Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “workload”

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 307 records · Page 17

Adoption of a token-based authentication model for the CMS Submission Infrastructure

The CMS Submission Infrastructure (SI) is the main computing resource provisioning system for CMS workloads. A number of HTCondor pools are employed to manage this infrastructure, which aggregates geographically distributed resources from the WLCG and other providers. Historically, the model of authentication among the diverse components of this infrastructure has relied on the Grid Security Infrastructure (GSI), based on identities and X509 certificates. In contrast, commonly used modern authentication standards are based on capabilities and tokens. The WLCG has identified this trend and aims at a transparent replacement of GSI for all its workload management, data transfer and storage access operations, to be completed during the current LHC Run 3. As part of this effort, and within the context of CMS computing, the Submission Infrastructure group is in the process of phasing out the GSI part of its authentication layers, in favor of IDTokens and Scitokens. The use of tokens is already well integrated into the HTCondor Software Suite, which has allowed us to fully migrate the authentication between internal components of SI. Additionally, recent versions of the HTCondor-CE support tokens as well, enabling CMS resource requests to Grid sites employing this CE technology to be granted by means of token exchange. After a rollout campaign to sites, successfully completed by the third quarter of 2022, the totality of HTCondor CEs in use by CMS are already receiving Scitoken-based pilot jobs. On the ARC CE side, a parallel campaign was launched to foster the adoption of the REST interface at CMS sites (required to enable token-based job submission via HTCondor-G), which is nearing completion as well. In this contribution, the newly adopted authentication model will be described. We will then report on the migration status and final steps towards complete GSI phase out in the CMS SI.

Pérez-Calero Yzquierdo, Antonio↗

Operational Analytics Studies for ATLAS Distributed Computing: Data Popularity Forecast and Utilization of the WLCG Centers

Operational analytics is the direction of research related to the analysis of the current state of computing processes and the prediction of future states in order to anticipate imbalances and take timely measures to stabilize a complex system. There are two relevant areas in ATLAS Distributed Computing that are currently the focus of studies: user physics analysis including the forecast of popularity of data samples among users, and evaluating WLCG centers for their readiness to process user analysis payloads. Studying these areas is challenging due to the complexity involved, as it requires a comprehensive understanding of numerous boundary conditions typically found in large-scale distributed computing infrastructures. Forecasts of data popularity are problematic without the categorization of user tasks by their types (data transformation or physics analysis), which do not always appear on the surface but may induce noise, which introduces significant distortions for predictive analysis. Evaluating the WLCG resources by their analysis workloads is also a challenging task as it is necessary to find a balance between the workload of the resource, its performance, the waiting time for jobs on it, as well as the volume of jobs that it processes. This is especially difficult in a heterogeneous computing environment, where legacy resources are used along with modern high-performance machines. We will look at these areas of research in detail and discuss what tools and methods are used in our work, demonstrating results already obtained.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Oscillatory Spreading and Inertia in Power Grids

The increase in variable renewable generators (VRGs) in power systems has altered the dynamics from a historical experience. VRGs introduce new sources of power oscillations, and the stabilizing response provided by synchronous generators (SGs, e.g., natural gas, coal, etc.), which help avoid some power fluctuations, will lessen as VRGs replace SGs. These changes have led to the need for new methods and metrics to quickly assess the likely oscillatory behavior for a particular network without performing computationally expensive simulations. This work studies the impact of a critical dynamical parameter - the inertia value - on the rest of a power system's oscillatory response to representative VRG perturbations. We use a known localization metric in a novel way to quantify the number of nodes responding to a perturbation and the magnitude of those responses. This metric allows us to relate the spread and severity of a system's power oscillations with inertia. We find that as inertia increases, the system response to node perturbations transitions from localized (only a few close nodes respond) to delocalized (many nodes across the network respond). We introduce a heuristic computed from the network Laplacian to relate this oscillatory transition to the network structure. We show that our heuristic accurately describes the spread of oscillations for a realistic power-system test case. Using a heuristic to determine the likely oscillatory behavior of a system given a set of parameters has wide applicability in power systems, and it could decrease the computational workload of planning and operation.

