Search NASA⌕ Search

SEARCH · Search NASA

Results for “Dynamic load balancing”

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.

At least 361 records · Page 20

Development of an analytical tool to study power quality of ac power systems for large spacecraft

A harmonic power flow program applicable to space power systems with sources of harmonic distortion is described. The algorithm is a modification of Electric Power Research Institute's HARMFLO program which assumes a three-phase, balanced, ac system with loads of harmonic distortion. The modified power flow program can be used with single phase, ac systems. Early results indicate that the required modifications and the models developed are quite adequate for the analysis of a 20-kHz testbed built by General Dynamics Corporation. This is demonstrated by the acceptable correlation of the present results with published data. Although the results are not exact, the discrepancies are relatively small.

Kraft, L. A.↗

Application and use of Multi Body Dynamic Simulations for In Space Mechanisms

The Thermal and Mechanical Analysis Branch of the Space System Department (ES22) is actively growing and expanding a new mechanical analysis capability. In recent years, there has been an increase in the demand for Multi Body Dynamics (MBD) analysis to support small ESPA ring size spacecraft, cubesats, lunar surface operations and large spacecraft with complex operations or deployables. Requirements can be balanced for both needs to ensure mechanism is verified for launch. Multi-body dynamics is the study of multiple rigid or flexible bodies that are linked together or are in contact and the associated kinematics and dynamics of the system. The typical outputs include performance information, loads, and deflections of mechanical systems. They routinely support the development of robotic mechanisms, deployment mechanisms, landing simulations, control systems, and conops planning.

Multi Body Dynamics Analysis↗

Monitoring the dynamics of miscible P3HT:PCBM blends: A quasi elastic neutron scattering study of organic photovoltaic active layers

Our work reports on the detailed molecular dynamic behavior of miscible blends of Poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) and their pure counterparts by quasi-elastic neutron scattering measurements (QENS). The study provides the measure of relaxation processes on pico-to-nanosecond time scales. A single relaxation process was observed in pure P3HT and PCBM while two relaxation processes, one fast and one slow, were observed in the blends. The fast process was attributed to the dynamics of P3HT while the slow process was correlated to the dynamics of PCBM. The results show that the relaxation process is a balance between two opposing effects: increased mobility due to thermal activation of P3HT molecules and decrease mobility due to the presence of PCBM which is correlated to the percent crystallinity of P3HT and local packing density of PCBM in the amorphous phase. The threshold for the domination of the thermally activated relaxation is between 5 and 9 vol.% of PCBM loading. Two distinct spatial dependences of the relaxation processes, in which the crossover length scale depends neither on temperature nor composition, were observed for all the samples. They were attributed to the collective motions of the hexyl side chains and the rotational motions of the C-C single bonds of the side chains. Finally, these results provide an understanding of the effects of PCBM loading and temperature on the dynamics of the polymer-fullerene blends which provides a tool to optimize the efficiency of charge carrier and exciton transport within the organic photovoltaic (OPV) active layer to improve the high performance of organic solar cells.

14 SOLAR ENERGY↗

A data-driven linear formulation of the optimal demand response scheduling problem for an industrial air separation unit

Demand response (DR) has become a key element in balancing the power grid as the contribution of time-varying renewable power generation increases. Chemical plants are appealing candidates for DR programs as they offer large, concentrated and flexible loads. DR participation calls for frequent production rate changes over time scales that overlap with the dominant dynamics of the plant. Production scheduling should therefore consider the process dynamics explicitly. Here we present a data-driven approach for modelling the scheduling-relevant dynamics based on historical closed-loop operating data using autoregressive with extra inputs (ARX) models. We introduce a new, linear scheduling problem formulation based on the ARX representation, and demonstrate its implementation on an industrial air separation unit.

