Commercialization of User-Friendly Performance Analysis & Profiling Tools to Enable High Performance Computing (HPC) Applications on the ARM Processor Architecture
Explore the source record for details and available documents.
42 ENGINEERING↗
SEARCH · Engineering Papers
Results for “Performance analysis”
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.
PATHA: Performance Analysis Tool for HPC Applications
Large science projects rely on complex workflows to analyze terabytes or petabytes of data. These jobs are often running over thousands of CPU cores and simultaneously performing data accesses, data movements, and computation. It is difficult to identify bottlenecks or to debug the performance issues in these large workflows. In order to address these challenges, we have developed Performance Analysis Tool for HPC Applications (PATHA) using the state-of-art open source big data processing tools. Our framework can ingest system logs to extract key performance measures, and apply the most sophisticated statistical tools and data mining methods on the performance data. Furthermore, it utilizes an efficient data processing engine to allow users to interactively analyze a large amount of different types of logs and measurements. To illustrate the functionality of PATHA, we conduct a case study on the workflows from an astronomy project known as the Palomar Transient Factory (PTF). This study processed 1.6 TB of system logs collected on the NERSC supercomputer Edison. Using PATHA, we were able to identify performance bottlenecks, which reside in three tasks of PTF workflow with the dependency on the density of celestial objects.
96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
Slider Crank WEC Performance Analysis with Adaptive Autoregressive Filtering
This paper investigates a performance analysis of wave excitation force prediction to extract wave power for a slider crank power take-off system (PTOS) based on auto regressive (AR) filters. To efficiently convert wave energy into electricity, the prediction of wave excitation forces to keep the generator and the wave excitation force in sync is important for maximum energy extraction. The study shows a prediction methodology of half period and zero crossings in the practical scenario of irregular ocean waves. The prediction has been tested for different wave periods and with different filter orders. The prediction results have been used in the PTOS simulation to analyze the energy extraction. It has been shown that the prediction accuracy in the wave half period between the truth data and the predicted data drives the WEC energy extraction efficiency. The amplitude of the wave force is not used and hence the prediction deviation in the wave force amplitude does not affect the PTOS energy extraction. Further analysis shows that the optimum energy can be extracted at 15 th order filter with moderate prediction horizon length.
autoregressive filter↗
Performance Analysis of Large Scale HPC Workflows for Earth System Models
Climate assessment reports, such as the Intergovernmental Panel on Climate Change’s 5th assesment report [IPCC AR5; IPCC et al., 2014] or the United States National Climate Assessment [U.S. Global Change Research Program, 2014], increasingly rely on many-ensemble climate projections from fully-coupled Earth system models (ESM) to asses possible future climate states. ESM participate in a large set of Coupled Model Intercomparson Projects (CMIP) to better understand their models and the climate system, so that they may provide credible climate projections. However, fully participating in these intercomparison exercises represents an extremely large computational personnel burden for modeling centers [Eyring et al., 2016, Williams et al., 2016a]. However, there has been little attention within the climate modeling community paid to the in situ performance of ESMs up until now. Recently a computational performance model intercomparison project (CPMIP) for CMIP6 has begun [Balaji et al., 2017], but there has not yet been an effort to gather, process, and analyze the historical performance data that has been saved within archived log files from previous modeling efforts, nor the development of performance analysis system capable of providing detailed statistical analyses or visualizations of this performance data. Therefore, we’ve begun a joint effort to analyze the historical and current performance data on both GFDL and OLCF modeling systems. Here, we have developed a preliminary performance analysis system to captue, analyize, and visualize end-to-end in situ performance data and applied it to NOAA/GFDL’s post-processing system. Through the course of the project a number of performance improvements have been made, with the steps necessary to handle the expected volume of data (Big Data) likely to come out of the CMIP6 exercises (TB to PB) provided. We have also begun developing a statistical workload model that will allow modeling centers to determined their expected scientific throughput in the context of expected resource (computational and personnel) availability and accounting for typical workflow disruptions. This will allow them to develop a CMIP participation strategy and maximize their level of participation in the exercises. We expect this infrastructure, and the data it produces, to ultimately allow us to identify performance issues and workflow errors that exist across Earth system modeling centers, and possibly HPC modeling centers in general. Center-independent solutions to workflow or performance issues would help alleviate the computational burden for CMIP6+ modeling projects and should realize both labor and cost savings benefits throughout the CMIP community.