dynamical systems↗

Distributed deep learning training using silicon photonic switched architectures

The scaling trends of deep learning models and distributed training workloads are challenging network capacities in today’s datacenters and high-performance computing (HPC) systems. We propose a system architecture that leverages silicon photonic (SiP) switch-enabled server regrouping using bandwidth steering to tackle the challenges and accelerate distributed deep learning training. In addition, our proposed system architecture utilizes a highly integrated operating system-based SiP switch control scheme to reduce implementation complexity. To demonstrate the feasibility of our proposal, we built an experimental testbed with a SiP switch-enabled reconfigurable fat tree topology and evaluated the network performance of distributed ring all-reduce and parameter server workloads. The experimental results show up to 3.6× improvements over the static non-reconfigurable fat tree. Our large-scale simulation results show that server regrouping can deliver up to 2.3× flow throughput improvement for a 2× tapered fat tree and a further 11% improvement when higher-layer bandwidth steering is employed. The collective results show the potential of integrating SiP switches into datacenters and HPC systems to accelerate distributed deep learning training.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Supporting multiple hardware architectures at CMS: the integration and validation of POWER9

Computing resources in the Worldwide LHC Computing Grid (WLCG) have been based entirely on the x86 architecture for more than two decades. In the near future, however, heterogeneous non-x86 resources, such as ARM, POWER and Risc-V, will become a substantial fraction of the resources that will be provided to the LHC experiments, due to their presence in existing and planned world-class HPC installations. The CMS experiment, one of the four large detectors at the LHC, has started to prepare for this situation, with the CMS software stack (CMSSW) already compiled for multiple architectures. In order to allow for a production use, the tools for workload management and job distribution need to be extended to be able to exploit heterogeneous architectures. Profiting from the opportunity to exploit the first sizable IBM Power9 allocation available on Marconi100 HPC system at CINECA, CMS developed all the needed modifications to the CMS workload management system. After a successful proof of concept, a full physics validation has been performed in order to bring the system in production. The experiences are of very high value, when it comes to commissioning of the similar (even larger) Summit HPC system at Oak Ridge, where CMS is also expecting a resource allocation. Moreover the compute power of those systems is being provided also via GPUs and this represents an extremely valuable opportunity to exploit the offloading capability already implemented in CMSSW. The status of the current integration including the exploitation of the GPUs, the results of the validation as well as the future plans will be shown and discussed.

Boccali, Tommaso [INFN, Pisa]↗

Large scale and low latency analysis facilities for the CMS experiment: Development and operational aspects

While a majority of CMS data analysis activities rely on the distributed computing infrastructure on the WLCG Grid, dedicated local computing facilities have been deployed to address particular requirements in terms of latency and scale. The CMS CERN Analysis Facility (CAF) was primarily designed to host a large variety of latency-critical workflows. These break down into alignment and calibration, detector commissioning and diagnosis, and high-interest physics analysis requiring fast turnaround. In order to reach the goal for fast turnaround tasks, the Workload Management group has designed a CRABServer based system to fit with two main needs: to provide a simple, familiar interface to the user (as used in the CRAB Analysis Tool[7]) and to allow an easy transition to the Tier-0 system. While the CRABServer component had been initially designed for Grid analysis by CMS end-users, with a few modifications it turned out to be also a very powerful service to manage and monitor local submissions on the CAF. Transition to Tier-0 has been guaranteed by the usage of the WMCore, a library developed by CMS to be the common core of workload management tools, for handing data driven workflow dependencies. This system is now being used with the first use cases, and important experience is being acquired. In addition to the CERN CAF facility, FNAL has CMS dedicated analysis resources at the FNAL LHC Physics Center (LPC). In the first few years of data collection FNAL has been able to accept a large fraction of CMS data. The remote centre is not well suited for the extremely low latency work expected of the CAF, but the presence of substantial analysis resources, a large resident community, and a large fraction of the data make the LPC a strong facility for resource intensive analysis. We present the building, commissioning and operation of these dedicated analysis facilities in the first year of LHC collisions, we also present the specific development to our software needed to allow for the use of these computing facilities in the special use cases of fast turnaround analyses.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A gLite FTS based solution for managing user output in CMS

The CMS distributed data analysis workflow assumes that jobs run in a different location from where their results are finally stored. Typically the user output must be transferred across the network from one site to another, possibly on a different continent or over links not necessarily validated for high bandwidth/high reliability transfer. This step is named stage-out and in CMS was originally implemented as a synchronous step of the analysis job execution. However, our experience showed the weakness of this approach both in terms of low total job execution efficiency and failure rates, wasting precious CPU resources. The nature of analysis data makes it inappropriate to use PhEDEx, the core data placement system for CMS. As part of the new generation of CMS Workload Management tools, the Asynchronous Stage-Out system (AsyncStageOut) has been developed to enable third party copy of the user output. The AsyncStageOut component manages glite FTS transfers of data from the temporary store at the site where the job ran to the final location of the data on behalf of that data owner. The tool uses python daemons, built using the WMCore framework, and CouchDB, to manage the queue of work and FTS transfers. CouchDB also provides the platform for a dedicated operations monitoring system. In this paper, we present the motivations of the asynchronous stage-out system. We give an insight into the design and the implementation of key features, describing how it is coupled with the CMS workload management system. Finally, we show the results and the commissioning experience.

Cinquilli, M.↗

CRAB3: Establishing a new generation of services for distributed analysis at CMS

In CMS Computing the highest priorities for analysis tools are the improvement of the end users ability to produce and publish reliable samples and analysis results as well as a transition to a sustainable development and operations model. To achieve these goals CMS decided to incorporate analysis processing into the same framework as data and simulation processing. This strategy foresees that all workload tools (TierO, Tier1, production, analysis) share a common core with long term maintainability as well as the standardization of the operator interfaces. The re-engineered analysis workload manager, called CRAB3, makes use of newer technologies, such as RESTFul based web services and NoSQL Databases, aiming to increase the scalability and reliability of the system. As opposed to CRAB2, in CRAB3 all work is centrally injected and managed in a global queue. A pool of agents, which can be geographically distributed, consumes work from the central services serving the user tasks. The new architecture of CRAB substantially changes the deployment model and operations activities. In this paper we present the implementation of CRAB3, emphasizing how the new architecture improves the workflow automation and simplifies maintainability. In particular, we will highlight the impact of the new design on daily operations.

Cinquilli, M.↗

Multi-core processing and scheduling performance in CMS

Commodity hardware is going many-core. We might soon not be able to satisfy the job memory needs per core in the current single-core processing model in High Energy Physics. In addition, an ever increasing number of independent and incoherent jobs running on the same physical hardware not sharing resources might significantly affect processing performance. It will be essential to effectively utilize the multi-core architecture. CMS has incorporated support for multi-core processing in the event processing framework and the workload management system. Multi-core processing jobs share common data in memory, such us the code libraries, detector geometry and conditions data, resulting in a much lower memory usage than standard single-core independent jobs. Exploiting this new processing model requires a new model in computing resource allocation, departing from the standard single-core allocation for a job. The experiment job management system needs to have control over a larger quantum of resource since multi-core aware jobs require the scheduling of multiples cores simultaneously. CMS is exploring the approach of using whole nodes as unit in the workload management system where all cores of a node are allocated to a multi-core job. Whole-node scheduling allows for optimization of the data/workflow management (e.g. I/O caching, local merging) but efficient utilization of all scheduled cores is challenging. Dedicated whole-node queues have been setup at all Tier-1 centers for exploring multi-core processing workflows in CMS. We present the evaluation of the performance scheduling and executing multi-core workflows in whole-node queues compared to the standard single-core processing workflows.

Hernandez, J. M.↗

Integration of Titan supercomputer at OLCF with ATLAS Production System

The PanDA (Production and Distributed Analysis) workload management system was developed to meet the scale and complexity of distributed computing for the ATLAS experiment. PanDA managed resources are distributed worldwide, on hundreds of computing sites, with thousands of physicists accessing hundreds of Petabytes of data and the rate of data processing already exceeds Exabyte per year. While PanDA currently uses more than 200,000 cores at well over 100 Grid sites, future LHC data taking runs will require more resources than Grid computing can possibly provide. Additional computing and storage resources are required. Therefore ATLAS is engaged in an ambitious program to expand the current computing model to include additional resources such as the opportunistic use of supercomputers. In this paper we will describe a project aimed at integration of ATLAS Production System with Titan supercomputer at Oak Ridge Leadership Computing Facility (OLCF). Current approach utilizes modified PanDA Pilot framework for job submission to Titan's batch queues and local data management, with lightweight MPI wrappers to run single node workloads in parallel on Titan's multi-core worker nodes. It provides for running of standard ATLAS production jobs on unused resources (backfill) on Titan. The system already allowed ATLAS to collect on Titan millions of core-hours per month, execute hundreds of thousands jobs, while simultaneously improving Titans utilization efficiency. We will discuss the details of the implementation, current experience with running the system, as well as future plans aimed at improvements in scalability and efficiency.

97 MATHEMATICS AND COMPUTING↗

Empirically-calibrated H100 node power models for accurate AI training energy estimation

Accurately quantifying the energy use of artificial intelligence (AI) training is critical for infrastructure planning, carbon accounting, and sustainable data center operation, but few studies have directly measured the power consumption of production workloads on contemporary hardware. By combining empirical measurements from Brookhaven National Laboratory during AI training on 8-graphics-processing-unit H100 systems with open-source benchmarking data, we develop statistical models relating computational intensity to node-level power consumption. We measure the gap between manufacturer-rated thermal design power (TDP) and actual power demand during AI training. Our analysis reveals that even computationally intensive workloads operate at only 76% of the 10.2 kW TDP rating. Our architecture-specific model, calibrated to floating-point operations, predicts energy consumption with 11.4% mean absolute percentage error, significantly outperforming TDP-based approaches (27%–37% error). We identified distinct power signatures between transformer and convolutional neural network architectures, with transformers showing characteristic fluctuations that may impact grid stability. These results provide a measurement-grounded basis for improving AI training energy estimates, enabling more reliable infrastructure sizing, cost projections, and environmental impact assessments.

Newkirk, Alex C↗

Exploring Hardware Support For Scaling Irregular Applications on Multi-node Multi-core Architectures

With the recent emergence of large-scale knowledge dis- covery, data mining and social network analysis, irregular applications have gained renewed interest. Classic cache-based high-performance architectures do not provide optimal performances with such kind of workloads, mainly due to the very low spatial and temporal locality of the irregular control and memory access patterns. In this paper, we present a multi-node, multi-core, fine-grained multi-threaded shared-memory system architecture specifically designed for the execution of large-scale irregular applications, and built on top of three pillars, that we believe are fundamental to support these workloads. First, we offer transparent hardware support for Partitioned Global Address Space (PGAS) to provide a large globally-shared address space with no software library overhead. Second, we employ multi-threaded multi-core processing nodes to achieve the necessary latency tolerance required by accessing global memory, which potentially resides in a remote node. Finally, we devise hardware support for inter-thread synchronization on the whole global address space. We first model the performances by using an analytical model that takes into account the main architecture and application characteristics. We describe the hardware design of the proposed cus- tom architectural building blocks that provide support for the above- mentioned three pillars. Finally, we present a limited-scale evaluation of the system on a multi-board FPGA prototype with typical irregular kernels and benchmarks. The experimental evaluation demonstrates the architecture performance scalability for different configurations of the whole system.

FPGA↗

A Runtime Environment for Supporting Research in Resilient HPC System Software & Tools

The high-performance computing (HPC) community continues to increase the size and complexity of hardware platforms that support advanced scientific workloads. The runtime environment (RTE) is a crucial layer in the software stack for these large-scale systems. The RTE manages the interface between the operating system and the application running in parallel on the machine. The deployment of applications and tools on large-scale HPC computing systems requires the RTE to manage process creation in a scalable manner, support sparse connectivity, and provide fault tolerance. We have developed a new RTE that provides a basis for building distributed execution environments and developing tools for HPC to aid research in system software and resilience. This paper describes the software architecture of the Scalable runTime Component Infrastructure (STCI), which is intended to provide a complete infrastructure for scalable start-up and management of many processes in large-scale HPC systems. We highlight features of the current implementation, which is provided as a system library that allows developers to easily use and integrate STCI in their tools and/or applications. The motivation for this work has been to support ongoing research activities in fault-tolerance for large-scale systems. We discuss the advantages of the modular framework employed and describe two use cases that demonstrate its capabilities: (i) an alternate runtime for a Message Passing Interface (MPI) stack, and (ii) a distributed control and communication substrate for a fault-injection tool.

Computer Science↗

Containers for Massive Ensemble of I/O Bound Hierarchical Coupled Simulations

We present our experience using containers to scale up a massive ensemble of coupled I/O bound workloads on the NERSC Cori supercomputer. We describe the design of a hierarchical simulation structure using the Integrated Plasma Simulator (IPS) that enables the flexible execution of coupled simulations at the system, node, and core level using the same coupling abstraction and API. The hierarchical design allows for the node-level execution to be efficiently executed using containers while not impacting the structure of the simulation at the system level. We demonstrate the viability of the approach by presenting experimental results from applications in coupled fusion plasma simulations that illustrate the performance impact of using containers to deploy the node-level workloads, in conjunction with the user mountable XFS file systems to ameliorate the load on the Lustre parallel file system. We also present results from production runs showing the ability of the ensemble simulations to scale to hundreds of Cori Haswell nodes, with little or no overhead.