42 ENGINEERING↗

Scaling Irregular Applications through Data Aggregation and Software Multithreading

Bioinformatics, data analytics, semantic databases, knowledge discovery are emerging high performance application areas that exploit dynamic, linked data structures such as graphs, unbalanced trees or unstructured grids. These data structures usually are very large, requiring significantly more memory than available on single shared memory systems. Additionally, these data structures are difficult to partition on distributed memory systems. They also present poor spatial and temporal locality, thus generating unpredictable memory and network accesses. The Partitioned Global Address Space (PGAS) programming model seems suitable for these applications, because it allows using a shared memory abstraction across distributed-memory clusters. However, current PGAS languages and libraries are built to target regular remote data accesses and block transfers. Furthermore, they usually rely on the Single Program Multiple Data (SPMD) parallel control model, which is not well suited to the fine grained, dynamic and unbalanced parallelism of irregular applications. In this paper we present {\bf GMT} (Global Memory and Threading library), a custom runtime library that enables efficient execution of irregular applications on commodity clusters. GMT integrates a PGAS data substrate with simple fork/join parallelism and provides automatic load balancing on a per node basis. It implements multi-level aggregation and lightweight multithreading to maximize memory and network bandwidth with fine-grained data accesses and tolerate long data access latencies. A key innovation in the GMT runtime is its thread specialization (workers, helpers and communication threads) that realize the overall functionality. We compare our approach with other PGAS models, such as UPC running using GASNet, and hand-optimized MPI code on a set of typical large-scale irregular applications, demonstrating speedups of an order of magnitude.

Aggregation↗

A Parallelized Oxidation-Driven Surface Recession Framework in DSMC Code, SPARTA

Spacecrafts rely on ablative thermal protection systems (TPS) made of composites consisting of a carbon-based reinforcement and a polymeric matrix. These materials are designed to withstand high-temperature oxidation and surface recession during re-entry into the Earth's atmosphere. However, ablation occurs due to a complex interplay of thermal, mechanical, and chemical factors, making it challenging to determine the individual impact of each on the TPS's overall degradation. In this study, we have developed an ablation model that can leverage a finite rate carbon oxidation model to predict material recession and surface states more accurately. Stochastic PArallel Rarified-gas Time-accurate Analyzer (SPARTA), a direct-simulation Monte Carlo (DSMC) code, is modified to allow oxidation-driven ablation of implicitly defined carbon surfaces. In SPARTA, implicit surfaces are generated from the grid corner point values via a marching cubes algorithm, therefore creating a new set of surface elements every time ablation is performed. The finite-rate oxidation model developed by Gopalan et. al can perform both gas-surface and pure-surface reactions and is now adapted to tally surface data on a per grid cell basis. The ablation functionality was also adjusted so once the reactions have occurred, the number of reactions leading to CO formation can be converted to corner point reduction values; therefore, carbon removal is directly proportional to surface recession. We also briefly discuss some unique challenges associated with parallelizing this dynamic surface state and geometry. Finally, we analyze the performance of this parallelized implicit chemistry model with simple 2D and 3D benchmark cases by producing surface state statistics, area changes over time, and visualization across a range of surface temperatures and processors with and without load-balancing.

DSMC↗

Plant Operation and Control Strategy for the PGSFR

The plant operating strategy and control system design for the PGSFR plant are being developed to meet functional and design requirements for the plant. The functional requirements are derived from the mission of the reactor while the design requirements are related to the reactor technologies that underlie the PGSFR. This paper describes the operation and control strategy developed to meet these requirements and the implementation of that strategy. Representative results are presented for the dynamic behavior of the nuclear steam supply system (NSSS) during operational transients. A supervisory control system has been designed for the PGSFR for Startup, Reactor Leading, and Load Following modes and it performance assessed for operational transients in the Load Following mode. The control system extends into the balance of plant to include regulation of turbine inlet pressure and steam generator pressure and includes a feedwater pump with prototypic operating characteristics. The results indicate a very stable plant response with minimal deviation from quasi-static conditions for the transients simulated. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Application of Intelligent Load Control to Manage Building Loads to Support Rapid Growth of Distributed Renewable Generation