54 ENVIRONMENTAL SCIENCES↗
SYMBIOSYS: A Methodology for Performance Analysis of Composable HPC Data Services
Microservices are a powerful new way of building, customizing, and deploying distributed services owing to their flexibility and maintainability. Several large-scale distributed platforms have emerged to serve the growing needs of data-centric workloads and services in commercial computing. Concurrently, high-performance computing (HPC) systems and software are rapidly evolving to meet the demands of diversified applications and heterogeneity. The interplay of hardware factors, software configuration parameters, and the flexibility offered with a microservice architecture makes it nontrivial to estimate the optimal service instantiation for a given application workload. Further, this problem is exacerbated when considering that these services operate in a dynamic and heterogeneous HPC environment. An optimally integrated service can be vastly more performant than a haphazardly integrated one. Existing performance tools for HPC either fail to understand the request-response model of communication inherent to microservices or they operate within a narrow scope, limiting the insight that can be gleaned from employing them in isolation. We propose a methodology for integrated performance analysis of HPC microservices frameworks and applications called SYMBIOSYS. We describe its design and implementation within the context of the Mochi framework. This integration is achieved by combining distributed callpath profiling and tracing with a performance data exchange strategy that collects fine-grained, low-level metrics from the RPC communication library and network layers. The result is a portable, low-overhead performance analysis setup that provides a holistic profile of the dependencies among microservices and how they interact with the Mochi RPC software stack. Using HEPnOS, a production-quality Mochi data service, we demonstrate the low-overhead operation of SYMBIOSYS at scale and use it to identify the root causes of poorly performing service configurations.
microservices↗
Performance analysis and data reduction for exascale scientific workflows
Chimbuko is the first in situ, scalable, workflow-level performance analysis tool for trace-level analysis and visualization of application performance. This tool was developed by the Co-design Center for Online Data Analysis and Reduction and funded by the U.S. Department of Energy’s Exascale Computing Project. We provide a detailed description of Chimbuko’s architecture and illustrate our online and offline visualization with multiple use cases. We also present results for the deployment and scalability of the tool as applied to a high-energy physics workflow running at large scale on the Frontier supercomputer.
97 MATHEMATICS AND COMPUTING↗
Thermal performance analysis of optimized hexagonal finned heat sinks in impinging air jet
In this study, thermal performance analysis of hexagonal finned heat sinks which optimized according to the experimental design and optimization method of Taguchi were investigated. Experiments of air jet impingement on heated hexagonal finned heat sinks were carried out adhering to the L{sub 18}(2{sup 1*}3{sup 6}) orthogonal array test plan. Optimum geometries were determined and named OH-1, OH-2. Enhancement efficiency with the first law of thermodynamics was analyzed for optimized heat sinks with 100, 150, 200 mm heights of hexagonal fin. Nusselt correlations were found out and variations of enhancement efficiency with Reynolds number presented in η–Re graphics.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Performance Analysis, Modeling and Scaling of HPC Applications and Tools
E cient use of supercomputers at DOE centers is vital for maximizing system throughput, mini- mizing energy costs and enabling science breakthroughs faster. This requires complementary e orts along several directions to optimize the performance of scienti c simulation codes and the under- lying runtimes and software stacks. This in turn requires providing scalable performance analysis tools and modeling techniques that can provide feedback to physicists and computer scientists developing the simulation codes and runtimes respectively. The PAMS project is using time allocations on supercomputers at ALCF, NERSC and OLCF to further the goals described above by performing research along the following fronts: 1. Scaling Study of HPC applications; 2. Evaluation of Programming Models; 3. Hardening of Performance Tools; 4. Performance Modeling of Irregular Codes; and 5. Statistical Analysis of Historical Performance Data. We are a team of computer and computational scientists funded by both DOE/NNSA and DOE/ ASCR programs such as ECRP, XStack (Traleika Glacier, PIPER), ExaOSR (ARGO), SDMAV II (MONA) and PSAAP II (XPACC). This allocation will enable us to study big data issues when analyzing performance on leadership computing class systems and to assist the HPC community in making the most e ective use of these resources.
97 MATHEMATICS AND COMPUTING↗
A Thermal Model for the Performance Analysis of Large Scale Spent Fuel Cooling Pond - 19173
This paper presents a thermal model for the performance analysis of a large scale cooling pond for the long term storage of spent nuclear fuel. The model has been developed using the knowledge gained from fundamental investigation using CFD and has been validated by comparing against reliable experimental and/or numerical data. It was then applied in the context of operation of a large scale pond to ascertain the sensitivity of various operational and environmental conditions. To develop the model, the water body along with the heating elements was considered as one zone and humid air inside the building as another zone, both of which are considered as well-mixed zones. Various input conditions namely, total heat load from the heat source, environmental temperature and relative humidity, ventilation rate, efficiency of the cooling tower, recirculation and make-up water flow rates have been programmed in the model. The statistical method AVONA has been used to assess the influence of the above variables. The results reveal that the efficiency of the cooling tower is the most influential parameter on the cooling performance of cooling pond within the considered values of heat load. It has also been found that the majority of the heat is taken by the recirculating water for very high heat load while the other modes of heat transfer are small but not insignificant. The model is capable of simulating the scenario of loss of cooling water and suggests that it can be used for managing a large pond efficiently as well as recovery from an accident situation. (authors)
12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
Transient fuel performance analysis for the preliminary fuel concept of general atomics fast modular reactor
This study investigates the transient fuel performance of General Atomics Fast Modular Reactor (GA-FMR) during accident scenarios, focusing on the behavior of its innovative fuel system that combines high-assay low enriched uranium dioxide (HALEUO2) fuel with SiGA® ceramic matrix composite silicon carbide cladding. The preliminary fuel design’s response was analyzed during reactivity-initiated accidents (RIA) and loss of coolant accidents (LOCA) using BISON fuel performance analysis code, which included both the diffusion enhanced and BISON-FASTGRASS coupled UO 2 models. The RIA analysis demonstrated that effective reactivity control reduced fuel temperature, though with transient fission gas release resulting in additional tensile stress state on the cladding. LOCA simulations revealed differing predictions between the two models: the BISON UO 2 model showed more transient fission gas release but minimal pellet expansion, while the BISON-FASTGRASS UO 2 model predicted less pronounced fission gas release but more fuel swelling and thermal expansion, potentially leading to pellet-cladding mechanical interaction. Here, these findings highlight critical areas for fuel design optimization and identify knowledge gaps requiring further experimental and computational investigation to advance GA-FMR fuel development.
Lee, Soon K. [Argonne National Laboratory (ANL), A↗
Performance Analysis of PIConGPU: Particle-in-Cell on GPUs using NVIDIA’s NSight Systems and NSight Compute
PIConGPU, Particle In Cell on GPUs, is an open source simulations framework for plasma and laser-plasma physics used to develop advanced particle accelerators for radiation therapy of cancer, high energy physics and photon science. While PIConGPU has been optimized for at least 5 years to run well on NVIDIA GPU-based clusters, there has been limited exploration by the development team of potential scalability bottlenecks using recently updated and new tools including NVIDIA’s NVProf tool and the brand-new NVIDIA NSight Suite (Systems and Compute) tools. PIConGPU is a highly optimized application that runs production jobs at scale on a system Oak Ridge Leadership Facility’s (OLCF) Summit supercomputer (using the full machine at 4600 nodes; at 98% of GPU utilization on all ~28000 NVIDIA Volta GPUs). PIConGPU has been selected as one of the the eight applications for OLCF’s coveted Center for Accelerated Application Readiness (CAAR) program aimed at the facility’s Frontier supercomputer (OLCF’s first exascale system to launch in 2021), to partner with our vendors (primary vendors: AMD and Cray/HPE) ensuring that Frontier will be able to perform large-scale science when it opens to users in 2022. To this effect, performance engineers on the PIConGPU team wanted to dive deep into the application to understand at the finest granularity, which