Wireless Sensor Network Configuration -- Part II:
This is the second of a two-part investigation of
SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
This is the second of a two-part investigation of
The PRIMA-X (Performance Retargeting of Instrumentation, Measurement, and Analysis Technologies for Exascale Computing) project is the successor of the DOE PRIMA (Performance Refactoring of Instrumentation, Measurement, and Analysis Technologies for Petascale Computing) project, which addressed the challenge of creating a core measurement infrastructure that would serve as a common platform for both integrating leading parallel performance systems (notably TAU and Scalasca) and developing next-generation scalable performance tools. The PRIMA-X project shifts the focus away from refactorization of robust performance tools towards a re-targeting of the parallel performance measurement and analysis architecture for extreme scales. The massive concurrency, asynchronous execution dynamics, hardware heterogeneity, and multi-objective prerequisites (performance, power, resilience) that identify exascale systems introduce fundamental constraints on the ability to carry forward existing performance methodologies. In particular, there must be a deemphasis of per-thread observation techniques to significantly reduce the otherwise unsustainable flood of redundant performance data. Instead, it will be necessary to assimilate multi-level resource observations into macroscopic performance views, from which resilient performance metrics can be attributed to the computational features of the application. This requires a scalable framework for node-level and system-wide monitoring and runtime analyses of dynamic performance information. Also, the interest in optimizing parallelism parameters with respect to performance and energy drives the integration of tool capabilities in the exascale environment further. Initially, PRIMA-X was a collaborative project between the University of Oregon (lead institution) and the German Research School for Simulation Sciences (GRS). Because Prof. Wolf, the PI at GRS, accepted a position as full professor at Technische Universität Darmstadt (TU Darmstadt) starting February 1st, 2015, the project ended at GRS on January 31st, 2015. This report reflects the work accomplished at GRS until then. The work of GRS is expected to be continued at TU Darmstadt. The first main accomplishment of GRS is the design of different thread-level aggregation techniques. We created a prototype capable of aggregating the thread-level information in performance profiles using these techniques. The next step will be the integration of the most promising techniques into the Score-P measurement system and their evaluation. The second main accomplishment is a substantial increase of Score-P’s scalability, achieved by improving the design of the system-tree representation in Score-P’s profile format. We developed a new representation and a distributed algorithm to create the scalable system tree representation. Finally, we developed a lightweight approach to MPI wait-state profiling. Former algorithms either needed piggy-backing, which can cause significant runtime overhead, or tracing, which comes with its own set of scaling challenges. Our approach works with local data only and, thus, is scalable and has very little overhead.
Here, this paper considers a cooperative control problem in presence of unknown attacks. The attacker aims at destabilizing the consensus dynamics by intercepting the system’s communication network and corrupting its local state feedback. We first revisit the virtual network based resilient control proposed in our previous work and provide a new interpretation and insights into its implementation. Based on these insights, a novel distributed algorithm is presented to detect and identify the compromised communication links. It is shown that it is not possible for the adversary to launch a harmful and stealthy attack by only manipulating the physical states being exchanged via the network. In addition, a new virtual network is proposed which makes it more difficult for the adversary to launch a stealthy attack even though it is also able to manipulate information being exchanged via the virtual network. A numerical example demonstrates that the proposed control framework achieves simultaneously resilient operation and real-time attack identification.
Reliable broadcast is an essential tool to disseminate information among a set of nodes in the presence of failures. We present a novel study of reliable broadcast in interdependent networks, in which the failures in one network may cascade to another network. In particular, we focus on the interdependency between the communication network and power grid network, where the power grid depends on the signals from the communication network for control and the communication network depends on the grid for power. In this paper, we build a resilient solution to handle crash failures in the communication network that may cause cascading failures and may even partition the network. In order to guarantee that all the correct nodes deliver the messages, we use soft links, which are inactive backup links to non-neighboring nodes that are only active when failures occur. At the core of our work is a fully distributed algorithm for the nodes to predict and collect the information of cascading failures so that soft links can be maintained to correct nodes prior to the failures. In the presence of failures, soft links are activated to guarantee message delivery and new soft links are built accordingly for long term robustness. Our evaluation results show that the algorithm achieves low packet drop rate and handles cascading failures with little overhead.