Electricity utilities are faced with the mounting challenge of providing a stable supply of power to meet the growing demand while also integrating rapid growth in distributed variable renewable generation. Traditional means of balancing short- and long-term supply and demand imbalance will be expensive. Alternative approaches of using flexible loads in buildings are needed to mitigate the imbalance at a lower cost. This paper shows how the intelligent load control (ILC) process can be used to manage loads in buildings by dynamically prioritizing loads for curtailment using both quantitative and qualitative criteria. The ILC process can be deployed on low-cost computing platforms without the need for any additional sensing. ILC was first validated in a simulation environment to provide two grid service use cases: (1) managing monthly peak electricity demand and (2) managing buildings’ electricity consumption during a capacity bidding event. After being successfully tested in a simulation environment, ILC was deployed on real buildings to manage electricity consumption to provide two different use cases under different outdoor operating conditions. Both the simulation tests and the real building experiments were deployed using VOLTTRON™, a distributed sensing and control platform. The results from the tests and experiments showed that ILC was able to manage the controllable loads (heat pumps) in the building to maintain the electricity consumption at the desired level without a significant impact on occupant comfort. Overall, the results demonstrate that the ILC allows coordination of the controllable loads and provides a more intelligent means of load management than the traditional duty-cycling approach.

Kim, Woohyun↗

Research and test facilities

A description is given of each of the following Langley research and test facilities: 0.3-Meter Transonic Cryogenic Tunnel, 7-by 10-Foot High Speed Tunnel, 8-Foot Transonic Pressure Tunnel, 13-Inch Magnetic Suspension & Balance System, 14-by 22-Foot Subsonic Tunnel, 16-Foot Transonic Tunnel, 16-by 24-Inch Water Tunnel, 20-Foot Vertical Spin Tunnel, 30-by 60-Foot Wind Tunnel, Advanced Civil Transport Simulator (ACTS), Advanced Technology Research Laboratory, Aerospace Controls Research Laboratory (ACRL), Aerothermal Loads Complex, Aircraft Landing Dynamics Facility (ALDF), Avionics Integration Research Laboratory, Basic Aerodynamics Research Tunnel (BART), Compact Range Test Facility, Differential Maneuvering Simulator (DMS), Enhanced/Synthetic Vision & Spatial Displays Laboratory, Experimental Test Range (ETR) Flight Research Facility, General Aviation Simulator (GAS), High Intensity Radiated Fields Facility, Human Engineering Methods Laboratory, Hypersonic Facilities Complex, Impact Dynamics Research Facility, Jet Noise Laboratory & Anechoic Jet Facility, Light Alloy Laboratory, Low Frequency Antenna Test Facility, Low Turbulence Pressure Tunnel, Mechanics of Metals Laboratory, National Transonic Facility (NTF), NDE Research Laboratory, Polymers & Composites Laboratory, Pyrotechnic Test Facility, Quiet Flow Facility, Robotics Facilities, Scientific Visualization System, Scramjet Test Complex, Space Materials Research Laboratory, Space Simulation & Environmental Test Complex, Structural Dynamics Research Laboratory, Structural Dynamics Test Beds, Structures & Materials Research Laboratory, Supersonic Low Disturbance Pilot Tunnel, Thermal Acoustic Fatigue Apparatus (TAFA), Transonic Dynamics Tunnel (TDT), Transport Systems Research Vehicle, Unitary Plan Wind Tunnel, and the Visual Motion Simulator (VMS).

Source record↗

The mechanisms and seasonal differences of the impact of aerosols on daytime surface urban heat island effect

