Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

to be done in this area. Face recognition is a problem that scarcely clamoured for attention before the computer age but, having surfaced, has involved a wide range of techniques and has attracted the attention of some fine minds (David Mumford was a Fields Medallist in 1974). This singular application of mathematical modelling to a messy applied problem of obvious utility and importance but with no unique solution is a pretty one to share with students: perhaps, returning to the source of our opening quotation, we may invert Duncan's earlier observation, 'There is an art to find the mind's construction in the face!'.

Linear and nonlinear waves, by G. B. Whitham. Pp.636. £50. 1999. ISBN 0 471 35942 4 (Wiley).

1: Magnetic Particles: Preparation, Properties and Applications: M. Ozaki. 2: Maghemite (gamma-Fe2O3): A Versatile Magnetic Colloidal Material C.J. Serna, M.P. Morales. 3: Dynamics of Adsorption and Oxidation of Organic Molecules on Illuminated Titanium Dioxide Particles Immersed in Water M.A. Blesa, R.J. Candal, S.A. Bilmes. 4: Colloidal Aggregation in Two-Dimensions A. Moncho-Jorda, F. Martinez-Lopez, M.A. Cabrerizo-Vilchez, R. Hidalgo Alvarez, M. Quesada-PMerez. 5: Kinetics of Particle and Protein Adsorption Z. Adamczyk.

Surface and Colloid Science

Computational Fluid Dynamics: Principles and Applications, Third Edition presents students, engineers, and scientists with all they need to gain a solid understanding of the numerical methods and principles underlying modern computation techniques in fluid dynamics By providing complete coverage of the essential knowledge required in order to write codes or understand commercial codes, the book gives the reader an overview of fundamentals and solution strategies in the early chapters before moving on to cover the details of different solution techniques

	This updated edition includes new worked programming examples, expanded coverage and recent literature regarding incompressible flows, the Discontinuous Galerkin Method, the Lattice Boltzmann Method, higher-order spatial schemes, implicit Runge-Kutta methods and parallelization

	An accompanying companion website contains the sources of 1-D and 2-D Euler and Navier-Stokes flow solvers (structured and unstructured) and grid generators, along with tools for Von Neumann stability analysis of 1-D model equations and examples of various parallelization techniques



	
		Will provide you with the knowledge required to develop and understand modern flow simulation codes
	
		Features new worked programming examples and expanded coverage of incompressible flows, implicit Runge-Kutta methods and code parallelization, among other topics
	
		Includes accompanying companion website that contains the sources of 1-D and 2-D flow solvers as well as grid generators and examples of parallelization techniques

Computational Fluid Dynamics: Principles and Applications Ed. 3

The mechanisms of sound generation in a Mach 0.9, Reynolds number 3600 turbulent jet are investigated by direct numerical simulation. Details of the numerical method are briefly outlined and results are validated against an experiment at the same flow conditions (Stromberg, McLaughlin & Troutt 1980). Lighthill's theory is used to define a nominal acoustic source in the jet, and a numerical solution of Lighthill's equation is compared to the simulation to verify the computational procedures. The acoustic source is Fourier transformed in the axial coordinate and time and then filtered in order to identify and separate components capable of radiating to the far field. This procedure indicates that the peak radiating component of the source is coincident with neither the peak of the full unfiltered source nor that of the turbulent kinetic energy. The phase velocities of significant components range from approximately 5% to 50% of the ambient sound speed which calls into question the commonly made assumption that the noise sources convect at a single velocity. Space–time correlations demonstrate that the sources are not acoustically compact in the streamwise direction and that the portion of the source that radiates at angles greater than 45° is stationary. Filtering non-radiating wavenumber components of the source at single frequencies reveals that a simple modulated wave forms for the source, as might be predicted by linear stability analysis. At small angles from the jet axis the noise from these modes is highly directional, better described by an exponential than a standard Doppler factor.

Noise sources in a low-Reynolds-number turbulent jet at Mach 0.9

This article provides a critical review of computational techniques for flow-noise prediction and the underlying theories. Hybrid approaches, in which the turbulent noise source field is computed and/or modeled separately from the far-field calculation, are afforded particular attention. Numerical methods and modern flow simulation techniques are discussed in terms of their suitability and accuracy for flow-noise calculations. Other topics highlighted include some important formulation and computational issues in the application of aeroacoustic theories, generalized acoustic analogies with better accounts of flow-sound interaction, and recent computational investigations of noise-control strategies. The review ends with an analysis of major challenges and key areas for improvement in order to advance the state of the art of computational aeroacoustics.

/pdf/computational-prediction-of-flow-generated-sound-1xk8onzdl0.pdf

Computational prediction of flow-generated sound

A simple new zonal boundary condition has been proposed. It is based upon the addition of dissipative and convective terms to the compressible Navier Stokes equations

Proposed Inflow/Outflow Boundary Condition for Direct Computation of Aerodynamic Sound

The cellular detail of blood is an essential factor in its flow, especially in vessels or devices with size comparable to that of its suspended cells. This article motivates and reviews numerical simulation techniques that provide a realistic description of cell-scale blood flow by explicitly representing its coupled fluid and solid mechanics. Red blood cells are the principal focus because of their importance and because of their remarkable deformability, which presents particular simulation challenges. Such simulations must couple discretizations of the large-deformation elasticity of the cells with the viscous flow mechanics of the suspension. The Reynolds numbers are low, so the effectively linear fluid mechanics is amenable to a wide range of simulation methods, although the constitutive models and geometric factors of the coupled system introduce challenging nonlinearity. Particular emphasis is given to the relative merits of several fundamentally different simulation methods. The detailed description provided by such simulations is invaluable for advancing our scientific understanding of blood flow, and their ultimate impact will be in the design of biomedical tools and interventions.

/pdf/numerical-simulation-of-flowing-blood-cells-3mer6skahy.pdf

Numerical Simulation of Flowing Blood Cells

A perspective is presented on how machine learning methods might advance fluid mechanics. Current limitations are discussed, though the potential impact is deemed high, so long as outcomes are held to the long-standing critical standards that should guide studies of flow physics.

https://link.aps.org/accepted/10.1103/PhysRevFluids.4.100501

Jonathan B. Freund

Papers

Noise sources in a low-Reynolds-number turbulent jet at Mach 0.9

Computational prediction of flow-generated sound

Proposed Inflow/Outflow Boundary Condition for Direct Computation of Aerodynamic Sound

Numerical Simulation of Flowing Blood Cells

Perspective on machine learning for advancing fluid mechanics