Engineering Papers⌕ Search

Engineering topics

Holzman, Burt

Publications and source records attributed to Holzman, Burt.

30 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

HL-LHC Computing Review: Common Tools and Community Software

Common and community software packages, such as ROOT, Geant4 and event generators have been a key part of the LHC's success so far and continued development and optimisation will be critical in the future. The challenges are driven by an ambitious physics programme, notably the LHC accelerator upgrade to high-luminosity, HL-LHC, and the corresponding detector upgrades of ATLAS and CMS. In this document we address the issues for software that is used in multiple experiments (usually even more widely than ATLAS and CMS) and maintained by teams of developers who are either not linked to a particular experiment or who contribute to common software within the context of their experiment activity. We also give space to general considerations for future software and projects that tackle upcoming challenges, no matter who writes it, which is an area where community convergence on best practice is extremely useful.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2019

FPGA-Accelerated Machine Learning Inference as a Service for Particle Physics Computing

Large-scale particle physics experiments face challenging demands for high-throughput computing resources both now and in the future. New heterogeneous computing paradigms on dedicated hardware with increased parallelization, such as Field Programmable Gate Arrays (FPGAs), offer exciting solutions with large potential gains. The growing applications of machine learning algorithms in particle physics for simulation, reconstruction, and analysis are naturally deployed on such platforms. We demonstrate that the acceleration of machine learning inference as a web service represents a heterogeneous computing solution for particle physics experiments that potentially requires minimal modification to the current computing model. As examples, we retrain the ResNet-50 convolutional neural network to demonstrate state-of-the-art performance for top quark jet tagging at the LHC and apply a ResNet-50 model with transfer learning for neutrino event classification. Using Project Brainwave by Microsoft to accelerate the ResNet-50 image classification model, we achieve average inference times of 60 (10) ms with our experimental physics software framework using Brainwave as a cloud (edge or on-premises) service, representing an improvement by a factor of approximately 30 (175) in model inference latency over traditional CPU inference in current experimental hardware. A single FPGA service accessed by many CPUs achieves a throughput of 600–700 inferences per second using an image batch of one, comparable to large batch-size GPU throughput and significantly better than small batch-size GPU throughput. Deployed as an edge or cloud service for the particle physics computing model, coprocessor accelerators can have a higher duty cycle and are potentially much more cost-effective.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2019

HPC resource integration into CMS Computing via HEPCloud

The higher energy and luminosity from the LHC in Run 2 have put increased pressure on CMS computing resources. Extrapolating to even higher luminosities (and thus higher event complexities and trigger rates) beyond Run 3, it becomes clear that simply scaling up the the current model of CMS computing alone will become economically unfeasible. High Performance Computing (HPC) facilities, widely used in scientific computing outside of HEP, have the potential to help fill the gap. Here we describe the U.S.CMS efforts to integrate US HPC resources into CMS Computing via the HEPCloud project at Fermilab. We present advancements in our ability to use NERSC resources at scale and efforts to integrate other HPC sites as well. We present experience in the elastic use of HPC resources, quickly scaling up use when so required by CMS workflows. We also present performance studies of the CMS multi-threaded framework on both Haswell and KNL HPC resources.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

HEPCloud, an Elastic Hybrid HEP Facility using an Intelligent Decision Support System

HEPCloud is rapidly becoming the primary system for provisioning compute resources for all Fermilab-affiliated experiments. In order to reliably meet the peak demands of the next generation of High Energy Physics experiments, Fermilab must plan to elastically expand its computational capabilities to cover the forecasted need. 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 at which 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. In this paper, we discuss the goals and architecture of the HEPCloud Facility, the architecture of the IDSS, and our early experience in using the IDSS for automated facility expansion both at Fermi and Brookhaven National Laboratory.

Mhashilkar, Parag↗
DOE OSTI2019

FPGAs as a Service to Accelerate Machine Learning Inference [PowerPoint]

Large-scale particle physics experiments face challenging demands for high-throughput computing resources both now and in the future. New heterogeneous computing paradigms on dedicated hardware with increased parallelization, such as Field Programmable Gate Arrays (FPGAs), offer exciting solutions with large potential gains. The growing applications of machine learning algorithms in particle physics for simulation, reconstruction, and analysis are naturally deployed on such platforms. We demonstrate that the acceleration of machine learning inference as a web service represents a heterogeneous computing solution for particle physics experiments that requires minimal modification to the current computing model. As examples, we retrain the ResNet-50 convolutional neural network to demonstrate state-of-the-art performance for top quark jet tagging at the LHC, and apply a ResNet-50 model with transfer learning for neutrino event classification. Using Project Brainwave b y Microsoft to accelerate the ResNet-50 image classification model, we achieve average inference times of 60 (10) milliseconds with our experimental physics software framework using Brainwave as a cloud (edge or on-premises) service, representing an improvement by a factor of approximately 30 (175) in model inference latency over traditional CPU inference in current experimental hardware. A single FPGA service accessed by many CPUs achieves a throughput of 600--700 inferences per second using an image batch of one, comparable to large batch-size GPU throughput and significantly better than small batch-size GPU throughput. Deployed as an edge or cloud service for the particle physics computing model, coprocessor accelerators can have a higher duty cycle and are potentially much more cost-effective.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2019

