Proceedings of ExaMPI 2020: Exascale MPI Workshop, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis
Multithreaded MPI applications are gaining popularity in scientific and high-performance computing. While the combination of programming models is suited to support current parallel hardware, it moves threading models and their interaction with MPI into focus. With the advent of new threading libraries, the flexibility to select threading implementations of choice is becoming an important usability feature. Open MPI has traditionally avoided componentizing its threading model, relying on code inlining and static initialization to minimize potential impacts on runtime fast paths and synchronization. This paper describes the implementation of a generic threading runtime support in Open MPI using the Opal Modular Component Architecture. This architecture allows the programmer to select a threading library at compile-or run-time, providing both static initialization of threading primitives as well as dynamic instantiation of threading objects. In this work, we present the implementation, define required interfaces, and discuss trade-offs of dynamic and static initialization.
This report summarizes the work performed under the project "Linear Programming in Strongly Polynomial Time." Linear programming (LP) is a classic combinatorial optimization problem heavily used directly and as an enabling subroutine in integer programming (IP). Specifically IP is the same as LP except that some solution variables must take integer values (e.g. to represent yes/no decisions). Together LP and IP have many applications in resource allocation including general logistics, and infrastructure design and vulnerability analysis. The project was motivated by the PI's recent success developing methods to efficiently sample Voronoi vertices (essentially finding nearest neighbors in high-dimensional point sets) in arbitrary dimension. His method seems applicable to exploring the high-dimensional convex feasible space of an LP problem. Although the project did not provably find a strongly-polynomial algorithm, it explored multiple algorithm classes. The new medial simplex algorithms may still lead to solvers with improved provable complexity. We describe medial simplex algorithms and some relevant structural/complexity results. We also designed a novel parallel LP algorithm based on our geometric insights and implemented it in the Spoke-LP code. A major part of the computational step is many independent vector dot products. Our parallel algorithm distributes the problem constraints across processors. Current commercial and high-quality free LP solvers require all problem details to fit onto a single processor or multicore. Our new algorithm might enable the solution of problems too large for any current LP solvers. We describe our new algorithm, give preliminary proof-of-concept experiments, and describe a new generator for arbitrarily large LP instances.
Predictions of the bulk scale thermal conductivity of solids using non-equilibrium molecular dynamics (MD) simulations have relied on the linear extrapolation of the thermal resistivity versus the reciprocal of the system length in the simulations. Several studies have reported deviation of the extrapolation from linearity near the micro-scale, raising a concern of its applicability to large systems. To investigate this issue, present work conducted extensive MD simulations of silicon with two different potentials (EDIP and Tersoff-II) for unprecedented length scales up to 10.3 μm and simulation times up to 530 ns. For large systems ≥0.35 μm in size the non-linearity of the extrapolation of the reciprocal of the thermal conductivity is mostly due to ignoring the dependence of the thermal conductivity on temperature. To account for such dependence, the present analysis fixes the temperature range for determining the gradient for calculating the thermal conductivity values. However, short systems ≤0.23 μm in size show significant non-linearity in the calculated thermal conductivity values using a temperature window of 500 ± 10 K from the simulations results with the EDIP potential. Since these system sizes are shorter than the mean phonon free path in EDIP (~0.22 μm), the nonlinearity may be attributed to phonon transport. For the MD simulations with the Tersoff-II potential there is no significant non-linearity in the calculated thermal conductivity values for systems ranging in size from 0.05 to 5.4 μm.
Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis
Dominguez-Trujillo, Jered; Haskins, Keira; Khouzani, Soheila J.; Leap, Christopher; Tashakkori, Sahba; Wofford, Quincy; Estrada, Trilce; Bridges, Patrick G.; Widener, Patrick W.
Performance variation deriving from hardware and software sources is common in modern scientific and data-intensive computing systems, and synchronization in parallel and distributed programs often exacerbates their impacts at scale. The decentralized and emergent effects of such variation are, unfortunately, also difficult to systematically measure, analyze, and predict; modeling assumptions which are stringent enough to make analysis tractable frequently cannot be guaranteed at meaningful application scales, and longitudinal methods at such scales can require the capture and manipulation of impractically large amounts of data. This paper describes a new, scalable, and statistically robust approach for effective modeling, measurement, and analysis of large-scale performance variation in HPC systems. Our approach avoids the need to reason about complex distributions of runtimes among large numbers of individual application processes by focusing instead on the maximum length of distributed workload intervals. We describe this approach and its implementation in MPI which makes it applicable to a diverse set of HPC workloads. We also present evaluations of these techniques for quantifying and predicting performance variation carried out on large-scale computing systems, and discuss the strengths and limitations of the underlying modeling assumptions.
Research shows that individuals often overestimate their knowledge and performance without realizing they have done so, which can lead to faulty technical outcomes. This phenomenon is known as the Dunning-Kruger effect (Kruger & Dunning, 1999). This research sought to determine if some individuals were more prone to overestimating their performance due to underlying personality and cognitive characteristics. To test our hypothesis, we first collected individual difference measures. Next, we asked participants to estimate their performance on three performance tasks to assess the likelihood of overestimation. We found that some individuals may be more prone to overestimating their performance than others, and that faulty problem-solving abilities and low skill may be to blame. Encouraging individuals to think critically through all options and to consult with others before making a high-consequence decision may reduce overestimation.