Elwasif, Wael↗

Overlapped checkpointing with hardware assist

We present a new approach to handling the demanding I/O workload incurred during checkpoint writes encountered in High Performance Computing. Prior efforts to improve performance have been primarily bound by mechanical limitations of the hard drive. Our research surpasses this limitation by providing a method to: (1) write checkpoint data to a high-speed, non-volatile buffer, and (2) asynchronously write this data to permanent storage while resuming computation. This removes the hard drive from the critical data path because our I/O node based buffers isolate the compute nodes from the storage servers. This solution is feasible because of industry declines in cost for high-capacity, non-volatile storage technologies. Testing was conducted on a small-scale cluster to prove the design, and then scaled at Los Alamos National Laboratory. Results show a definitive speedup factor for select workloads over writing directly to a typical global parallel file system; the Panasas ActiveScale File System.

97 MATHEMATICS AND COMPUTING↗

On the Feasibility of Dynamic Power Steering

While high performance has always been the primary constraint behind large-scale system design, future systems will be built with increasing energy efficiency in mind. Mechanisms such as fine-grained power scaling and gating will provide tools to system-software and application developers to ensure the most efficient use of tightly constrained power budgets. Such approaches to-date have been focused on node level optimizations to impact overall system energy efficiency. In this work we introduce Dynamic Power Steering, in which power can be dynamically routed across a system to resources where it will be of most benefit and away from other resources to maintain a near-constant overall power budget. This, a higher level algorithmic approach to improving energy efficiency, considers the whole extent of a system being used by an application. It can be used for applications in which there is load imbalance that varies over its execution. Using two classes of applications, namely those that contain a wave front type processing, and a particle-in-cell, we quantify the benefit of Dynamic Power Steering for a variety of workload characteristics and derive some insight into the ways in which workload behavior affect Power Steering applicability.

Barker, Kevin J.↗

GCoD: Graph Convolutional Network Acceleration via Dedicated Algorithm and Accelerator Co-Design

Graph Convolutional Networks (GCNs) have emerged as the state-of-the-art graph learning model. However, it remains notoriously challenging to inference GCNs over large graph datasets, limiting their application to large real-world graphs and hindering the exploration of deeper and more sophisticated GCN graphs. This is because real-world graphs can be extremely large and sparse. Furthermore, the node degree of GCNs tends to follow the power-law distribution and therefore have highly irregular adjacency matrices, resulting in prohibitive inefficiencies in both data processing and movement and thus substantially limiting the achievable GCN acceleration efficiency. To this end, this paper proposes the first GCN algorithm and accelerator Co-Design framework dubbed GCoD which can largely alleviate the aforementioned GCN irregularity and boost GCNs' inference efficiency. Specifically, on the algorithm level, GCoD integrates a divide and conquer GCN training strategy that polarizes the graphs to be either denser or sparser in local neighborhoods without compromising the model accuracy, resulting in graph adjacency matrices that (mostly) have merely two levels of workload and enjoys largely enhanced regularity and thus ease of acceleration. On the hardware level, we further develop a dedicated two-pronged accelerator with a separated engine to process each of the aforementioned workloads, further boosting the overall utilization and acceleration efficiency. Extensive experiments and ablation studies validate that our GCoD consistently outperforms state-of-the-art designs in terms of accelerator efficiency while maintaining or even improving the task accuracy. Additionally, we visualize GCoD trained graph adjacency matrices to better understand its advantages. All codes and pre-trained models will be released upon acceptance.

You, Haoran↗

JACC.shared: Leveraging HPC Metaprogramming and Performance Portability for Computations That Use Shared Memory GPUs

In this work, we present JACC.shared, a new feature of Julia for ACCelerators (JACC), which is the performanceportable and metaprogramming model of the just-in-time and LLVM-based Julia language. This new feature allows JACC applications to leverage the high-performance computing (HPC) capabilities of high-bandwidth, on-chip GPU memory. Historically, exploiting high-bandwidth, shared-memory GPUs has not been a priority for high-level programming solutions. JACC.shared covers that gap for the first time, thereby providing a highlevel, portable, and easy-to-use solution for programmers to exploit this memory and supporting all current major accelerator architectures. Well-known HPC and AI workloads, such as multi/hyperspectral imaging and AI convolutions, have been used to evaluate JACC.shared on two exascale GPU architectures hosted by some of the most powerful US Department of Energy supercomputers: Perlmutter (NVIDIA A100) and Frontier (AMD MI250X). The performance evaluation reports speedup of up to 3.5× by adding only one line of code to the base codes, thus providing important accelerators in a simple, portable, and transparent way and elevating the programming productivity and performance-portability capabilities for Julia/JACC HPC, AI, and scientific applications.

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