Application performance is determined by a combination of many choices: hardware platform, runtime environment, languages and compilers used, algorithm choice and implementation, and more. In this complicated environment, we find that the use of mini-applications - small self-contained proxies for real applications - is an excellent approach for rapidly exploring the parameter space of all these choices. Furthermore, use of mini-applications enriches the interaction between application, library and computer system developers by providing explicit functioning software and concrete performance results that lead to detailed, focused discussions of design trade-offs, algorithm choices and runtime performance issues. In this paper we discuss a collection of mini-applications and demonstrate how we use them to analyze and improve application performance on new and future computer platforms.

/pdf/improving-performance-via-mini-applications-141kz27nto.pdf

Improving Performance via Mini-applications

Parallel strategies for crash and impact simulations

A computer system for analyzing and automatically modifying existing process system models that includes (i) an Initial Data stage (410, 420, 430, 450) that prompts the user to enter qualitative and quantitative objectives in designing or re-designing a process system; (ii) a Component model stage that displays, edits, and/or completes an image of the sequence of components produced or added during the process; (iii) a Process (Functional) Analysis stage (225) designating each function as harmful or useful, ranking each useful function, and performing a value link analysis of each function; (iv) a trimming routine (245) for eliminating or simplifying an operation in a process, and (v) a trimming routine that generates a recommended list, in priority, of problems to be solved to achieve the function trimming.

Computer based system for imaging and analyzing a process system and indicating values of specific design changes

Natural convection between concentric spheres

Dynamic load-balancing of finite element applications with the DRAMA library

We describe our parallelization of PRONTO, Sandia's transient solid dynamics code, via a novel algorithmic approach that utilizes multiple decompositions for different key segments of the computations, including the material contact calculation This latter calculation is notoriously difficult to perform well in parallel, because it involves dynamically changing geometry, global searches for elements in contact, and unstructured communications among the compute nodes Our approach scales to at least 3600 compute nodes on problems involving millions of finite elements We can simulate models using more than ten million elements in a few tenths of a second per timestep

/pdf/transient-solid-dynamics-simulations-on-the-sandia-intel-4m2pkknsm9.pdf

Transient Solid Dynamics Simulations on the Sandia/Intel Teraflop Computer

Prandtl number effects on the stability of natural convection between spherical shells

A process for removing hulls from soybeans comprises the steps of placing the soybeans in an atmosphere of reduced pressure, subjecting the beans to microwave energy while the beans are within the atmosphere of reduced pressure, removing a portion of the moisture from the beans until the moisture content of the soybeans is about 746 to about 10% on a wet basis, removing the soybeans from the atmosphere of reduced pressure, the temperature of the soybeans leaving the reduced pressure atmosphere being at least 110° F and less than 160° F, immediately cracking the hulls of the hot soybeans without tempering, and removing the hulls from the hot soybeans

Process for removing soybean hulls

Natural convection in a Boussinesq fluid filling the narrow gap between two isothermal, concentric spheres at different temperatures depends strongly on radius ratio, Prandtl number, and Grashof number. When the inner sphere has a higher temperature than the outer sphere, and for fixed values of radius ratio and Prandtl number, experiments show the flow to be steady and axisymmetric for sufficiently small Grashof number and quasi-periodic and axisymmetric for Grashof numbers greater than a critical value. It is our hypothesis that the observed transition is a flow bifurcation. This hypothesis is examined by solving an appropriate eigenvalue problem. The critical Grashof number, critical eigenvalues, and corresponding eigenvectors are obtained as functions of the radius ratio, Prandtl number, and longitudinal wavenumber. Critical Grashof numbers range from 1.18 × 10 4 to 2.63 × 10 3 as Prandtl number Pr increases from zero to 0.7, for radius ratios of 0.900 and 0.950. A transitional Prandtl number Pr t exists such that for Pr Pr t the bifurcation is time-periodic and axisymmetric. For Pr > Pr t the bifurcation is steady and non-axisymmetric with wavenumber two A first approximation to the bifurcated flow is obtained using the critical eigenvectors. For Pr Pr t the bifurcation sets in as a cluster of relatively strong cells with alternating directions of rotation. The cells remain fixed in location, but pulsate with time. The cluster moves toward the top of the annulus as Pr increases toward Pr t . An important feature of the non-axisymmetric bifurcation for Pr > Pr t is a set of four cells located at each pole of the annulus in which the radial velocity alternates direction in moving from any one cell to an adjacent one. For fixed radius ratio, the average Nusselt number at criticality varies only slightly with Prandtl number.

https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S0022112090003500

Linear stability of natural convection in spherical annuli

Current supercomputers use large parallel arrays of tightly coupled processors to achieve levels of performance far surpassing conventional vector supercomputers. Shock-wave physics codes have been developed for these new supercomputers at Sandia National Laboratories and elsewhere. These parallel codes run fast enough on many simulations to consider using them to study the effects of varying design parameters on the performance of models of conventional munitions and other complex systems. Such studies maybe directed by optimization software to improve the performance of the modeled system. Using a shaped-charge jet design as an archetypal test case and the CTH parallel shock-wave physics code controlled by the Dakota optimization software, we explored the use of automatic optimization tools to optimize the design for conventional munitions. We used a scheme in which a lower resolution computational mesh was used to identify candidate optimal solutions and then these were verified using a higher resolution mesh. We identified three optimal solutions for the model and a region of the design domain where the jet tip speed is nearly optimal, indicating the possibility of a robust design. Based on this study we identified some of the difficulties in using high-fidelity models with optimization software to develop improved designs. These include developing robust algorithms for the objective function and constraints and mitigating the effects of numerical noise in them. We conclude that optimization software running high-fidelity models of physical systems using parallel shock wave physics codes to find improved designs can be a valuable tool for designers. While current state of algorithm and software development does not permit routine, ''black box'' optimization of designs, the effort involved in using the existing tools may well be worth the improvement achieved in designs.

/pdf/the-optimization-of-a-shaped-charge-design-using-parallel-f50w6o4njk.pdf

David R. Gardner

Papers

Transient Solid Dynamics Simulations on the Sandia/Intel Teraflop Computer

Prandtl number effects on the stability of natural convection between spherical shells

Process for removing soybean hulls

Linear stability of natural convection in spherical annuli

The Optimization of a Shaped-Charge Design Using Parallel Computers