The urban heat island intensity (UHII) is the temperature difference between urban areas and their rural surroundings. It is commonly attributed to changes in the underlying surface structure caused by urbanization. Air pollution caused by aerosol particles can affect the UHII through changing (1) the surface energy balance by the aerosol radiative effect (ARE) and (2) planetary-boundary-layer (PBL) stability and airflow intensity by modifying thermodynamic structure, which is referred to as the aerosol dynamic effect (ADE). By analyzing satellite data and ground-based observations collected from 2001 to 2010 at 35 cities in China and using the WRF-Chem model, we find that the impact of aerosols on UHII differs considerably: reducing the UHII in summer but increasing the UHII in winter. This seasonal contrast is proposed to be caused by the different strengths of the ARE and ADE between summer and winter. In summer, the ARE on UHII is dominant over the ADE, cooling down surface temperature more strongly in urban areas than in rural areas because of much higher aerosol loading, and offsets the urban heating, therefore weakening UHII. In winter, however, the ADE is more dominant, because aerosols stabilize the PBL more in the polluted condition, weakening the near-surface heat transport over urban areas in both vertical and horizontal directions. This means that the heat accumulated in urban areas is dispersed less effectively, and thus the UHII is enhanced. These findings shed new light on the impact of the interaction between urbanization-induced surface changes and air pollution on urban climate.urban areas and their rural surroundings. It is commonly attributed to changes in the underlying surface structure caused by urbanization. Air pollution caused by aerosol particles can affect the UHII through changing (1) the surface energy balance by the aerosol radiative effect (ARE) and (2) planetary-boundary-layer (PBL) stability and airflow intensity by modifying thermodynamic structure, which is referred to as the aerosol dynamic effect (ADE). By analyzing satellite data and ground-based observations collected from 2001 to 2010 at 35 cities in China and using the WRF-Chem model, we find that the impact of aerosols on UHII differs considerably: reducing the UHII in summer but increasing the UHII in winter. This seasonal contrast is proposed to be caused by the different strengths of the ARE and ADE between summer and winter. In summer, the ARE on UHII is dominant over the ADE, cooling down surface temperature more strongly in urban areas than in rural areas because of much higher aerosol loading, and offsets the urban heating, therefore weakening UHII. In winter, however, the ADE is more dominant, because aerosols stabilize the PBL more in the polluted condition, weakening the near-surface heat transport over urban areas in both vertical and horizontal directions. This means that the heat accumulated in urban areas is dispersed less effectively, and thus the UHII is enhanced. These findings shed new light on the impact of the interaction between urbanization-induced surface changes and air pollution on urban climate.

54 ENVIRONMENTAL SCIENCES↗

Enabling Staged Pressurized Oxy Combustion (SPOC): Improving Flexibility and Performance at Reduced Cost

Staged Pressurized Oxy-Combustion (SPOC) is a coal combustion power technology being developed by Washington University in St. Louis (WUSTL) that offers the potential to deliver low-carbon power at a reduced cost. Traditional oxy-combustion plants remove most of the nitrogen in air prior to combustion, thereby burning fuel in oxygen instead of air, producing a flue gas containing primarily CO 2 and water that allows relatively simple CO 2 capture at amounts > 90%. This technology relies on flue gas recycle (FGR) to reduce the peak temperature and radiation that would otherwise occur in a fuel/oxygen only flame. SPOC reduces the peak temperatures of combustion by utilizing two or more pressurized boiler modules connected in series to produce fuel staging; hence, only a portion of the fuel is combusted in any given furnace module. This means that the thermal energy released at each stage can be captured and removed from the gases prior to subsequent stages, when more fuel is introduced. This allows the SPOC process to operate with minimal FGR, avoiding the associated efficiency losses and additional costs. The SPOC process operates at an elevated gas-side pressure, reducing boiler size, enhancing heat transfer to achieve a compact boiler configuration as compared to an atmospheric-pressure boiler design, and allowing the recovery of the latent heat of the water from the flue gas at a temperature useful to the steam cycle. The resultant net efficiency of the system is up to 6 percentage points greater than traditional atmospheric-pressure oxy-combustion, representing a step-change improvement over first-generation oxy-combustion technologies. WUSTL has teamed up with the Electric Power Research Institute, Inc., American Air Liquide, Inc., Doosan Babcock Limited, and the U.S. Department of Energy to develop a practicable and workable boiler design concept associated with the application of SPOC. In particular, due to the staged nature of the heat release, the team has identified the potential for enhanced process flexibility for controlling power generation over a wider load range than is normally available to conventional coal-fired power plants. Improved flexibility is a critical need as coal power plants are being tasked with providing grid stability to balance the growth of intermittent renewables. Not only is flexible operation important, achieving efficient power generation at reduced loads will be key as the plants are likely to operate in this regime for significant periods of time as more renewables become available. At reduced loads, coal-fired steam generators typically face challenges in maintaining temperature control of the reheat steam and to a lesser extent the main steam due to the heat transfer characteristics of these circuits when flue gas quantities are lower. This results in inefficient operation, both in terms of the boiler efficiency and steam turbine heat rate. Since the SPOC process can be configured to deliver targeted heat input to each boiler circuit by adjusting the relative fuel firing rate of each stage, better temperature control can be achieved during dynamic conditions. Oxygen supply flexibility is also a key consideration for the overall flexibility of the SPOC process given the operating constraints of conventional air separation units. The boiler design concept assessment is being carried out to identify the maximum permissible compactness of the SPOC boiler heating surfaces. The assessment will determine the minimum overall height that will deliver appropriate tube operating metal temperatures at full load and achieve rated reheat steam temperatures at low operating loads, balanced against the needs of efficient coal combustion, and the resultant slagging and ash environments. The design parameters are being validated by combustion testing in the 100-kWth pressurized combustion test rig at WUSTL. Testing results and model validation that form the basis of a full-scale boiler design that will encompass both efficiency and flexibility improvements over conventional oxy-combustion processes will be presented.

