Institute for Computational and Mathematical Engineering

Anderson, United States

Institute for Computational and Mathematical Engineering

Anderson, United States
Time filter
Source Type

Abdul-Wahid B.,University of Notre Dame | Abdul-Wahid B.,The Interdisciplinary Center | Yu L.,University of Notre Dame | Rajan D.,University of Notre Dame | And 7 more authors.
2012 IEEE 8th International Conference on E-Science, e-Science 2012 | Year: 2012

Molecular modeling is a field that traditionally has large computational costs. Until recently, most simulation techniques relied on long trajectories, which inherently have poor scalability. A new class of methods is proposed that requires only a large number of short calculations, and for which minimal communication between computer nodes is required. We considered one of the more accurate variants called Accelerated Weighted Ensemble Dynamics (AWE) and for which distributed computing can be made efficient. We implemented AWE using the Work Queue framework for task management and applied it to an all atom protein model (Fip35 WW domain). We can run with excellent scalability by simultaneously utilizing heterogeneous resources from multiple computing platforms such as clouds (Amazon EC2, Microsoft Azure), dedicated clusters, grids, on multiple architectures (CPU/GPU, 32/64bit), and in a dynamic environment in which processes are regularly added or removed from the pool. This has allowed us to achieve an aggregate sampling rate of over 500 ns/hour. As a comparison, a single process typically achieves 0.1 ns/hour. ©2012 IEEE.

Dror R.O.,D E Shaw Research | Dror R.O.,Institute for Computational and Mathematical Engineering | Mildorf T.J.,D E Shaw Research | Hilger D.,Stanford University | And 14 more authors.
Science | Year: 2015

G protein-coupled receptors (GPCRs) relay diverse extracellular signals into cells by catalyzing nucleotide release from heterotrimeric G proteins, but the mechanism underlying this quintessential molecular signaling event has remained unclear. Here we use atomic-level simulations to elucidate the nucleotide - release mechanism. We find that the G protein a subunit Ras and helical domains-previously observed to separate widely upon receptor binding to expose the nucleotide-binding site - separate spontaneously and frequently even in the absence of a receptor. Domain separation is necessary but not sufficient for rapid nucleotide release. Rather, receptors catalyze nucleotide release by favoring an internal structural rearrangement of the Ras domain that weakens its nucleotide affinity. We use double electron-electron resonance spectroscopy and protein engineering to confirm predictions of our computationally determined mechanism. © 2015, American Association for the Advancement of Science. All rights reserved.

Saibaba A.K.,Tufts University | Kitanidis P.K.,Stanford University | Kitanidis P.K.,Institute for Computational and Mathematical Engineering
Advances in Water Resources | Year: 2015

We consider the computational challenges associated with uncertainty quantification involved in parameter estimation such as seismic slowness and hydraulic transmissivity fields. The reconstruction of these parameters can be mathematically described as inverse problems which we tackle using the geostatistical approach. The quantification of uncertainty in the geostatistical approach involves computing the posterior covariance matrix which is prohibitively expensive to fully compute and store. We consider an efficient representation of the posterior covariance matrix at the maximum a posteriori (MAP) point as the sum of the prior covariance matrix and a low-rank update that contains information from the dominant generalized eigenmodes of the data misfit part of the Hessian and the inverse covariance matrix. The rank of the low-rank update is typically independent of the dimension of the unknown parameter. The cost of our method scales as O(mlogm) where m dimension of unknown parameter vector space. Furthermore, we show how to efficiently compute measures of uncertainty that are based on scalar functions of the posterior covariance matrix. The performance of our algorithms is demonstrated by application to model problems in synthetic travel-time tomography and steady-state hydraulic tomography. We explore the accuracy of the posterior covariance on different experimental parameters and show that the cost of approximating the posterior covariance matrix depends on the problem size and is not sensitive to other experimental parameters. © 2015 Elsevier Ltd.