Purpose: The purpose of this study is to develop a new calculation algorithm that is satisfactory in terms of the requirements for both accuracy and calculation time for a simulation of imaging of the proton-irradiated volume in a patient body in clinical proton therapy. Methods: The activity pencil beam algorithm (APB algorithm), which is a new technique to apply the pencil beam algorithm generally used for proton dose calculations in proton therapy to the calculation of activity distributions, was developed as a calculation algorithm of the activity distributions formed by positron emitter nuclei generated from target nuclear fragment reactions. In the APB algorithm, activity distributions are calculated using an activity pencil beam kernel. In addition, the activity pencil beam kernel is constructed using measured activity distributions in the depth direction and calculations in the lateral direction. {sup 12}C, {sup 16}O, and {sup 40}Ca nuclei were determined as the major target nuclei that constitute a human body that are of relevance for calculation of activity distributions. In this study, ''virtual positron emitter nuclei'' was defined as the integral yield of various positron emitter nuclei generated from each target nucleus by target nuclear fragment reactions with irradiated proton beam. Compounds, namely, polyethylene, water (including some gelatin) and calcium oxide, which contain plenty of the target nuclei, were irradiated using a proton beam. In addition, depth activity distributions of virtual positron emitter nuclei generated in each compound from target nuclear fragment reactions were measured using a beam ON-LINE PET system mounted a rotating gantry port (BOLPs-RGp). The measured activity distributions depend on depth or, in other words, energy. The irradiated proton beam energies were 138, 179, and 223 MeV, and measurement time was about 5 h until the measured activity reached the background level. Furthermore, the activity pencil beam data were made using the activity pencil beam kernel, which was composed of the measured depth data and the lateral data including multiple Coulomb scattering approximated by the Gaussian function, and were used for calculating activity distributions. Results: The data of measured depth activity distributions for every target nucleus by proton beam energy were obtained using BOLPs-RGp. The form of the depth activity distribution was verified, and the data were made in consideration of the time-dependent change of the form. Time dependence of an activity distribution form could be represented by two half-lives. Gaussian form of the lateral distribution of the activity pencil beam kernel was decided by the effect of multiple Coulomb scattering. Thus, the data of activity pencil beam involving time dependence could be obtained in this study. Conclusions: The simulation of imaging of the proton-irradiated volume in a patient body using target nuclear fragment reactions was feasible with the developed APB algorithm taking time dependence into account. With the use of the APB algorithm, it was suggested that a system of simulation of activity distributions that has levels of both accuracy and calculation time appropriate for clinical use can be constructed.
We design and implement an efficient parallel approximation algorithm for the problem of maximum weight perfect matching in bipartite graphs, i.e. the problem of finding a set of non-adjacent edges that covers all vertices and has maximum weight. This problem differs from the maximum weight matching problem, for which scalable approximation algorithms are known. It is primarily motivated by finding good pivots in scalable sparse direct solvers before factorization where sequential implementations of maximum weight perfect matching algorithms, such as those available in MC64, are widely used due to the lack of scalable alternatives. To overcome this limitation, we propose a fully parallel distributed memory algorithm that first generates a perfect matching and then searches for weightaugmenting cycles of length four in parallel and iteratively augments the matching with a vertex disjoint set of such cycles. For most practical problems the weights of the perfect matchings generated by our algorithm are very close to the optimum. An efficient implementation of the algorithm scales up to 256 nodes (17,408 cores) on a Cray XC40 supercomputer and can solve instances that are too large to be handled by a single node using the sequential algorithm.
This paper presents scalable algorithms and data structures for adaptive mesh refinement computations. We describe a novel mesh restructuring algorithm for adaptive mesh refinement computations that uses a constant number of collectives regardless of the refinement depth. To further increase scalability, we describe a localized hierarchical coordinate-based block indexing scheme in contrast to traditional linear numbering schemes, which incur unnecessary synchronization. In contrast to the existing approaches which take O(P) time and storage per process, our approach takes only constant time and has very small memory footprint. With these optimizations as well as an efficient mapping scheme, our algorithm is scalable and suitable for large, highly-refined meshes. We present strong-scaling experiments up to 2k ranks on Cray XK6, and 32k ranks on IBM Blue Gene/Q.
