Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel systems”

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 523 records · Page 29

Scalability under a Power Bound using the GREMLINs Framework

With the move towards exascale, system and software developers will have to deal with issues of extreme parallelism. The system properties affected most by the increase in node and core count are the shared resources on node and across the system. The increase in parallelism leads to reduced memory and bandwidth when regarding individual cores. Since power is a limiting factor for supercomputers, and power is not fully utilized in current systems, overprovisioning compute resources is a viable approach to maximized power utilization. To maximize system performance in regard to these changing conditions, it is necessary to understand how resource restrictions impact performance and system behavior. For the purpose of understanding anticipated system properties the GREMLINs Framework was developed. The framework gives the opportunity to add power restrictions, hinder memory properties and introduce faults to study resilience, among others. These features give the opportunity to use current petascale technology to study problems system designers and software developers will have to face when moving towards exascale and beyond. This work describes the initial release of the GREMLINs Framework, developed for this work, and shows how it can be used to study the scaling behavior of proxy applications. These proxy applications represent a selection of HPC workloads important to the scientific community. The proxy applications studied are AMG2013, an algebraic multi-grid linear system solver, CoMD, a classical molecular dynamics proxy application and NEKBONE, an application that uses a high order spectral element method to solve the Navier-Stokes equations. The main interest of these studies lies in analysis regarding their power behavior at scale under a power bound. These findings show how the GREMLINs Framework can help systems and software designers to attain better application performance and can also be used as basis for CPU power balancing tools to use power more efficiently.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Enabling parallel simulation of large-scale HPC network systems

Here, with the increasing complexity of today’s high-performance computing (HPC) architectures, simulation has become an indispensable tool for exploring the design space of HPC systems—in particular, networks. In order to make effective design decisions, simulations of these systems must possess the following properties: (1) have high accuracy and fidelity, (2) produce results in a timely manner, and (3) be able to analyze a broad range of network workloads. Most state-of-the-art HPC network simulation frameworks, however, are constrained in one or more of these areas. In this work, we present a simulation framework for modeling two important classes of networks used in today’s IBM and Cray supercomputers: torus and dragonfly networks. We use the Co-Design of Multi-layer Exascale Storage Architecture (CODES) simulation framework to simulate these network topologies at a flit-level detail using the Rensselaer Optimistic Simulation System (ROSS) for parallel discrete-event simulation. Our simulation framework meets all the requirements of a practical network simulation and can assist network designers in design space exploration. First, it uses validated and detailed flit-level network models to provide an accurate and high-fidelity network simulation. Second, instead of relying on serial time-stepped or traditional conservative discrete-event simulations that limit simulation scalability and efficiency, we use the optimistic event-scheduling capability of ROSS to achieve efficient and scalable HPC network simulations on today’s high-performance cluster systems. Third, our models give network designers a choice in simulating a broad range of network workloads, including HPC application workloads using detailed network traces, an ability that is rarely offered in parallel with high-fidelity network simulations

97 MATHEMATICS AND COMPUTING↗

T-FSM: A Scalable Distributed Task-Based System for Frequent Subgraph Pattern Mining from a Big Graph

Finding frequent subgraph patterns in a big graph is an important problem with many applications such as classifying chemical compounds and building indexes to speed up graph queries. Since this problem is NP-hard, some recent parallel and distributed systems have been developed to accelerate the mining. However, they often have a huge memory cost, very long running time, suboptimal load balancing, poor scale-out capability, and possibly inaccurate results. In this article, we propose an efficient system called T-FSM for parallel mining of frequent subgraph patterns in a big graph. T-FSM supports a new anti-monotonic frequentness measure called Fraction-Score, which is more accurate than the widely used MNI measure. The execution engine of T-FSM supports both intra-machine parallelism and inter-machine parallelism. For intra-machine parallelism, T-FSM adopts a novel task-based execution model to ensure high multithreading concurrency, bounded memory consumption, and effective load balancing. For inter-machine parallelism, T-FSM ensures good scale-out performance with a lightweight pattern rebalancing approach that reduces workload skewness of pattern evaluations among machines. To avoid recomputing the contexts for migrated patterns, we design a novel context cache table to support concurrent and asynchronous requesting and caching of remote context data, which can timely evict and garbage collect used pattern contexts that are no longer needed to keep memory consumption bounded. Extensive experiments show that T-FSM is orders of magnitude faster than existing state-of-the-art parallel systems (more than 10×, 51×, 131×, 55× speedup over ScaleMine, DistGraph, Pangolin and Peregrine, respectively) and distributed systems (more than 42× and 88× over ScaleMine and DistGraph, respectively) for frequent subgraph pattern mining, and it scales out satisfactorily to 512 CPU cores on the Polaris supercomputer at Argonne National Laboratory.

97 MATHEMATICS AND COMPUTING↗

Parallel compression of data chunks of a shared data object using a log-structured file system

Techniques are provided for parallel compression of data chunks being written to a shared object. A client executing on a compute node or a burst buffer node in a parallel computing system stores a data chunk generated by the parallel computing system to a shared data object on a storage node by compressing the data chunk; and providing the data compressed data chunk to the storage node that stores the shared object. The client and storage node may employ Log-Structured File techniques. The compressed data chunk can be de-compressed by the client when the data chunk is read. A storage node stores a data chunk as part of a shared object by receiving a compressed version of the data chunk from a compute node; and storing the compressed version of the data chunk to the shared data object on the storage node.

97 MATHEMATICS AND COMPUTING↗

Operational characteristics of a parallel jet MILD combustion burner system

This study describes the performance and stability characteristics of a parallel jet MILD (Moderate or Intense Low-oxygen Dilution) combustion burner system in a laboratory-scale furnace, in which the reactants and exhaust ports are all mounted on the same wall. Thermal field measurements are presented for cases with and without combustion air preheat, in addition to global temperature and emission measurements for a range of equivalence ratio, heat extraction, air preheat and fuel dilution levels. The present furnace/burner configuration proved to operate without the need for external air preheating, and achieved a high degree of temperature uniformity. Based on an analysis of the temperature distribution and emissions, PSR model predictions, and equilibrium calculations, the CO formation was found to be related to the mixing patterns and furnace temperature rather than reaction quenching by the heat exchanger. The critical equivalence ratio, or excess air level, which maintains low CO emissions is reported for different heat exchanger positions, and an optimum operating condition is identified. Results of CO and NO{sub x} emissions, together with visual observations and a simplified two-dimensional analysis of the furnace aerodynamics, demonstrate that fuel jet momentum controls the stability of this multiple jet system. A stability diagram showing the threshold for stable operation is reported, which is not explained by previous stability criteria. (author)

42 ENGINEERING↗

Sisal

Sisal (Streams and Iterations in a Single Assignment Language) was a general-purpose applicative language developed for shared-memory and vector supercomputer systems. It provided an hierarchical intermediate form, parallel runtime system, optimizing compiler, and programming environment. The language was strongly typed. It supported both array and stream data structures, and had both iterative and parallel loop constructs.

Feo, John T↗

Description and operating performance of a parallel-rail electric-arc system with helium driver gas for the Langley 6-inch expansion tube

A parallel-rail arc-discharge system to heat and pressurize the initial helium driver gas of the Langley 6-inch expansion tube is described. This system was designed for a 2.44-m-long driver vessel rated at 138 MPa, with a distance between rails of 20.3 cm. Electric energy was obtained from a capacitor storage system rated at 12,000 V with a maximum energy of 5 MJ. Tests were performed over a range of energy from 1.74 MJ to the maximum value. The operating experience and system performance are discussed, along with results from a limited number of expansion-tube tests with air and carbon dioxide as test gases.

Moore, J. A.↗

Small File Aggregation with PLFS [Slides]

Today’s computational science demands have resulted in ever larger parallel computers, and storage systems have grown to match these demands. Parallel file systems used in this environment are increasingly specialized to extract the highest possible performance for large I/O operations, at the expense of other potential workloads. While some applications have adapted to I/O best practices and can obtain good performance on these systems, the natural I/O patterns of many applications result in the generation of a huge number of small files, the creation of which is poorly served by current parallel file systems at very large scale. This paper describes a new technique for optimizing small file access in parallel file systems for these very large scale systems. The idea is to use a virtual parallel log-structure file system on the compute nodes in order to aggregate large numbers of small files in compute node memory and then stream their data sequentially to a much smaller number of physical files on an underlying parallel file system. The technique is implemented and evaluated using PLFS as the aggregating middleware. We evaluate our system with micro-benchmarks on a local OSX filesystem and with an MPI extension of the standard Postmark to provide results at scale on both Lustre and PanFS parallel filesystems. We observe as much as a 33x improvement in small file create rates on a single host, and 30x improvement in small file write rates, compared to a baseline Lustre configuration on a leadership computing platform using 16,384 cores and achieve an unprecedented create rate of 200 million files per second.

97 MATHEMATICS AND COMPUTING↗

The Scalable Checkpoint/Restart Library

The Scalable Checkpoint/Restart (SCR) library provides an interface that codes may use to worite our and read in application-level checkpoints in a scalable fashion. In the current implementation, checkpoint files are cached in local storage (hard disk or RAM disk) on the compute nodes. This technique provides scalable aggregate bandwidth and uses storage resources that are fully dedicated to the job. This approach addresses the two common drawbacks of checkpointing a large-scale application to a shared parallel file system, namely, limited bandwidth and file system contention. In fact, on current platforms, SCR scales linearly with the number of compute nodes. It has been benchmarked as high as 720GB/s on 1094 nodes of Atlas, which is nearly two orders of magnitude faster thanthe parallel file system.

ECP↗

Parallel computing for power system climate resiliency: Solving a large-scale stochastic capacity expansion problem with mpi-sppy

Here we propose a nodal stochastic generation and transmission expansion planning model that incorporates the output from high-resolution global climate models through load and generation availability scenarios. We implement our model in Pyomo and perform computational studies on a realistically-sized test case of the California electric grid in a high performance computing environment. We propose model reformulations and algorithm tuning to efficiently solve this large problem using a variant of the Progressive Hedging Algorithm. We utilize the parallelization capabilities and overall versatility of mpi-sppy, exploiting its hub-and-spoke architecture to concurrently obtain inner and outer bounds on an optimal expansion plan. Initial results show that instances with 360 representative days on a system with over 8,000 buses can be solved to within 5% of optimality in under 4 h of wall clock time, a first step towards solving a large-scale power system expansion planning problem across a wide range of climate-informed operational scenarios.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Improving the Availability of Supercomputer Job Input Data Using Temporal Replication

Storage systems in supercomputers are a major reason for service interruptions. RAID solutions alone cannot provide sufficient protection as (1) growing average disk recovery times make RAID groups increasingly vulnerable to disk failures during reconstruction, and (2) RAID does not help with higher-level faults such failed I/O nodes. This paper presents a complementary approach based on the observation that files in the supercomputer scratch space are typically accessed by batch jobs whose execution can be anticipated. Therefore, we propose to transparently, selectively, and temporarily replicate 'active' job input data by coordinating the parallel file system with the batch job scheduler. We have implemented the temporal replication scheme in the popular Lustre parallel file system and evaluated it with real-cluster experiments. Our results show that the scheme allows for fast online data reconstruction, with a reasonably low overall space and I/O bandwidth overhead.

99 GENERAL AND MISCELLANEOUS↗

Parallel Implementation of Power System Dynamic Simulation

Dynamic simulation of power system transient stability is important for planning, monitoring, operation, and control of electrical power systems. However, modeling the system dynamics and network involves the computationally intensive time-domain solution of numerous differential and algebraic equations (DAE). This results in a transient stability implementation that may not maintain the real-time constraints of an online security assessment. This paper presents a parallel implementation of the dynamic simulation on a high-performance computing (HPC) platform using parallel simulation algorithms and computation architectures. It enables the simulation to run even faster than real time, enabling the “look-ahead” capability of upcoming stability problems in the power grid.

dynamic simulation↗

HPC Global File System Performance Analysis Using A Scientific-Application Derived Benchmark

With the exponential growth of high-fidelity sensor and simulated data, the scientific community is increasingly reliant on ultrascale HPC resources to handle its data analysis requirements. However, to use such extreme computing power effectively, the I/O components must be designed in a balanced fashion, as any architectural bottleneck will quickly render the platform intolerably inefficient. To understand I/O performance of data-intensive applications in realistic computational settings, we develop a lightweight, portable benchmark called MADbench2, which is derived directly from a large-scale Cosmic Microwave Background (CMB) data analysis package. Our study represents one of the most comprehensive I/O analyses of modern parallel file systems, examining a broad range of system architectures and configurations, including Lustre on the Cray XT3, XT4, and Intel Itanium2 clusters; GPFS on IBM Power5 and AMD Opteron platforms; a BlueGene/P installation using GPFS and PVFS2 file systems; and CXFS on the SGI Altix\-3700. We present extensive synchronous I/O performance data comparing a number of key parameters including concurrency, POSIX- versus MPI-IO, and unique-versus shared-file accesses, using both the default environment as well as highly-tuned I/O parameters. Finally, we explore the potential of asynchronous I/O and show that only the two of the nine evaluated systems benefited from MPI-2's asynchronous MPI-IO. On those systems, experimental results indicate that the computational intensity required to hide I/O effectively is already close to the practical limit of BLAS3 calculations. Overall, our study quantifies vast differences in performance and functionality of parallel file systems across state-of-the-art platforms -- showing I/O rates that vary up to 75x on the examined architectures -- while providing system designers and computational scientists a lightweight tool for conducting further analysis.

97 MATHEMATICS AND COMPUTING↗

SDS: A Framework for Scientific Data Services

Large-scale scientific applications typically write their data to parallel file systems with organizations designed to achieve fast write speeds. Analysis tasks frequently read the data in a pattern that is different from the write pattern, and therefore experience poor I/O performance. In this paper, we introduce a prototype framework for bridging the performance gap between write and read stages of data access from parallel file systems. We call this framework Scientific Data Services, or SDS for short. This initial implementation of SDS focuses on reorganizing previously written files into data layouts that benefit read patterns, and transparently directs read calls to the reorganized data. SDS follows a client-server architecture. The SDS Server manages partial or full replicas of reorganized datasets and serves SDS Clients' requests for data. The current version of the SDS client library supports HDF5 programming interface for reading data. The client library intercepts HDF5 calls and transparently redirects them to the reorganized data. The SDS client library also provides a querying interface for reading part of the data based on user-specified selective criteria. We describe the design and implementation of the SDS client-server architecture, and evaluate the response time of the SDS Server and the performance benefits of SDS.

59 BASIC BIOLOGICAL SCIENCES↗

Biological neural networks as model systems for designing future parallel processing computers

One of the more interesting debates of the present day centers on whether human intelligence can be simulated by computer. The author works under the premise that neurons individually are not smart at all. Rather, they are physical units which are impinged upon continuously by other matter that influences the direction of voltage shifts across the units membranes. It is only the action of a great many neurons, billions in the case of the human nervous system, that intelligent behavior emerges. What is required to understand even the simplest neural system is painstaking analysis, bit by bit, of the architecture and the physiological functioning of its various parts. The biological neural network studied, the vestibular utricular and saccular maculas of the inner ear, are among the most simple of the mammalian neural networks to understand and model. While there is still a long way to go to understand even this most simple neural network in sufficient detail for extrapolation to computers and robots, a start was made. Moreover, the insights obtained and the technologies developed help advance the understanding of the more complex neural networks that underlie human intelligence.

Ross, Muriel D.↗

End-to-end I/O portfolio for the summit supercomputing ecosystem

The I/O subsystem for the Summit supercomputer, No. 1 on the Top500 list, and its ecosystem of analysis platforms is composed of two distinct layers, namely the in-system layer and the center-wide parallel file system layer (PFS), Spider 3. The in-system layer uses node-local SSDs and provides 26.7 TB/s for reads, 9.7 TB/s for writes, and 4.6 billion IOPS to Summit. The Spider 3 PFS layer uses IBM's Spectrum Scale™ and provides 2.5 TB/s and 2.6 million IOPS to Summit and other systems. While deploying them as two distinct layers was operationally efficient, it also presented usability challenges in terms of multiple mount points and lack of transparency in data movement. To address these challenges, we have developed novel end-to-end I/O solutions for the concerted use of the two storage layers. We present the I/O subsystem architecture, the end-to-end I/O solution space, their design considerations and our deployment experience.

Oral, Sarp↗

Polarized 3He gas circulating technologies for neutron analyzers

We outline our project to develop a circulating polarized helium-3 system for developing of large, quasi-continuously operating neutron analyzers. The project consisted of four areas: 1) Development of robust external cavity narrowed diode laser output with spectral line width < 0.17 nm and power of 2000 W. 2) Development of large glass polarizing cells using cell surface treatments to obtain long relaxation lifetimes. 3) Refinements of the circulation system with an emphasis on gas purification and materials testing. 4) Design/fabrication of a new polarizer system. 5) Preliminary testing of the new polarizer. 1. Developed Robust High-Power Narrowed Laser The optical configuration of the laser was discussed in the proposal and will be reviewed in the body of this report. The external cavity is configured to mutually lock the wavelength of five 10-bar laser stacks. All the logistical milestones were been met and critical subsystems- laser stack manifold and power divider, external laser cavity, and output telescope- were assembled and tested at low power. Each individual bar is narrowed to ~0.05 nm; when combined the laser has a cumulative spectral width of 0.17 nm across the entire beam due to variations of the bars central wavelength by +/- 0.1 nm, which is similar to that of Volume Bragg Grating narrowed laser bars. This configuration eliminates the free-running “pedestal” that occurs in other external cavity diode lasers. The full-scale laser was completed in 2016 and was used in both the older and newer helium polarizers. This laser was operated at 75% power for periods of up to 8 hours. Once installed, the spectrum became slightly broader (~.25 nm) at full power; this is likely due to very slight misalignments that occurred during handling. 2. Developed the processes to create uniform sintered sol-gel coatings. Our work on cell development comprised: 1) Production of large GE180 cells and explore different means of cell preparation, and 2) Development of apply sol-gel coatings to the interior of both borosilicate and aluminosilicate cells. We applied six sol-gel coatings. By modifying the mixture and developing procedures to drain and dry the cell, we produced visually uniform coatings on the interior of the cells. We now have perfected that process as described below in our report. We were able to accelerate the testing of cells using an ex situ method that avoids installing each cell into a polarizer. In the project’s last year, we conducted 38 external tests of 8 different cells. We also installed two sol-gel coated cells in our polarizers. We created cell with long ex situ relaxation lifetimes, one of which exceeded 40 hours. However, when installed in the polarizer the measured lifetime is 8 hours or less. 3. Demonstrated cycling of polarized gas and ex situ cell testing We are now cycling polarized gas from the polarizer to glass vessels and back. This has allowed us, for the first time, to make ex situ T1 measurements of polarizing cells without installing them into the polarizer itself. This has greatly improved our productivity in producing cells and evaluating our cell preparation processes. We continued development of the gas handling system in parallel with fabricating new polarizer. The integrated system was tested by the end of 2016. We now regularly cycle gas into and out of the polarizer. 4. Completed new polarizer infrastructure and control systems. We completed the new polarizer infrastructure in November 2016. The polarizer subsystems are 1) the frame, 2) the oil flow system, 3) the gas handling system, 4) the pressure vessel, with embedded solenoid, 5) cell mounting hardware with heat spreaders, and 6) electrical power and instrumentation. 5. Carried out initial tests of polarizer. We completed initial testing of the polarizer in April and May of 2017. These tests were carried out for periods up to 6 hours with laser power between 750 and 1300 Watts. The laser performed well and the polarization with asymptotic to 45 percent, which was below expectations. This low value resulted from a stationary thermal inversion in the cell that caused most of the laser power to be absorbed near the laser inlet window and deprived the lower portions of the cell of pumping laser light. Possible solutions to this problem include enhanced cooling of the cell near the laser entry and slight detuning of the laser. 6. Ongoing work. Our polarizer development efforts are ongoing to pursue our interest in neutron analyzers, nuclear targets, and providing helium for medical imaging. Current tests in the pipeline include: 1. Testing cooling enhancements to improve laser penetration of spectrally narrow lasers; 2. Testing of a cell with isolation valves that minimizes diffusive contact with gas handling hardware during polarization; 3. Testing of smaller hybrid cells with reduced alkali loading; 4. Producing polarized helium-3 for MRI imaging at the University of Missouri.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Reliable inverter systems

Base driver with common-load-current feedback protects paralleled inverter systems from open or short circuits. Circuit eliminates total system oscillation that can occur in conventional inverters because of open circuit in primary transformer winding. Common feedback signal produced by functioning modules forces operating frequency of failed module to coincide with clock drive so module resumes normal operating frequency in spite of open circuit.

Nagano, S.↗