Wang K.,Stanford University | Gretarsson J.T.,Stanford University | Main A.,Stanford University | Farhat C.,Institute for Computational and Mathematical Engineering
20th AIAA Computational Fluid Dynamics Conference 2011 | Year: 2011

A robust, accurate, and computationally efficient interface tracking algorithm is a key component of an embedded/immersed computational framework for the solution of fluid-structure interaction problems with complex and deformable geometries. To a large extent, the design of such an algorithm has focused on the case of a closed embedded interface and a Cartesian Computational Fluid Dynamics (CFD) grid. Here, two robust and efficient interface tracking computational algorithms capable of operating on structured as well as unstructured three-dimensional CFD grids are presented. The first one is based on a projection approach, whereas the second one is based on a collision approach. The first algorithm is faster. However, it is restricted to closed interfaces and resolved enclosed volumes. The second algorithm is therefore slower. However, it can handle open shell surfaces and underresolved enclosed volumes. Both computational algorithms exploit the bounding box hierarchy technique and its parallel distributed implementation to efficiently store and retrieve the elements of the discretized embedded interface. They are illustrated, and their respective performances are assessed and contrasted, with the solution of three-dimensional, nonlinear, dynamic fluid-structure interaction problems pertaining to aeroelastic and underwater implosion applications. © 2011 by Kevin Wang.

Bakhos T.,Institute for Computational and Mathematical Engineering | Cardiff M.,University of Wisconsin - Madison | Barrash W.,Boise State University | Kitanidis P.K.,Institute for Computational and Mathematical Engineering | Kitanidis P.K.,Stanford University
Journal of Hydrology | Year: 2014

Characterizing the subsurface is important for many hydrogeologic projects such as site remediation and groundwater resource exploration. Methods based on the analysis of conventional pumping tests have the notable disadvantage that at a certain distance, the signal is small relative to the noise due to the effects of recharge, pumping in neighboring wells, change in the level or adjacent streams, and other common disturbances. This work focuses on oscillatory pumping tests in which fluid is extracted for half a period, then reinjected. We discuss a major advantage of oscillatory pumping tests: small amplitude signals can be recovered from noisy data measured at observation wells and quantify the uncertainties in the estimates. We demonstrate results from a joint inversion of storativity and transmissivity. We conclude with an analysis of the duration of the initial transient, providing lower bounds on the length of elapsed time until the effects of the transient can be neglected. © 2014 The Authors.

Marsden A.L.,Institute for Computational and Mathematical Engineering | Feinstein J.A.,Stanford University
Current Opinion in Pediatrics | Year: 2015

Purpose of review Recent methodological advances in computational simulations are enabling increasingly realistic simulations of hemodynamics and physiology, driving increased clinical utility. We review recent developments in the use of computational simulations in pediatric and congenital heart disease, describe the clinical impact in modeling in single-ventricle patients, and provide an overview of emerging areas. Recent findings Multiscale modeling combining patient-specific hemodynamics with reduced order (i.e., mathematically and computationally simplified) circulatory models has become the de-facto standard for modeling local hemodynamics and 'global' circulatory physiology. We review recent advances that have enabled faster solutions, discuss new methods (e.g., fluid structure interaction and uncertainty quantification), which lend realism both computationally and clinically to results, highlight novel computationally derived surgical methods for single-ventricle patients, and discuss areas in which modeling has begun to exert its influence including Kawasaki disease, fetal circulation, tetralogy of Fallot (and pulmonary tree), and circulatory support. Summary Computational modeling is emerging as a crucial tool for clinical decision-making and evaluation of novel surgical methods and interventions in pediatric cardiology and beyond. Continued development of modeling methods, with an eye towards clinical needs, will enable clinical adoption in a wide range of pediatric and congenital heart diseases. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Amsallem D.,Stanford University | Cortial J.,Stanford University | Cortial J.,Institute for Computational and Mathematical Engineering | Farhat C.,Stanford University | Farhat C.,Institute for Computational and Mathematical Engineering
AIAA Journal | Year: 2010

This paper describes a computational-fluid-dynamics-based computational methodology for fast on-demand aeroelastic predictions of the behavior of a full aircraft configuration at variable flight conditions and demonstrates its feasibility. The methodology relies onthe offline precomputationof a database of reduced-order bases and models associated with a discrete set of flight parameters, and its training for an interpolation method suitable for reducedorder information. The potential of this near-real-time computational methodology for assisting flutter flight testing is highlighted with the aeroelastic identification of an F-16 configuration in the subsonic, transonic, and supersonic regimes. Copyright © 2010.

Loading Institute for Computational and Mathematical Engineering collaborators
Loading Institute for Computational and Mathematical Engineering collaborators