and 80M tetrahedral elements. A simulation of 60 seconds of shaking can generate on the order of 6 TB of ....

BAO, H., BIELAK, J., GHATTAS, O., KALLIVOKAS, L., O'HALLARON, D., SHEWCHUK, J., AND XU, J. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering 152 (Jan. 1998), 85--102.

....systems. For example, the Quake project at Carnegie Mellon University uses a sequence of more than 16,000 sparse matrix vector product (SMVP) operations to simulate the motion of the ground during the first 40 seconds of an aftershock of the 1994 Northridge earthquake in the San Fernando Valley [4, 3]. The sparse matrix consists of over 13 million rows and 180 million nonzero entries, where each nonzero entry is a dense 3 Theta 3 submatrix of double precision floating point numbers. Repeated executions of a single SMVP routine account for over 70 of the simulation s computation time. A ....

BAO, H., BIELAK, J., GHATTAS, O., O'HALLARON, D., KALLIVOKAS, L., SHEWCHUK, J., AND XU, J. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering (1997). To appear.

....controls, called application specific quality parameters, that can be adjusted to change the amount of computation and communication resources a task requires. 3 QuakeViz The Quake project developed a toolchain capable of detailed simulation of large geographic areas during strong earthquakes [2]. Thorough assessment of the seismicity associated with a geographic region requires accurate, interactive visualization of the data produced by the simulation. Visualization of this data is complex because the full data for even a small region such as the San Fernando Valley requires ....

BAO, H., BIELAK, J., GHATTAS, O., KALLIVOKAS, L. F., O'HALLARON, D. R., SHEWCHUK, J. R., AND XU, J. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering 152, 1--2 (Jan. 1998), 85--102.

....in many tasks missing their deadlines. These parameters can also be used as knobs to adjust the behavior of the application to meet the user s quality metrics. 3 QuakeViz The Quake project developed a toolchain capable of detailed simulation of large geographic areas during strong earthquakes [2]. For example, a simulation was developed of the response of the San Fernando Valley to an aftershock of the 1994 Northridge Earthquake. The finite element representation of this simulation contains up to 77 million tetrahedrons producing over 40 million unknowns. Storing all of the data for each ....

BAO, H., BIELAK, J., GHATTAS, O., KALLIVOKAS, L. F., O'HALLARON, D. R., SHEWCHUK, J. R., AND XU, J. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering 152, 1--2 (Jan. 1998), 85--102.

....unstructured finite element simulations. The Quake applications, described in Section 2, simulate the motion of the ground during strong earthquakes. They were developed as part of an ongoing project at Carnegie Mellon to model earthquakes in the Los Angeles Basin and other alluvial valleys [2]. The running time of the Quake applications is dominated by a sparse matrix vector product (SMVP) operation that is repeated thousands of times, and the SMVP is the only operation besides I O that requires the transfer of data between processors. Thus, as modelers we can focus on understanding ....

....focus of future efforts to engineer effective communication networks and software. 2 The family of Quake applications The Quake applications are unstructured finite element codes that were developed to predict ground motion in the San FernandoValley of Southern California during earthquakes [2]. There are four Quake applications, denoted sf10, sf5, sf2, and sf1. The sf is an abbreviation for San Fernando, and the number indicates the period (in seconds) of the highest frequency wave that the simulation is able to resolve. For example, sf10 resolves waves with 10 second periods, sf5 ....

[Article contains additional citation context not shown here]

H. Bao, J. Bielak, O. Ghattas, L. Kallivokas, D. O'Hallaron, J. Shewchuk, and J. Xu, Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers, Computer Methods in Applied Mechanics and Engineering 152 (1998), no. 1--2, 85--102.

No context found.

Bao H, Bielak J, Ghattas O, Kallivokas LF, O'Hallaron D.R, Shewchuk JR, Xu J. Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers. Computer Methods in Applied Mechanics and Engineering 1998; 152(1--2):85--102.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. F. Kallivokas, D. R. O'Hallaron, J. R. Shewchuk, and J. Xu. Large-scale Simulation of Elastic Wave Propagation in Heterogeneous Media on Parallel Computers. Computer Methods in Applied Mechanics and Engineering, 152(1--2):85--102, Jan. 1998.

No context found.

H. Bao, J. Bielak, O. Ghattas, L. F. Kallivokas, D. R. O'Hallaron, J. R. Shewchuk, and J. Xu, "Large-scale simulation of elastic wave propagation in heterogeneous media on parallel computers," Computer Methods in Applied Mechanics and Engineering, vol. 152, no. 1-2, pp.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC