Engineering Papers⌕ Search

Engineering topics

Paterno, Marc

Publications and source records attributed to Paterno, Marc.

34 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2019

Parallel Event Selection on HPC Systems

In their recent measurement of the neutrino oscillation parameters,NOvA uses a sample of approximately 25 million reconstructed spills to searchfor electron-neutrino appearance events. These events are stored in an n-tupleformat, in 250 thousand ROOT files. File sizes range from a few hundred KiB toa few MiB; the full dataset is approximately 1.4 TiB. These millions of eventsare reduced to a few tens of events by the application of strict event selectioncriteria, and then summarized by a handful of numbers each, which are used inthe extraction of the neutrino oscillation parameters.The NOvA event selection code is currently a serial C++ program that readsthese n-tuples. The current table data format and organization and the selection/reduction processing involved provides us with an opportunity to explorealternate approaches to represent the data and implement the processing. Werepresent our n-tuple data in HDF5 format that is optimized for the HPC environmentand which allows us to use the machine’s high-performance parallelI/O capabilities. We use MPI, numpy and h5py to implement our approach andcompare the performance with the existing approach. We study the performanceimplications of using thousands of small files of different sizes as comparedwith one large file using HPC resources. This work has been done as part of theSciDAC project, “HEP analytics on HPC” in collaboration with the ASCR teamsat ANL and LBNL.

Paterno, Marc↗
DOE OSTI2018

HEP Software Foundation Community White Paper Working Group - Data Processing Frameworks

Data processing frameworks are an essential part of HEP experiments' software stacks. Frameworks provide a means by which code developers can undertake the essential tasks of physics data processing, accessing relevant inputs and storing their outputs, in a coherent way without needing to know the details of other domains. Frameworks provide essential core services for developers and help deliver a configurable working application to the experiments' production systems. Modern HEP processing frameworks are in the process of adapting to a new computing landscape dominated by parallel processing and heterogeneity, which pose many questions regarding enhanced functionality and scaling that must be faced without compromising the maintainability of the code. In this paper we identify a program of work that can help further clarify the key concepts of frameworks for HEP and then spawn R&D activities that can focus the community's efforts in the most efficient manner to address the challenges of the upcoming experimental program.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2018

Supporting HEP Data as the Exascale Era Approaches

We describe our work on the use of HPC data formats and storage technologies to efficiently describe, store and access High Energy Physics (HEP) data. In the HEP community, important data sets are large, currently at the scale of tens of PiB, and are expected to grow dramatically in the next decade. These data require different data representations—ranging from complex C++ objects to tabular data—that are used at the many varied stages of data analysis. Traditionally, these data have been stored in large collections of files at each stage. This approach does not scale well on HPC resources. To obtain good scalability, without requiring that analysts have parallel programming expertise, our work stresses the use of HPCoptimized tools and implicit data-parallelism. We describe data organization and the performance and scaling of data processing using examples from three HEP experiments (LArIAT, NOvA and DUNE), and for three different stages of analysis. We compare the performance of our HPC-oriented solutions to the traditional HEP approaches.

Buchanan, Norman↗
DOE OSTI2018

Data-parallel Python for High Energy Physics Analyses

In this paper, we explore features available in Python which are useful for data reduction tasks in High Energy Physics (HEP). Highlevel abstractions in Python are convenient for implementing data reduction tasks. However, in order for such abstractions to be practical, the efficiency of their performance must also be high. Because the data sets we process are typically large, we care about both I/O performance and in-memory processing speed. In particular, we evaluate the use of data-parallel programming, using MPI and numpy, to process a large experimental data set (42 TiB) stored in an HDF5 file. We measure the speed of processing of the data, distinguishing between the time spent reading data and the time spent processing the data in memory, and demonstrate the scalability of both, using up to 1200 KNL nodes (76800 cores) on Cori at NERSC

Paterno, Marc↗
DOE OSTI2018

A Methodology for the Rapid Development of Scalable HPC Data Services

There is growing evidence in the scientifc computing community that parallel file systems are not sufficient for all HPC storage workloads. This realization has motivated extensive research in new storage system designs. The question of which alternative we should turn towards implies that there could be a single answer satisfying a wide range of very diverse applications. We argue that such a generic solution does not exist. Instead, custom data services should be designed and tailored to the needs of specific applications on specific hardware. Furthermore, custom data services should be designed in close collaboration with users. In this paper, we present a methodology for the rapid development of such data services. This methodology promotes the design of reusable building blocks that can be composed together efficiently through a runtime based on high-performance threading, tasking, and remote procedure calls. We illustrate the success of our methodology by showcasing three examples of data services designed from the same building blocks, yet targeting entirely different applications.