In this study, we introduce new intersection-distribution-based remapping tools for indirect staggered arbitrary Lagrangian-Eulerian (ALE) simulations of multi-material shock hydrodynamics on arbitrary meshes. In addition to conserving momentum and total energy, the three-stage remapper proposed in this work preserves non-negativity of the internal energy. At the first stage, we construct slope-limited piecewise-linear reconstructions of all conserved quantities on zones of the source mesh and perform intersection-based remap to obtain bound-preserving zonal quantities on the target mesh. At the second stage, we define bound-preserving nodal quantities of the staggered ALE discretization as convex combinations of corner quantities. The nodal internal energy is corrected in a way which keeps it non-negative, while providing exact conservation of total energy. At the final stage, we distribute the non-negative nodal internal energy to corners, zones and materials using non-negative weights. Proofs of positivity preservation are provided for each stage. This work is a natural extension of our paper [14] in which a similar intersection-distribution-based remapping procedure was employed. The original version used a nodal kinetic energy fix which did not provably ensure positivity preservation for the zonal internal energy after the final distribution stage. The new algorithm cures this potential drawback by using ‘coordinated’ limiters for piecewise-linear reconstructions, remapping the internal energy to nodes and correcting it before redistribution. The effectiveness of the new nodal fix is illustrated by numerical examples.
Here, this paper considers the problem of designing a resilient distributed voltage control algorithm for distribution systems with high penetration of distributed energy resources in the presence of an unknown cyber-attack. The purpose of the attack is to force the system to violate the operating voltage limit by intercepting its communication channels and inserting exogenous signals to perturb and/or modify the information being exchanged. We first review the cooperative voltage control proposed in our previous work and provide a new stability analysis for it. Next, we present a resilient cooperative voltage control algorithm by introducing a virtual system interconnected with the original system such that the voltage can be maintained within the operational limit under unknown attacks. The resiliency of the proposed algorithm is demonstrated via simulations on the IEEE 8500-node system when subjected to an attack which consists of corrupting the data being exchanged in the communication network between two generation units.
Large-scale datasets in computational chemistry typically require distributed-memory parallel methods to perform a special operation known as tensor contraction. Tensors are multidimensional arrays, and a tensor contraction is akin to matrix multiplication with special types of permutations. Creating an efficient algorithm and optimized implementation in this domain is complex, tedious, and error-prone. To address this, we develop a notation to express data distributions so that we can apply use automated methods to find optimized implementations for tensor contractions. We consider the spin-adapted coupled cluster singles and doubles method from computational chemistry and use our methodology to produce an efficient implementation. Experiments per- formed on the IBM Blue Gene/Q and Cray XC30 demonstrate impact both improved performance and reduced memory consumption.
Graph matching and coloring constitute two fundamental classes of combinatorial problems having numerous established as well as emerging applications in computational science and engineering, high-performance computing, and informatics. We provide a snapshot of an on-going work on the design and implementation of new highly-scalable distributed-memory parallel algorithms for two prototypical problems from these classes, edge-weighted matching and distance-1 vertex coloring. Graph algorithms in general have low concurrency and poor data locality, making it challenging to achieve scalability on massively parallel machines. We overcome this challenge by employing a variety of techniques, including approximation, speculation and iteration, optimized communication, and randomization, in concert. We present preliminary results on weak and strong scalability studies conducted on an IBM Blue Gene/P machine employing up to tens of thousands of processors. The results show that the algorithms hold strong potential for computing at petascale.
Not Available
Purpose: Recently, much research on imaging the clinical proton-irradiated volume using positron emitter nuclei based on target nuclear fragment reaction has been carried out. The purpose of this study is to develop an activity pencil beam (APB) algorithm for a simulation system for proton-activated positron-emitting imaging in clinical proton therapy using spread-out Bragg peak (SOBP) beams.Methods: The target nuclei of activity distribution calculations are {sup 12}C nuclei, {sup 16}O nuclei, and {sup 40}Ca nuclei, which are the main elements in a human body. Depth activity distributions with SOBP beam irradiations were obtained from the material information of ridge filter (RF) and depth activity distributions of compounds of the three target nuclei measured by BOLPs-RGp (beam ON-LINE PET system mounted on a rotating gantry port) with mono-energetic Bragg peak (MONO) beam irradiations. The calculated data of depth activity distributions with SOBP beam irradiations were sorted in terms of kind of nucleus, energy of proton beam, SOBP width, and thickness of fine degrader (FD), which were verified. The calculated depth activity distributions with SOBP beam irradiations were compared with the measured ones. APB kernels were made from the calculated depth activity distributions with SOBP beam irradiations to construct a simulation system using the APB algorithm for SOBP beams.Results: The depth activity distributions were prepared using the material information of RF and the measured depth activity distributions with MONO beam irradiations for clinical therapy using SOBP beams. With the SOBP width widening, the distal fall-offs of depth activity distributions and the difference from the depth dose distributions were large. The shapes of the calculated depth activity distributions nearly agreed with those of the measured ones upon comparison between the two. The APB kernels of SOBP beams were prepared by making use of the data on depth activity distributions with SOBP beam irradiations that were made from the depth activity distributions with MONO beam irradiations and sorted in terms of energy, SOBP width, and thickness of FD. The data on APB kernels of SOBP beams were determined as installment data for the simulation system using the APB algorithm for SOBP beam irradiations.Conclusions: A method of obtaining the depth activity distributions and the APB algorithm for clinical use of SOBP beams have been developed. It is suggested that the simulation system for imaging the clinical irradiated volume with the APB algorithm can be used in clinical proton therapy using SOBP beams by preparing and investigating the data on APB kernels of SOBP beams.
