Highly accelerated simulations of glassy dynamics using GPUs: Caveats on limited floating-point precision

Modern graphics processing units (GPUs) provide impressive computing resources, which can be accessed conveniently through the CUDA programming interface. We describe how GPUs can be used to considerably speed up molecular dynamics (MD) simulations for system sizes ranging up to about 1 million particles. Particular emphasis is put on the numerical long-time stability in terms of energy and momentum conservation, and caveats on limited floating-point precision are issued. Strict energy conservation over 10⁸ MD steps is obtained by double-single emulation of the floating-point arithmetic in accuracy-critical parts of the algorithm. For the slow dynamics of a supercooled binary Lennard-Jones mixture, we demonstrate that the use of single-floating point precision may result in quantitatively and even physically wrong results. For simulations of a Lennard-Jones fluid, the described implementation shows speedup factors of up to 80 compared to a serial implementation for the CPU, and a single GPU was found to compare with a parallelised MD simulation using 64 distributed cores.     

A probabilistic simulator for population dynamics of quasispecies

We constructed a probabilistic simulator that allows all the events in population dynamics such as death, birth, mutation, and suppression/stimulation to be described by probabilistic rules. The simulator also facilitates a lattice used for expressing distribution and diversity (number of distinct strains) of quasispecies. The simulator is used to investigate the diversity threshold in HIV and T-cell interaction. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Exploiting hierarchy parallelism for molecular dynamics on a petascale heterogeneous system

Heterogeneous systems with nodes containing more than one type of computation units, e.g., central processing units (CPUs) and graphics processing units (GPUs), are becoming popular because of their low cost and high performance. In this paper, we have developed a Three-Level Parallelization Scheme (TLPS) for molecular dynamics (MD) simulation on heterogeneous systems. The scheme exploits multi-level parallelism combining (1) inter-node parallelism using spatial decomposition via message passing, (2) intra-node parallelism using spatial decomposition via dynamically scheduled multi-threading, and (3) intra-chip parallelism using multi-threading and short vector extension in CPUs, and employing multiple CUDA threads in GPUs. By using a hierarchy of parallelism with optimizations such as communication hiding intra-node, and memory optimizations in both CPUs and GPUs, we have implemented and evaluated a MD simulation on a petascale heterogeneous supercomputer TH-1A. The results show that MD simulations can be efficiently parallelized with our TLPS scheme and can benefit from the optimizations.     

Automatizing the online filter test management for a general-purpose particle detector

This paper presents a software environment to automatically configure and run online triggering and dataflow farms for the ATLAS experiment at the Large Hadron Collider (LHC). It provides support for a broad set of users, with distinct knowledge about the online triggering system, ranging from casual testers to final system deployers. This level of automatization improves the overall ATLAS TDAQ work flow for software and hardware tests and speeds-up system modifications and deployment.     

A hardware-in-the-loop dynamics simulator for motorcycle rapid controller prototyping

The purpose of this research is to develop an innovative hardware-in-the-loop simulator for the purpose of motorcycle power train rapid controller prototyping. Proposed control algorithms can be validated using the developed setup. Such an in-lab validation saves time and development cost and thus is preferable during power train controller development process. The developed simulator includes an engine, a transmission, and a rear wheel from a real motorcycle. A powder brake is rigidly coupled to the rear wheel for road loading generation. A central computer is used to control the operation of the system and the measurement of system dynamics variables. An xPC system is also integrated in the system to provide the feasibility of power train control algorithm rapid prototyping. Comparison between field tests and simulation results show that the simulator can be used to evaluate the performance of different control algorithms for controller rapid prototyping in the laboratory. An example of power train control algorithm development featuring engine fuel injection control using the above rapid prototyping setup is also described.     