Dorier, Mattieu↗
DOE OSTI2018

Intelligently-Automated Facilities Expansion with the HEPCloud Decision Engine

The next generation of High Energy Physics experiments are expected to generate exabytes of data---two orders of magnitude greater than the current generation. In order to reliably meet peak demands, facilities must either plan to provision enough resources to cover the forecasted need, or find ways to elastically expand their computational capabilities. Commercial cloud and allocation-based High Performance Computing (HPC) resources both have explicit and implicit costs that must be considered when deciding when to provision these resources, and to choose an appropriate scale. In order to support such provisioning in a manner consistent with organizational business rules and budget constraints, we have developed a modular intelligent decision support system (IDSS) to aid in the automatic provisioning of resources---spanning multiple cloud providers, multiple HPC centers, and grid computing federations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2017

Fermilab HEPCloud Facility Decision Engine Design

The Decision Engine is a critical component of the HEP Cloud Facility. It provides the functionality of resource scheduling for disparate resource providers, including those which may have a cost or a restricted allocation of cycles. Along with the architecture, design, and requirements for the Decision Engine, this document will provide the rationale and explanations for various design decisions. In some cases, requirements and interfaces for a limited subset of external services will be included in this document. This document is intended to be a high level design. The design represented in this document is not complete and does not break everything down in detail. The class structures and pseudo-code exist for example purposes to illustrate desired behaviors, and as such, should not be taken literally. The protocols and behaviors are the important items to take from this document. This project is still in prototyping mode so flaws and inconsistencies may exist and should be noted and treated as failures.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2017

Spark and HPC for High Energy Physics Data Analyses

A full High Energy Physics (HEP) data analysis is divided into multiple data reduction phases. Processing within these phases is extremely time consuming, therefore intermediate results are stored in files held in mass storage systems and referenced as part of large datasets. This processing model limits what can be done with interactive data analytics. Growth in size and complexity of experimental datasets, along with emerging big data tools are beginning to cause changes to the traditional ways of doing data analyses. Use of big data tools for HEP analysis looks promising, mainly because extremely large HEP datasets can be represented and held in memory across a system, and accessed interactively by encoding an analysis using highlevel programming abstractions. The mainstream tools, however, are not designed for scientific computing or for exploiting the available HPC platform features. We use an example from the Compact Muon Solenoid (CMS) experiment at the Large Hadron Collider (LHC) in Geneva, Switzerland. The LHC is the highest energy particle collider in the world. Our use case focuses on searching for new types of elementary particles explaining Dark Matter in the universe. We use HDF5 as our input data format, and Spark to implement the use case. We show the benefits and limitations of using Spark with HDF5 on Edison at NERSC.

Sehrish, Saba↗
DOE OSTI2017

Processing Contexts for Experimental HEP Data

This document provides, for those not closely associated with the experimental High Energy Physics (HEP) community, an introduction to data input and output requirements for a variety of data processing tasks. Examples in it are drawn from the art event processing framework, and from experiments and projects using art, most notably the LArSoft and NuTools projects.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Study of a Docker Use-Case for HEP

Over the past few years, container technology has become increasingly promising as a means to seamlessly make our software available across a wider range of platforms. In December 2015, we decided to put together a set of docker images that serve as a demonstration of this container technology for managing a run-time environment for art-related software projects, and also serve as a set of test cases for evaluation of performance. Docker[1] containers provide a way to “wrap up a piece of software in a complete filesystem that contains everything it needs to run”. In combination with Shifter[2], such containers provide a way to run software developed and deployed on “typical” HEP platforms (such as SLF 6, in common use at Fermilab and on OSG platforms) on HPC facilities at NERSC. Docker containers provide a means of delivering software that can be run on a variety of hosts without needing to be compiled specially for each OS to be supported. This could substantially reduce the effort required to create and validate a new release, since one build could be suitable for use on both grid machines (both FermiGrid and OSG) as well as any machine capable of running the Docker container. In addition, docker containers may provide for a quick and easy way for users to install and use a software release in a standardized environment. This report contains the results and status of this demonstration and evaluation.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Exploring the Performance of Spark for a Scientific Use Case