20 FOSSIL-FUELED POWER PLANTS↗

Data Locality Enhancement of Dynamic Simulations for Exascale Computing (Final Report)

The development of modern processors exhibits two trends that complicate the optimizations of modern software. The first is the increasing sensitivity of processors' throughput to irregularities in computation. With more processors produced through a massive integration of simple cores, future systems will increasingly favor regular data-level parallel computations, but deviate from the needs of applications with complex patterns. Some evidences are already shown on Graphic Processing Units (GPU): Irregular data accesses (e.g., indirect references A[D[i]]) and conditional branches are limiting many GPU applications' performance at a level an order of magnitude lower than the peak of GPU. The second hardware trend is the growing gap between memory bandwidth and the aggregate speed—that is, the sum of all cores' computing power—of a Chip Multiprocessor (CMP). Despite the capped growth of the peak CPU speed, the aggregate speed of a CMP keeps increasing as more cores get into a single chip. It is expected that by 2018, node concurrency in an exascale system will increase by hundreds of times, whereas, memory bandwidth will expand by only 10 to 20 times. Consequently, data movement and storage is expected to consume more than 70% of the total system power. Bridging this gap is difficult; the complexities of modern CMP memory hierarchy make it even harder: Data cache becomes shared among computing units, and the sharing is often non-uniform—whether two computing units share a cache depends on their proximity and the level of the cache. On the recent IBM Power7 architecture, for instance, four hardware contexts (or SMT threads) in a core share the entire memory hierarchy, all cores in one chip share an on-chip L3 cache, and cores across chips share L3 and main memory through off-chip connections. These two trends complicate the translation of computing power into performance, especially for a program with either intensive data accesses or complex patterns in data accesses or control flow paths. Unfortunately, both attributes present and will persist in a class of important applications. For instance, many scientific simulations deal with a large volume of data. And meanwhile, as most real-world processes are non-uniform and evolving (e.g., the evolution of a galaxy or the process of a drug injection), both the computations and data accesses of these programs tend to be irregular and dynamically changing. Currently, the lack of support to these applications on modern CMP severely limits their performance. On GPU, as our recent study shows and other studies echo, performance enhancement of a factor of integers is possible when memory accesses or control flows are streamlined for a set of GPU applications. On multicore CPU, our studies show that traditional locality enhancement, for being oblivious to the new features of multicore memory hierarchy, may even cause large slowdown to data-intensive dynamic applications. The severity of the issues is expected to worsen as the two hardware trends continue. Some recent studies try to match software with the trends, but in a limited scope or manner. For irregularities on GPU, most studies focus on irregularities analyzable through static analysis (e.g., data accesses in regular loops). Dynamic irregularities are harder to address because the needed analysis and transformations typically have to happen at run time. Some other research resorts to hardware extensions, an actual adoption of which is unclear for the entailed space cost and complexity. For data locality, recent years have seen some exploitations of the new memory hierarchy on multicore for performance, but most of them are on process or thread scheduling, rather than program transformations. Our recent study reveals that program-level transformations may magnify the scheduling benefits by a factor of seven, concluding that program-level transformation should play a central role for data locality enhancement on modern CMP. But research in this direction has been sparse, and most have focused on data layout or cache performance modeling, rather than program transformations to match with the new memory hierarchy features. Overall, it is still an open question how to bridge the gap between dynamic computations and the two prominent properties of modern processors. The goal of this project is to develop a set of techniques and software tools to enhance the matching between memory accesses in dynamic simulations and the prominent features of modern and future CMP, alleviating the memory performance issues for petascale and exascale computing. This report summarizes the discoveries and products produced throughout this project. It includes free launch, a new software approach to overcoming the