Search

With the increasing number of geographically distributed scientific collaborations and the scale of the data size growth, it has become more challenging for users to achieve the best possible network performance on a shared network. We have developed a forecast model to predict expected bandwidth utilization for high-bandwidth wide area network. The forecast model can improve the efficiency of resource utilization and scheduling data movements on high-bandwidth network to accommodate ever increasing data volume for large-scale scientific data applications. Univariate model is developed with STL and ARIMA on SNMP path utilization data. Compared with traditional approach such as Box-Jenkins methodology, our forecast model reduces computation time by 83.2percent. It also shows resilience against abrupt network usage change. The accuracy of the forecast model is within the standard deviation of the monitored measurements.

With the tight power budgets in modern large-scale chips and the unpredictability of application traffic, on-chip network designers are faced with the dilemma of designing for worst- case bandwidth demands and incurring high static power overheads, or designing for an average traffic pattern and risk degrading performance. This paper proposes adaptive bandwidth networks (ABNs) which divide channels and switches into lanes such that the network provides just the bandwidth necessary in each hop. ABNs also activate input virtual channels (VCs) individually and take advantage of drowsy SRAM cells to eliminate false VC activations. In addition, ABNs readily apply to silicon defect tolerance with just the extra cost for detecting faults. For application traffic, ABNs reduce total power consumption by an average of 45percent with comparable performance compared to single-lane power-gated networks, and 33percent compared to multi-network designs.

VPIN (Volume synchronized Probability of Informed trading) is a leading indicator of liquidity-induced volatility. It is best known for having produced a signal more than hours before the Flash Crash of 2010. On that day, the market saw the biggest one-day point decline in the Dow Jones Industrial Average, which culminated to the market value of $1 trillion disappearing, but only to recover those losses twenty minutes later (Lauricella 2010).The computation of VPIN requires the user to set up a handful of free parameters. The values of these parameters significantly affect the effectiveness of VPIN as measured by the false positive rate (FPR). An earlier publication reported that a brute-force search of simple parameter combinations yielded a number of parameter combinations with FPR of 7percent. This work is a systematic attempt to find an optimal parameter set using an optimization package, NOMAD (Nonlinear Optimization by Mesh Adaptive Direct Search) by Audet, le digabel, and tribes (2009) and le digabel (2011). We have implemented a number of techniques to reduce the computation time with NOMAD. Tests show that we can reduce the FPR to only 2percent.To better understand the parameter choices, we have conducted a series of sensitivity analysis via uncertainty quantification on the parameter spaces using UQTK (Uncertainty Quantification Toolkit). Results have shown dominance of 2 parameters in the computation of FPR. Using the outputs from NOMAD optimization and sensitivity analysis, We recommend A range of values for each of the free parameters that perform well on a large set of futures trading records.

The raw computational power of GPU accelerators enables fast denoising of 3D MR images using bilateral filtering, anisotropic diffusion, and non-local means. In practice, applying these filtering operations requires setting multiple parameters. This study was designed to provide better guidance to practitioners for choosing the most appropriate parameters by answering two questions: what parameters yield the best denoising results in practice? And what tuning is necessary to achieve optimal performance on a modern GPU? To answer the first question, we use two different metrics, mean squared error (MSE) and mean structural similarity (MSSIM), to compare denoising quality against a reference image. Surprisingly, the best improvement in structural similarity with the bilateral filter is achieved with a small stencil size that lies within the range of real-time execution on an NVIDIA Tesla M2050 GPU. Moreover, inappropriate choices for parameters, especially scalingparameters, can yield very poor denoising performance. To answer the second question, we perform an autotuning study to empirically determine optimal memory tiling on the GPU. The variation in these results suggests that such tuning is an essential step in achieving real-time performance. These results have important implications for the real-time application of denoising to MR images in clinical settings that require fast turn-around times.

This document provides an overview of the benchmark ? HPGMG ? for ranking large scale general purpose computers for use on the Top500 list [8]. We provide a rationale for the need for a replacement for the current metric HPL, some background of the Top500 list and the challenges of developing such a metric; we discuss our design philosophy and methodology, and an overview of the specification of the benchmark. The primary documentation with maintained details on the specification can be found at hpgmg.org and the Wiki and benchmark code itself can be found in the repository https://bitbucket.org/hpgmg/hpgmg.

The Energy Sciences Network (ESnet) is the primary provider of network connectivity for the U.S. Department of Energy (DOE) Office of Science (SC), the single largest supporter of basic research in the physical sciences in the United States. In support of SC programs, ESnet regularly updates and refreshes its understanding of the networking requirements needed by instruments, facilities, scientists, and science programs that it serves. This focus has helped ESnet to be a highly successful enabler of scientific discovery for over 25 years.In August 2013, ESnet and the DOE SC Offices of High Energy Physics (HEP) and Nuclear Physics (NP) organized a review to characterize the networking requirements of the programs funded by the HEP and NP program offices.Several key findings resulted from the review. Among them:1. The Large Hadron Collider?s ATLAS (A Toroidal LHC Apparatus) and CMS (Compact Muon Solenoid) experiments are adopting remote input/output (I/O) as a core component of their data analysis infrastructure. This will significantly increase their demands on the network from both a reliability perspective and a performance perspective.2. The Large Hadron Collider (LHC) experiments (particularly ATLAS and CMS) are working to integrate network awareness into the workflow systems that manage the large number of daily analysis jobs (1 million analysis jobs per day for ATLAS), which are an integral part of the experiments. Collaboration with networking organizations such as ESnet, and the consumption of performance data (e.g., from perfSONAR [PERformance Service Oriented Network monitoring Architecture]) are critical to the success of these efforts.3. The international aspects of HEP and NP collaborations continue to expand. This includes the LHC experiments, the Relativistic Heavy Ion Collider (RHIC) experiments, the Belle II Collaboration, the Large Synoptic Survey Telescope (LSST), and others. The international nature of these collaborations makes them heavily reliant on transoceanic connectivity, which is subject to longer term service disruptions than terrestrial connectivity. The network engineering aspects of undersea connectivity will continue to be a significant part of the planning, deployment, and operation of the data analysis infrastructure for HEP and NP experiments for the foreseeable future. Given their critical dependency on networking services, the experiments have expressed the need for tight integration (both technically and operationally) of the domestic and the transoceanic parts of the network infrastructure that supports the experiments.4. The datasets associated with simulations continue to increase in size, and the need to move these datasets between analysis centers is placing ever-increasing demands on networks and on data management systems at the supercomputing centers. In addition, there is a need to harmonize cybersecurity practice with the data transfer performance requirements of the science.This report expands on these points, and addresses others as well. The report contains a findings section in addition to the text of the case studies discussed during the review.

We present a numerical integration scheme for evaluating theconvolution of a Green's function with a screened Coulomb potentialon the real axis in the GW approximation of the self energy.Our scheme takes the zero broadening limit in Green's functionfirst, replaces the numerator of the integrand with a piecewisepolynomial approximation, and performs principal value integrationon subintervals analytically. We give the error bound of ournumerical integration scheme and show by numerical examples thatit is more reliable and accurate than the standard quadrature rulessuch as the composite trapezoidal rule. We also discuss the benefit of usingdifferent self energy expressions to perform the numerical convolutionat different frequencies.

The National Energy Research Scientific Computing Center (NERSC) is the primary computing center for the DOE Office of Science, serving approximately 4,500 users working on some 650 projects that involve nearly 600 codes in a wide variety of scientific disciplines. In March 2013, NERSC, DOE?s Office of Advanced Scientific Computing Research (ASCR) and DOE?s Office of Fusion Energy Sciences (FES) held a review to characterize High Performance Computing (HPC) and storage requirements for FES research through 2017. This report is the result.

We demonstrate the global convergence of the alternating projection algorithm to a unique solution up to a global phase factor. Additionally, for the ptychographic imaging problem, we discuss phase synchronization and connection graph Laplacian, and show how to construct an accurate initial guess to accelerate convergence speed to handle the big imaging data in the coming new light source era.

UPC is an explicitly parallel extension to the ISO C 99 Standard. UPC follows the partitioned global address space programming model. This document is the formal specification for the UPC language and library syntax and semantics, and supersedes prior specification version 1.2 (LBNL-59208).

The ever-increasing scale of scientific data has become a significant challenge for researchers that rely on networks to interact with remote computing systems and transfer results to collaborators worldwide. Despite the availability of high-capacity connections, scientists struggle with inadequate cyberinfrastructure that cripples data transfer performance, and impedes scientific progress. The Science DMZ paradigm comprises a proven set of network design patterns that collectively address these problems for scientists. We explain the Science DMZ model, including network architecture, system configuration, cybersecurity, and performance tools, that creates an optimized network environment for science. We describe use cases from universities, supercomputing centers and research laboratories, highlighting the effectiveness of the Science DMZ model in diverse operational settings. In all, the Science DMZ model is a solid platform that supports any science workflow, and flexibly accommodates emerging network technologies. As a result, the Science DMZ vastly improves collaboration, accelerating scientific discovery.

Plasma-based particle accelerators can produce and sustain thousands of times stronger acceleration fields than conventional particle accelerators, providing a potential solution to the problem of the growing size and cost of conventional particle accelerators. To facilitate scientific knowledge discovery from the ever growing collections of accelerator simulation data generated by accelerator physicists to investigate next-generation plasma-based particle accelerator designs, we describe a novel approach for automatic detection and classification of particle beams and beam substructures due to temporal differences in the acceleration process, here called acceleration features. The automatic feature detection in combination with a novel visualization tool for fast, intuitive, query-based exploration of acceleration features enables an effective top-down data exploration process, starting from a high-level, feature-based view down to the level of individual particles. We describe the application of our analysis in practice to analyze simulations of single pulse and dual and triple colliding pulse accelerator designs, and to study the formation and evolution of particle beams, to compare substructures of a beam and to investigate transverse particle loss.

Particle advection is an important vector field visualization technique that is difficult to apply to very large data setsin a distributed setting due to scalability limitations in existing algorithms. In this paper, we report on several experiments usingwork requesting dynamic scheduling which achieves balanced work distribution on arbitrary problems with minimal communicationoverhead. We present a corresponding prototype implementation, provide and analyze benchmark results, and compare our resultsto an existing algorithm.

Electricity is a vital part of our daily life; therefore it is important to avoid irregularities such as the California Electricity Crisis of 2000 and 2001. In this work, we seek to predict anomalies using advanced machine learning algorithms. These algorithms are effective, but computationally expensive, especially if we plan to apply them on hourly electricity market data covering a number of years. To address this challenge, we significantly accelerate the computation of the Gaussian Process (GP) for time series data. In the context of a Change Point Detection (CPD) algorithm, we reduce its computational complexity from O($n^{5}$) to O($n^{2}$). Our efficient algorithm makes it possible to compute the Change Points using the hourly price data from the California Electricity Crisis. By comparing the detected Change Points with known events, we show that the Change Point Detection algorithm is indeed effective in detecting signals preceding major events.

Mass spectrometry imaging (MSI) enables researchers to directly probe endogenous molecules directly within the architecture of the biological matrix. Unfortunately, efficient access, management, and analysis of the data generated by MSI approaches remain major challenges to this rapidly developing field. Despite the availability of numerous dedicated file formats and software packages, it is a widely held viewpoint that the biggest challenge is simply opening, sharing, and analyzing a file without loss of information. Here we present OpenMSI, a software framework and platform that addresses these challenges via an advanced, high-performance, extensible file format and Web API for remote data access (http://openmsi.nersc.gov). The OpenMSI file format supports storage of raw MSI data, metadata, and derived analyses in a single, self-describing format based on HDF5 and is supported by a large range of analysis software (e.g., Matlab and R) and programming languages (e.g., C++, Fortran, and Python). Careful optimization of the storage layout of MSI data sets using chunking, compression, and data replication accelerates common, selective data access operations while minimizing data storage requirements and are critical enablers of rapid data I/O. The OpenMSI file format has shown to provide >2000-fold improvement for image access operations, enabling spectrum and image retrieval in less than 0.3 s across the Internet even for 50 GB MSI data sets. To make remote high-performance compute resources accessible for analysis and to facilitate data sharing and collaboration, we describe an easy-to-use yet powerful Web API, enabling fast and convenient access to MSI data, metadata, and derived analysis results stored remotely to facilitate high-performance data analysis and enable implementation of Web based data sharing, visualization, and analysis.

Data producers typically optimize the layout of data files to minimize the write time. In most cases, data analysis tasks read these files in access patterns different from the write patterns causing poor read performance. In this paper, we introduce Scientific Data Services (SDS), a framework for bridging the performance gap between writing and reading scientific data. SDS reorganizes data to match the read patterns of analysis tasks and enables transparent data reads from the reorganized data. We implemented a HDF5 Virtual Object Layer (VOL) plugin to redirect the HDF5 dataset read calls to the reorganized data. To demonstrate the effectiveness of SDS, we applied two parallel data organization techniques: a sort-based organization on a plasma physics data and a transpose-based organization on mass spectrometry imaging data. We also extended the HDF5 data access API to allow selection of data based on their values through a query interface, called SDS Query. We evaluated the execution time in accessing various subsets of data through existing HDF5 Read API and SDS Query. We showed that reading the reorganized data using SDS is up to 55X faster than reading the original data.

The Energy Sciences Network (ESnet) is the primary provider of network connectivity for the U.S. Department of Energy (DOE) Office of Science (SC), the single largest supporter of basic research in the physical sciences in the United States. In support of SC programs, ESnet regularly updates and refreshes its understanding of the networking requirements of the instruments, facilities, scientists, and science programs that it serves. This focus has helped ESnet be a highly successful enabler of scientific discovery for over 25 years.In November 2012, ESnet and the Office of Biological and Environmental Research (BER) of the DOE SC organized a review to characterize the networking requirements of the programs funded by the BER program office.Several key findings resulted from the review. Among them:1) The scale of data sets available to science collaborations continues to increase exponentially. This has broad impact, both on the network and on the computational and storage systems connected to the network.2) Many science collaborations require assistance to cope with the systems and network engineering challenges inherent in managing the rapid growth in data scale.3) Several science domains operate distributed facilities that rely on high-performance networking for success. Key examples illustrated in this report include the Earth System Grid Federation (ESGF) and the Systems Biology Knowledgebase (KBase).This report expands on these points, and addresses others as well. The report contains a findings section as well as the text of the case studies discussed at the review.

The Mori-Zwanzig formalism of statistical mechanics is used to estimatethe uncertainty caused by underresolution in the solution of a nonlineardynamical system. A general approach is outlined and applied to a simpleexample. The noise term that describes the uncertainty turns out tobe neither Markovian nor Gaussian. It is argued that this is the generalsituation.

Understanding the microstructure of the financial market requires the processing of a vast amount of data related to individual trades, and sometimes even multiple levels of quotes. Analyzing such a large volume of data requires tremendous computing power that is not easily available to financial academics and regulators. Fortunately, public funded High Performance Computing (HPC) power is widely available at the National Laboratories in the US. In this paper we demonstrate that the HPC resource and the techniques for data-intensive sciences can be used to greatly accelerate the computation of an early warning indicator called Volume-synchronized Probability of Informed trading (VPIN). The test data used in this study contains five and a half year?s worth of trading data for about 100 most liquid futures contracts, includes about 3 billion trades, and takes 140GB as text files. By using (1) a more efficient file format for storing the trading records, (2) more effective data structures and algorithms, and (3) parallelizing the computations, we are able to explore 16,000 different ways of computing VPIN in less than 20 hours on a 32-core IBM DataPlex machine. Our test demonstrates that a modest computer is sufficient to monitor a vast number of trading activities in real-time ? an ability that could be valuable to regulators.Our test results also confirm that VPIN is a strong predictor of liquidity-induced volatility. With appropriate parameter choices, the false positive rates are about 7percent averaged over all the futures contracts in the test data set. More specifically, when VPIN values rise above a threshold (CDF > 0.99), the volatility in the subsequent time windows is higher than the average in 93percent of the cases.

Current steady state rolling tire calculations often do not include treads because treadsdestroy the rotational symmetry of the tire. We describe two methodologies to compute timeperiodic solutions of a two-dimensional viscoelastic tire with treads: solving a minimizationproblem and solving a system of equations. We also expand on work by Oden and Lin onfree spinning rolling elastic tires in which they disovered a hierachy of N-peak steady statestanding wave solutions. In addition to discovering a two-dimensional hierarchy of standingwave solutions that includes their N-peak hiearchy, we consider the effects of viscoelasticityon the standing wave solutions. Finally, a commonplace model of viscoelasticity used in ournumerical experiments led to non-physical elastic energy growth for large tire speeds. Weshow that a viscoelastic model of Govindjee and Reese remedies the problem.

The Belle experiment, part of a broad-based search for new physics, is a collaboration of~;;400 physicists from 55 institutions across four continents. The Belle detector is locatedat the KEKB accelerator in Tsukuba, Japan. The Belle detector was operated at theasymmetric electron-positron collider KEKB from 1999-2010. The detector accumulatedmore than 1 ab-1 of integrated luminosity, corresponding to more than 2 PB of data near10 GeV center-of-mass energy. Recently, KEK has initiated a $400 million acceleratorupgrade to be called SuperKEKB, designed to produce instantaneous and integratedluminosity two orders of magnitude greater than KEKB. The new internationalcollaboration at SuperKEKB is called Belle II. The first data from Belle II/SuperKEKB isexpected in 2015.In October 2012, senior members of the Belle-II collaboration gathered at PNNL todiscuss the computing and neworking requirements of the Belle-II experiment withESnet staff and other computing and networking experts. The day-and-a-half-longworkshop characterized the instruments and facilities used in the experiment, theprocess of science for Belle-II, and the computing and networking equipment andconfiguration requirements to realize the full scientific potential of the collaboration?swork.

To support interactive visualization and analysis of complex, large-scale climate data sets, UV-CDAT integrates a powerful set of scientific computing libraries and applications to foster more efficient knowledge discovery. Connected through a provenance framework, the UV-CDAT components can be loosely coupled for fast integration or tightly coupled for greater functionality and communication with other components. This framework addresses many challenges in the interactive visual analysis of distributed large-scale data for the climate community.

We present for the first time an efficient iterative method to directly solve the four-component Dirac-Kohn-Sham (DKS) density functional theory. Due to the existence of the negative energy continuum in the DKS operator, the existing iterative techniques for solving the Kohn-Sham systems cannot be efficiently applied to solve the DKS systems. The key component of our method is a novel filtering step (F) which acts as a preconditioner in the framework of the locally optimal block preconditioned conjugate gradient (LOBPCG) method. The resulting method, dubbed the LOBPCG-F method, is able to compute the desired eigenvalues and eigenvectors in the positive energy band without computing any state in the negative energy band. The LOBPCG-F method introduces mild extra cost compared to the standard LOBPCG method and can be easily implemented. We demonstrate our method in the pseudopotential framework with a planewave basis set which naturally satisfies the kinetic balance prescription. Numerical results for Pt$_{2}$, Au$_{2}$, TlF, and Bi$_{2}$Se$_{3}$ indicate that the LOBPCG-F method is a robust and efficient method for investigating the relativistic effect in systems containing heavy elements.