We present an evaluation of the performance of a Spark implementation of a classification algorithm in the domain of High Energy Physics (HEP). Spark is a general engine for in-memory, large-scale data processing, and is designed for applications where similar repeated analysis is performed on the same large data sets. Classification problems are one of the most common and critical data processing tasks across many domains. Many of these data processing tasks are both computation- and data-intensive, involving complex numerical computations employing extremely large data sets. We evaluated the performance of the Spark implementation on Cori, a NERSC resource, and compared the results to an untuned MPI implementation of the same algorithm. While the Spark implementation scaled well, it is not competitive in speed to our MPI implementation, even when using significantly greater computational resources.

Sehrish, Saba↗
DOE OSTI2015

Geant4 Computing Performance Benchmarking and Monitoring

Performance evaluation and analysis of large scale computing applications is essential for optimal use of resources. As detector simulation is one of the most compute intensive tasks and Geant4 is the simulation toolkit most widely used in contemporary high energy physics (HEP) experiments, it is important to monitor Geant4 through its development cycle for changes in computing performance and to identify problems and opportunities for code improvements. All Geant4 development and public releases are being profiled with a set of applications that utilize different input event samples, physics parameters, and detector configurations. Results from multiple benchmarking runs are compared to previous public and development reference releases to monitor CPU and memory usage. Observed changes are evaluated and correlated with code modifications. Besides the full summary of call stack and memory footprint, a detailed call graph analysis is available to Geant4 developers for further analysis. The set of software tools used in the performance evaluation procedure, both in sequential and multi-threaded modes, include FAST, IgProf and Open|Speedshop. In conclusion, the scalability of the CPU time and memory performance in multi-threaded application is evaluated by measuring event throughput and memory gain as a function of the number of threads for selected event samples.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

Exploring Two Approaches for an End-to-End Scientific Analysis Workflow

The advance of the scientific discovery process is accomplished by the integration of independently-developed programs run on disparate computing facilities into coherent workflows usable by scientists who are not experts in computing. For such advancement, we need a system which scientists can use to formulate analysis workflows, to integrate new components to these workflows, and to execute different components on resources that are best suited to run those components. In addition, we need to monitor the status of the workflow as components get scheduled and executed, and to access the intermediate and final output for visual exploration and analysis. Finally, it is important for scientists to be able to share their workflows with collaborators. Moreover we have explored two approaches for such an analysis framework for the Large Synoptic Survey Telescope (LSST) Dark Energy Science Collaboration (DESC), the first one is based on the use and extension of Galaxy, a web-based portal for biomedical research, and the second one is based on a programming language, Python. In our paper, we present a brief description of the two approaches, describe the kinds of extensions to the Galaxy system we have found necessary in order to support the wide variety of scientific analysis in the cosmology community, and discuss how similar efforts might be of benefit to the HEP community.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

Some good C++ practices for using the art framework

This document is intended for an audience that has some programming experience, at least beginning familiarity with C++, and at least beginning familiarity with the art2 framework. The intent of this document is to help the reader avoid some of the more common mistakes made by those with little experience in C++, or in use of the art framework, or both.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

CosmoSIS: A System for MC Parameter Estimation

Cosmological parameter estimation is entering a new era. Large collaborations need to coordinate high-stakes analyses using multiple methods; furthermore such analyses have grown in complexity due to sophisticated models of cosmology and systematic uncertainties. In this paper we argue that modularity is the key to addressing these challenges: calculations should be broken up into interchangeable modular units with inputs and outputs clearly defined. We present a new framework for cosmological parameter estimation, CosmoSIS, designed to connect together, share, and advance development of inference tools across the community. We describe the modules already available in Cosmo- SIS, including camb, Planck, cosmic shear calculations, and a suite of samplers. We illustrate it using demonstration code that you can run out-of-the-box with the installer available at http://bitbucket.org/joezuntz/cosmosis.

79 ASTRONOMY AND ASTROPHYSICS↗