shortcomings of both methods; coherence-free multiview, an approach that allows multiple views of a single data object to co-exist on GPU memory during a GPU kernel execution; algorithmic optimizations to data analytics problems, especially those that involve lots of distance calculations. Supporting dynamic parallelism is important for GPU to benefit a broad range of applications. There are currently two fundamental ways for programs to exploit dynamic parallelism on GPU: a software-based approach with software-managed worklists, and a hardware-based approach through dynamic subkernel launches. Neither is satisfactory. The former is complicated to program and is often subject to some load imbalance; the latter suffers large runtime overhead. In this work, we propose free launch, a new software approach to overcoming the shortcomings of both methods. It allows programmers to use subkernel launches to express dynamic parallelism. It employs a novel compiler-based code transformation named subkernel launch removal to replace the subkernel launches with the reuse of parent threads. Coupled with an adaptive task assignment mechanism, the transformation reassigns the tasks in the subkernels to the parent threads with a good load balance. The technique requires no hardware extensions, immediately deployable on existing GPUs. It keeps the programming convenience of the subkernel launch-based approach while avoiding its large runtime overhead. Meanwhile, its superior load balancing makes it outperform manual worklist-based techniques by 3X on average. The work was published at the 48th Annual IEEE/ACM International Symposium on Microarchitecture (Micro'2015). A Graphic Processing Unit (GPU) system is typically equipped with many types of memory (e.g., global, constant, texture, shared, cache). Data placement determines what data are placed on which type of memory, essential for GPU memory performance. Prior optimizations of data placement always require a single view of a data object on memory, which limits the optimization effectiveness. In this work, we propose coherence-free multiview, an approach that allows multiple views of a single data object to co-exist on GPU memory during a GPU kernel execution. We demonstrate that under certain conditions, the multiple views can remain incoherent while facilitating enhanced data placement. We present a theorem and some compiler support to ensure the soundness of the usage of coherence-free multiview. We further develop reference-discerning data placement, a new way to enhance data placements on GPU. It enables more flexible data placements by using coherence-free multiview to leverage the slack in coherence requirement of some GPU programs. Experiments on three types of GPU systems show that, with less than 200KB space cost, the new data placement technique can provide a 1.6X average (up to 4.27X) speedup. In addition, we have examined algorithmic optimizations to data analytics problems, especially those that involve lots of distance calculations. Computing distances among data points is an essential part of many important algorithms in data analytics, graph analysis, and other domains. In each of these domains, developers have spent significant manual effort optimizing algorithms, often through novel applications of the triangle equality, in order to minimize the number of distance computations in the algorithms. In this work, we observe that many algorithms across these domains can be generalized as an instance of a generic distance-related abstraction. Based on this abstraction, we derive seven principles for correctly applying the triangular inequality to optimize distance-related algorithms. Guided by the findings, we develop Triangular OPtimizer (TOP), the first software framework that is able to automatically produce optimized algorithms that either matches or outperforms manually designed algorithms for solving distance-related problems. TOP achieves up to 237x speedups and 2.5X on average. The work has been published at the 32nd International Conference on Machine Learning (ICML'15) and the 41st International Conference on Very Large Data Bases (VLDB'15).

42 ENGINEERING↗

Gamma radiation effects on the dynamic fatigue measurements of glass discs

Circular specimens of low iron soda lime silicate glass were blasted with grit after having a circular notch etched into their centers. After separation into two groups, one group was exposed to gamma radiation. The fracture strengths of all samples were then tested by the biaxial technique, i.e., specimens were balanced on three balls and loaded in the center by a piston. The irradiated samples had received a 140,000 Gy dose from a Co-60 source. An enhanced interaction between the ambient moisture and the grit-blasted central notch was observed in the irradiated samples, which displayed accelerated corrosion.