Path-following for linear systems with unstable zero dynamics

A constructive solution to the path-following problem for MIMO linear systems with unstable zero dynamics is developed. While the original control variable steers the system output along the path, the path parameter θ is used as an additional control to stabilize zero dynamics with a feedback law which is nonlinear due to the path constraint. A sufficient condition for solvability of the path-following problem is given in terms of the geometric properties of the path. When this condition is satisfied, an arbitrary small L₂ norm of path-following error can be achieved, thus avoiding performance limitations of the standard reference tracking problem imposed by unstable zero dynamics.     

Towards accelerating smoothed particle hydrodynamics simulations for free-surface flows on multi-GPU clusters

Starting from the single graphics processing unit (GPU) version of the Smoothed Particle Hydrodynamics (SPH) code DualSPHysics, a multi-GPU SPH program is developed for free-surface flows. The approach is based on a spatial decomposition technique, whereby different portions (sub-domains) of the physical system under study are assigned to different GPUs. Communication between devices is achieved with the use of Message Passing Interface (MPI) application programming interface (API) routines. The use of the sorting algorithm radix sort for inter-GPU particle migration and sub-domain “halo” building (which enables interaction between SPH particles of different sub-domains) is described in detail. With the resulting scheme it is possible, on the one hand, to carry out simulations that could also be performed on a single GPU, but they can now be performed even faster than on one of these devices alone. On the other hand, accelerated simulations can be performed with up to 32 million particles on the current architecture, which is beyond the limitations of a single GPU due to memory constraints. A study of weak and strong scaling behaviour, speedups and efficiency of the resulting program is presented including an investigation to elucidate the computational bottlenecks. Last, possibilities for reduction of the effects of overhead on computational efficiency in future versions of our scheme are discussed.     

Distributed event-triggered output consensus control for heterogeneous multi-agent system with general linear dynamics

This paper studies the distributed event-triggered output consensus problem of heterogeneous multi-agent system with general linear dynamics under an undirected connected graph. With the state-dependent triggering function, we design a novel distributed event-triggered output consensus controller for each agent to reach consensus with zero final consensus error. This strategy has several distinguishing features, including the fact that individual agent does not require continuous, or even periodic, communication with their neighbours to update the controller or monitor the triggering condition, and all parameters required by its implementation can be locally determined by the agent. We also prove that events cannot be triggered infinitely in finite time (i.e. no Zeno behaviour). Furthermore, the simulation examples are given to illustrate the theoretical analysis.     

A design methodology for switched discrete time linear systems with applications to automotive roll dynamics control

In this paper we consider the asymptotic stability of a class of discrete-time switching linear systems, where each of the constituent subsystems is Schur stable. We first present an example to motivate our study, which illustrates that the bilinear transform does not preserve the stability of a class of switched linear systems. Consequently, continuous time stability results cannot be transformed to discrete time analogs using this transformation. We then present a subclass of discrete-time switching systems that arise frequently in practical applications. We prove that global attractivity for this subclass can be obtained without requiring the existence of a common quadratic Lyapunov function (CQLF). Using this result, we present a synthesis procedure to construct switching stabilizing controllers for an automotive control problem, which is related to the stabilization of a vehicle’s roll dynamics subject to switches in the center of gravitys (CG) height.     

Distributed adaptive control and stability verification for linear multiagent systems with heterogeneous actuator dynamics and system uncertainties

The contribution of this paper is a control synthesis and stability verification framework for linear time-invariant multiagent systems with heterogeneous actuator dynamics and system uncertainties. In particular, we first propose a distributed adaptive control architecture in a leader-follower setting for this class of high-order multiagent systems. The proposed architecture uses a hedging method, which alters the ideal reference model dynamics of each agent in order to ensure correct adaptation in the presence of heterogeneous actuator dynamics of these agents. We then use Lyapunov stability theory and linear matrix inequalities to analyse the proposed architecture. This analysis reveals a stability condition, where evaluation of this condition with respect to a given graph topology allows stability verification of the controlled multiagent system. From a practical point of view, this condition also shows a fundamental tradeoff between heterogeneous agent actuation capabilities and unknown parameters in agent dynamics. Several illustrative numerical examples are also provided to demonstrate the efficacy of the proposed architecture.