Today?s scientific and business applications generate mas- sive data sets that need to be transferred to remote sites for sharing, processing, and long term storage. Because of increasing data volumes and enhancement in current net- work technology that provide on-demand high-speed data access between collaborating institutions, data handling and scheduling problems have reached a new scale. In this paper, we present a new data scheduling model with ad- vance resource provisioning, in which data movement operations are defined with earliest start and latest comple- tion times. We analyze time-dependent resource assign- ment problem, and propose a new methodology to improve the current systems by allowing researchers and higher-level meta-schedulers to use data-placement as-a-service, so they can plan ahead and submit transfer requests in advance. In general, scheduling with time and resource conflicts is NP-hard. We introduce an efficient algorithm to organize multiple requests on the fly, while satisfying users? time and resource constraints. We successfully tested our algorithm in a simple benchmark simulator that we have developed, and demonstrated its performance with initial test results.

The Energy Sciences Network (ESnet) is the primary provider of network connectivity forthe U.S. Department of Energy (DOE) Office of Science (SC), the single largest supporterof basic research in the physical sciences in the United States. In support of SCprograms, ESnet regularly updates and refreshes its understanding of the networkingrequirements of the instruments, facilities, scientists, and science programs that itserves. This focus has helped ESnet to be a highly successful enabler of scientificdiscovery for over 25 years. In October 2012, ESnet and the Office of Advanced Scientific Computing Research (ASCR) of the DOE SC organized a review to characterize the networking requirements of the programs funded by the ASCR program office.The requirements identified at the review are summarized in the Findings section, andare described in more detail in the body of the report.

Improved simulations and sensors are producing datasets whose increasing complexity exhausts our ability to visualize and comprehend them directly. To cope with this problem, we can detect and extract significant features in the data and use them asthe basis for subsequent analysis. Topological methods are valuable in this contextbecause they provide robust and general feature definitions.As the growth of serial computational power has stalled, data analysis is becoming increasingly dependent on massively parallel machines. To satisfy the computational demand created by complex datasets, algorithms need to effectively utilize these computer architectures. The main strength of topological methods, their emphasis on global information, turns into an obstacle during parallelization.We present two approaches to alleviate this problem. We develop a distributedrepresentation of the merge tree that avoids computing the global tree on a singleprocessor and lets us parallelize subsequent queries. To account for the increasingnumber of cores per processor, we develop a new data structure that lets us take advantage of multiple shared-memory cores to parallelize the work on a single node. Finally, we present experiments that illustrate the strengths of our approach as well as help identify future challenges.

Data visualization, as well as data analysis and data analytics, are all an integral part of the scientific process. Collectively, these technologies provide the means to gain insight into data of ever-increasing size and complexity. Over the past two decades, a substantial amount of visualization, analysis, and analytics R&D has focused on the challenges posed by increasing data size and complexity, as well as on the increasing complexity of a rapidly changing computational platform landscape. While some of this research focuses on solely on technologies, such as indexing and searching or novel analysis or visualization algorithms, other R&D projects focus on applying technological advances to specific application problems. Some of the most interesting and productive results occur when these two activities R&D and application are conducted in a collaborative fashion, where application needs drive R&D, and R&D results are immediately applicable to real world problems.

We discuss techniques for accelerating the self consistent field (SCF) iteration for solving the Kohn-Sham equations. These techniques are all based on constructing approximations to the inverse of the Jacobian associated with a fixed point map satisfied by the total potential. They can be viewed as preconditioners for a fixed point iteration. We point out different requirements for constructing preconditioners for insulating and metallic systems respectively, and discuss how to construct preconditioners to keep the convergence rate of the fixed point iteration independent of the size of the atomistic system. We propose a new preconditioner that can treat insulating and metallic system in a unified way. The new preconditioner, which we call an elliptic preconditioner, is constructed by solving an elliptic partial differential equation. The elliptic preconditioner is shown to be more effective in accelerating the convergence of a fixed point iteration than the existing approaches for large inhomogeneous systems at low temperature.