Ananaba, T. O. J.↗

SSME Investment in Turbomachinery Inducer Impeller Design Tools and Methodology

Within the rocket engine industry, SSME turbomachines are the de facto standards of success with regard to meeting aggressive performance requirements under challenging operational environments. Over the Shuttle era, SSME has invested heavily in our national inducer impeller design infrastructure. While both low and high pressure turbopump failures/anomaly resolution efforts spurred some of these investments, the SSME program was a major benefactor of key areas of turbomachinery inducer-impeller research outside of flight manifest pressures. Over the past several decades, key turbopump internal environments have been interrogated via highly instrumented hot-fire and cold-flow testing. Likewise, SSME has sponsored the advancement of time accurate and cavitating inducer impeller computation fluid dynamics (CFD) tools. These investments together have led to a better understanding of the complex internal flow fields within aggressive high performing inducers and impellers. New design tools and methodologies have evolved which intend to provide confident blade designs which strike an appropriate balance between performance and self induced load management.

Zoladz, Thomas↗

NO-LESS: Near optimal curtailment strategy selection for net load balancing in micro grids

The integration of renewable energy sources such as solar PVs into micro grids has increased in recent years. However, the intermittent and unpredictable nature of renewable energy generation makes supply uncertain, in turn requiring continuous net load balancing to ensure grid stability. One technique to achieve net load balancing is curtailment, however, selection of the optimal set of curtailment strategies to achieve a targeted curtailment is NP-hard. State-of-the-art curtailment selection techniques approach this problem by either developing computationally expensive optimal curtailment strategies or reduce accuracy to generate sub-optimal solutions quickly. In this work, we develop NO-LESS: a Near OptimaL CurtailmEnt Strategy Selection algorithm. NO-LESS is a novel Fully Polynomial Time Approximation Scheme (FPTAS) which determines bounded (near optimal) curtailment strategies in a small amount of time while simultaneously satisfying practical constraints on strategy switching overhead and curtailment fairness. We perform both theoretical analysis and practical evaluation to show that NO-LESS is a scalable solution for performing net load balancing through curtailment strategy selection in Micro Grids.

14 SOLAR ENERGY↗

Algorithms for parallel flow solvers on message passing architectures

The purpose of this project has been to identify and test suitable technologies for implementation of fluid flow solvers -- possibly coupled with structures and heat equation solvers -- on MIMD parallel computers. In the course of this investigation much attention has been paid to efficient domain decomposition strategies for ADI-type algorithms. Multi-partitioning derives its efficiency from the assignment of several blocks of grid points to each processor in the parallel computer. A coarse-grain parallelism is obtained, and a near-perfect load balance results. In uni-partitioning every processor receives responsibility for exactly one block of grid points instead of several. This necessitates fine-grain pipelined program execution in order to obtain a reasonable load balance. Although fine-grain parallelism is less desirable on many systems, especially high-latency networks of workstations, uni-partition methods are still in wide use in production codes for flow problems. Consequently, it remains important to achieve good efficiency with this technique that has essentially been superseded by multi-partitioning for parallel ADI-type algorithms. Another reason for the concentration on improving the performance of pipeline methods is their applicability in other types of flow solver kernels with stronger implied data dependence. Analytical expressions can be derived for the size of the dynamic load imbalance incurred in traditional pipelines. From these it can be determined what is the optimal first-processor retardation that leads to the shortest total completion time for the pipeline process. Theoretical predictions of pipeline performance with and without optimization match experimental observations on the iPSC/860 very well. Analysis of pipeline performance also highlights the effect of uncareful grid partitioning in flow solvers that employ pipeline algorithms. If grid blocks at boundaries are not at least as large in the wall-normal direction as those immediately adjacent to them, then the first processor in the pipeline will receive a computational load that is less than that of subsequent processors, magnifying the pipeline slowdown effect. Extra compensation is needed for grid boundary effects, even if all grid blocks are equally sized.