Implementation of Feldman-Cousins Corrections and Oscillation Calculations in the HPC Environment for the NOvA Experiment

Analysis of neutrino oscillation data involves a combination of complex fitting procedures and statistical correction techniques that are used to determine the full three-flavor PMNS parameters and constraint contours. These techniques rely on computationally intensive “multi-universe” stochastic modeling. The process of calculating these contours and corrections can dominate final stages of the data analysis and become a bottleneck for examining the ef-fect of systematic variations on the final results. As part of the DOE SciDAC-4 sponsored research program, we present a new implementation of a neutrino oscillation fitting and framework to carry out calculations of Feldman-Cousins corrections. This implementation is specifically designed and optimized to operate on modern High-Performance Computing (HPC) facilities. We present the performance of the system in calculating allowed regions for the NOvA experiment based on 8.9 x 1020 and 6.9 x 1020 protons-on-target (POT) neutrino and antineutrino datasets, and compare the performance of this new implementation run at the NERSC supercomputing facility with that from methods used previously by the NOvA collaboration running on grid computing facilities.

Sousa, Alexandre↗
DOE OSTI2019

Accelerated Machine Learning as a Service for Particle Physics Computing

Accelerated Machine Learning as a Service for Particle Physics Computing: • Amount and complexity of high-energy physics data increases dramatically from 2025 onward • Traditional algorithms will require too much CPU time • Machine learning can solve combinatorially-scaling problems in constant time, but must be fast enough

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
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

HEPCloud, a New Paradigm for HEP Facilities: CMS Amazon Web Services Investigation

Historically, high energy physics computing has been performed on large purpose-built computing systems. These began as single-site compute facilities, but have evolved into the distributed computing grids used today. Recently, there has been an exponential increase in the capacity and capability of commercial clouds. Cloud resources are highly virtualized and intended to be able to be flexibly deployed for a variety of computing tasks. There is a growing interest among the cloud providers to demonstrate the capability to perform large-scale scientific computing. In this paper, we discuss results from the CMS experiment using the Fermilab HEPCloud facility, which utilized both local Fermilab resources and virtual machines in the Amazon Web Services Elastic Compute Cloud. We discuss the planning, technical challenges, and lessons learned involved in performing physics workflows on a large-scale set of virtualized resources. Additionally, we will discuss the economics and operational efficiencies when executing workflows both in the cloud and on dedicated resources.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2014

Cloud Bursting with GlideinWMS: Means to satisfy ever increasing computing needs for Scientific Workflows

Scientific communities have been in the forefront of adopting new technologies and methodologies in the computing. Scientific computing has influenced how science is done today, achieving breakthroughs that were impossible to achieve several decades ago. For the past decade several such communities in the Open Science Grid (OSG) and the European Grid Infrastructure (EGI) have been using GlideinWMS to run complex application workflows to effectively share computational resources over the grid. GlideinWMS is a pilot-based workload management system (WMS) that creates on demand, a dynamically sized overlay HTCondor batch system on grid resources. At present, the computational resources shared over the grid are just adequate to sustain the computing needs. We envision that the complexity of the science driven by 'Big Data' will further push the need for computational resources. To fulfill their increasing demands and/or to run specialized workflows, some of the big communities like CMS are investigating the use of cloud computing as Infrastructure-As-A-Service (IAAS) with GlideinWMS as a potential alternative to fill the void. Similarly, communities with no previous access to computing resources can use GlideinWMS to setup up a batch system on the cloud infrastructure. To enable this, the architecture of GlideinWMS has been extended to enable support for interfacing GlideinWMS with different Scientific and commercial cloud providers like HLT, FutureGrid, FermiCloud and Amazon EC2. In this paper, we describe a solution for cloud bursting with GlideinWMS. The paper describes the approach, architectural changes and lessons learned while enabling support for cloud infrastructures in GlideinWMS.

Mhashilkar, Parag↗
DOE OSTI2014

CMS multicore scheduling strategy

In the next years, processor architectures based on much larger numbers of cores will be most likely the model to continue 'Moore's Law' style throughput gains. This not only results in many more jobs in parallel running the LHC Run 1 era monolithic applications, but also the memory requirements of these processes push the workernode architectures to the limit. One solution is parallelizing the application itself, through forking and memory sharing or through threaded frameworks. CMS is following all of these approaches and has a comprehensive strategy to schedule multicore jobs on the GRID based on the glideinWMS submission infrastructure. The main component of the scheduling strategy, a pilot-based model with dynamic partitioning of resources that allows the transition to multicore or whole-node scheduling without disallowing the use of single-core jobs, is described. This contribution also presents the experiences made with the proposed multicore scheduling schema and gives an outlook of further developments working towards the restart of the LHC in 2015.

Perez-Calero Yzquierdo, Antonio↗