Not Available
Ordering vertices of a graph is key to minimize fill-in and data structure size in sparse direct solvers, maximize locality in iterative solvers, and improve performance in graph algorithms. Except for naturally parallelizable ordering methods such as nested dissection, many important ordering methods have not been efficiently mapped to distributed-memory architectures. In this paper, we present the first-ever distributed-memory implementation of the reverse Cuthill-McKee (RCM) algorithm for reducing the profile of a sparse matrix. Our parallelization uses a two-dimensional sparse matrix decomposition. We achieve high performance by decomposing the problem into a small number of primitives and utilizing optimized implementations of these primitives. Our implementation shows strong scaling up to 1024 cores for smaller matrices and up to 4096 cores for larger matrices.
In this study, we investigate agent-based approach for system model identification with emphasis on power distribution system applications. Departing from conventional practices of relying on historical data for offline model identification, we adopt online update approach utilizing real-time data by employing the latest data points for gradient computation. This methodology offers advantages including a large reduction in the communication network's bandwidth requirements by minimizing the data exchanged at each iteration and enabling the model to adapt in real-time to disturbances. Furthermore, we extend our model identification process from linear frameworks to more complex non-linear convex models. This extension is validated through numerical studies demonstrating improved control performance for a synthetic IEEE test case.
Real-time cyber-physical systems and information processing clusters require system designers to consider the total latency involved in collecting and aggregating data. For example, applications such as wild-fire monitoring require data to be presented to users in a timely manner. However, most models and algorithms for sensor networks have focused on alternative metrics such as energy efficiency. In this paper, we present a new model of sensor network aggregation that focuses on total latency. Our model is flexible and enables users to configure varying transmission and computation time on a node-by-node basis, and thus enables the simulation of complex computational phenomena. In addition, we present results from three tasking algorithms that trade-off local communication for overall latency performance. These algorithms are evaluated in simulated networks of up to 200 nodes. We've presented an aggregation-focused model of sensor networks that can be used to study the trade-offs between computational coverage and total latency. Our model explicitly takes into account transmission and computation times, and enables users to define different values for the basestation. In addition, we've presented three different tasking algorithms that operate over model to produce aggregation schedules of varying quality. In the future, we expect to continue exploring distributed tasking algorithms for information processing systems. We've shown that the gap between highly optimized schedules that use global information is quite large relative to our distributed algorithms. This gives us encouragement that future distributed tasking algorithms can still make large gains.
This paper generalizes the parallel selected inversion algorithm called PSelInv to sparse non-symmetric matrices. We assume a general sparse matrix A has been decomposed as $PAQ$ = $LU$ on a distributed memory parallel machine, where L, U are lower and upper triangular matrices, and P, Q are permutation matrices, respectively. The PSelInvmethod computes selected elements of $A$ -1 . The selection is confined by the sparsity pattern of the matrix A T . Here our algorithm does not assume any symmetry properties of A, and our parallel implementation is memory efficient, in the sense that the computed elements of $A$ -T overwrites the sparse matrix $L + U$ in situ . PSelInv involves a large number of collective data communication activities within different processor groups of various sizes. In order to minimize idle time and improve load balancing, tree-based asynchronous communication is used to coordinate all such collective communication. Numerical results demonstrate that PSelInv can scale efficiently to 6,400 cores for a variety of matrices.