Vanderwijngaart, Rob F.↗

Polymer Chain Diffusion in All-Polymer Nanocomposites: Confinement vs Chain Acceleration

All-polymer nanocomposites, in which soft, polymer-based nanoparticles are dispersed in the polymer matrix, have received great interest lately due to their potential use in a range of applications, including drug delivery and self-healing materials. However, the impact of this new class of nanoparticles on the dynamics of a linear polymer matrix in an all-polymer nanocomposite is still largely unknown. In this work, we report that the addition of polystyrene soft nanoparticles accelerates the diffusion of high molecular weight linear PS chains over a range of nanoparticle loadings. Our results show that at nanoparticle loadings below 1%, the diffusion of the linear matrix increases, presumably via a constraint release mechanism. At loadings above 1%, the increase in diffusion is mitigated by confinement effects of the nanoparticles. Thus, the response of these all-polymer nanocomposites is dominated by the balance of entropic confinement of the chain, which slows diffusion and a constraint release mechanism that speeds up the diffusion. However, the diffusion of the linear chain in the all-polymer nanocomposite is faster than that of the same chain in the melt at all loadings, diverging from the behavior of most nanocomposites with hard, impenetrable nanoparticles. Thus, the mechanism that accelerates the chain diffusion dominates in these systems. This behavior is unusual and fundamentally different than what has been reported for nanocomposites with hard inorganic nanoparticles, indicating that new perspectives are needed for these materials where the control of loading can either accelerate or decelerate the dynamics of the matrix in a distinctive manner.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Efficient Machine Learning Approach for Optimizing Scientific Computing Applications on Emerging HPC Architectures

Efficient parallel implementations of scientific applications on multi-core CPUs with accelerators such as GPUs and Xeon Phis is challenging. This requires - exploiting the data parallel architecture of the accelerator along with the vector pipelines of modern x86 CPU architectures, load balancing, and efficient memory transfer between different devices. It is relatively easy to meet these requirements for highly structured scientific applications. In contrast, a number of scientific and engineering applications are unstructured. Getting performance on accelerators for these applications is extremely challenging because many of these applications employ irregular algorithms which exhibit data-dependent control-ow and irregular memory accesses. Furthermore, these applications are often iterative with dependency between steps, and thus making it hard to parallelize across steps. As a result, parallelism in these applications is often limited to a single step. Numerical simulation of charged particles beam dynamics is one such application where the distribution of work and memory access pattern at each time step is irregular. Applications with these properties tend to present significant branch and memory divergence, load imbalance between different processor cores, and poor compute and memory utilization. Prior research on parallelizing such irregular applications have been focused around optimizing the irregular, data-dependent memory accesses and control-ow during a single step of the application independent of the other steps, with the assumption that these patterns are completely unpredictable. We observed that the structure of computation leading to control-ow divergence and irregular memory accesses in one step is similar to that in the next step. It is possible to predict this structure in the current step by observing the computation structure of previous steps. In this dissertation, we present novel machine learning based optimization techniques to address the parallel implementation challenges of such irregular applications on different HPC architectures. In particular, we use supervised learning to predict the computation structure and use it to address the control-ow and memory access irregularities in the parallel implementation of such applications on GPUs, Xeon Phis, and heterogeneous architectures composed of multi-core CPUs with GPUs or Xeon Phis. We use numerical simulation of charged particles beam dynamics simulation as a motivating example throughout the dissertation to present our new approach, though they should be equally applicable to a wide range of irregular applications. The machine learning approach presented here use predictive analytics and forecasting techniques to adaptively model and track the irregular memory access pattern at each time step of the simulation to anticipate the future memory access pattern. Access pattern forecasts can then be used to formulate optimization decisions during application execution which improves the performance of the application at a future time step based on the observations from earlier time steps. In heterogeneous architectures, forecasts can also be used to improve the memory performance and resource utilization of all the processing units to deliver a good aggregate performance. We used these optimization techniques and anticipation strategy to design a cache-aware, memory efficient parallel algorithm to address the irregularities in the parallel implementation of charged particles beam dynamics simulation on different HPC architectures. Experimental result using a diverse mix of HPC architectures shows that our approach in using anticipation strategy is effective in maximizing data reuse, ensuring workload balance, minimizing branch and memory divergence, and in improving resource utilization.

Arumugam, Kamesh↗