portions of the code could be further optimized to exploit the hardware on Summit at it’s maximum potential and also to elucidate which key kernels should be tracked and optimized for the CAAR effort to port this code to Frontier. Any bottlenecks that are observed via performance profiling on Summit are likely to also impact scalability on the Frontier-dev system and the Frontier Early Access (EA) system. Additionally, the engineers wanted to take a closer look at the newest NVIDIA profiling tools which allows us to identify the most useful features on these tools and will provide an opportunity to compare it to new AMD and Cray’s performance analysis tool releases and provide feedback to our vendor partners on what features are most important and mission critical for CAAR efforts. The primary goal of this report is to focus on the evaluation of PIConGPU’s most time-intensive kernels using NVProf and NSight Suite. Three kernels, Current Deposition (also known as Compute Current), Particle Push (Move and Mark), and Shift Particles are known to be some of the most time-consuming kernels in PIConGPU. The Current xi Deposition kernel and Particle Push kernel both set up the particle attributes for running any physics simulation with PIConGPU, so it is crucial to improve the performance of these two kernels. In this report, we measure single GPU metrics for the three kernels, offer high level takeaways from the conducted analysis, and compare the profiling data from NSight Compute to that of NVProf. This analysis was performed using a grid size of 240 x 272 x 224, and 10 time steps with the Mid-November Figure of Merit (FOM) run setup. The Traveling Wave Electron Acceleration (TWEAC) science case used in this run is a representative science case for PIConGPU. This execution can also be used for baseline analysis on AMD MI50/ MI60 systems. As of the time of writing, the PIConGPU application has limited use for features of NSight Systems, so this report will mainly focus on insights garnered from NSight Compute. For this analysis, we run the “full” metric set available in NSight Compute version 2020.1.2 and use NSight Systems version 2020.3.1 to generate the application timeline.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Unveiling Temporal Performance Deviation: Leveraging Clustering in Microservices Performance Analysis
As the market for cloud computing continues to grow, an increasing number of users are deploying applications as microservices. The shift introduces unique challenges in identifying and addressing performance issues, particularly within large and complex infrastructures. To address this challenge, we propose a methodology that unveils temporal performance deviations in microservices by clustering containers based on their performance characteristics at different time intervals. Showcasing our methodology on the Alibaba dataset, we found both stable and dynamic performance patterns, providing a valuable tool for enhancing overall performance and reliability in modern application landscapes.
Clustering↗
CIGS Performance Analysis: Alternative Method to Fit the Sandia Array Performance Model.
Abstract not provided.
King, Bruce Hardison↗
Code artifact for "C.A.T.S.: Memory and Control Flow Tracing for Whole-Program Performance Analysis"
This software contains the code artifacts for the Supercomputing 2025 paper titled "C.A.T.S.: Memory and Control Flow Tracing for Whole-Program Performance Analysis". The included code contains analysis passes, dynamic program instrumentation, runtime, examples, and tests.
Tal, Ben Nun [Lawrence Livermore National Laborato↗
User-Centric Approach for Benchmark RDF Data Generator in Big Data Performance Analysis
This research focuses on user-centric approach of building such tools and proposes a flexible, extensible, and easy to use framework to support performance analysis of Big Data systems. Finally, case studies from two different domains are presented to validate the framework.
Purohit, Sumit↗
Application of Performance Analysis Tools on SNL ASC Codes
This milestone 1) exercised a broad set of performance profiling and analysis tools, including tools whose development has been promoted by the ASC program; 2) exercised the tools on two different SNL ASC codes, one Sierra code (Sierra/Aria, a C++ codebase) and one RAMSES code (ITS, a Fortran codebase); and 3) exercised the tools on multiple platforms, including the CTS-1 (e.g., Serrano) and ATS-1 Trinity (e.g., Mutrino) platforms. The milestone generated a plethora of strong and weak scaling, trend and profile data for multiple versions and problem cases for each of the two codes. A wealth of experience was gained with the various tools that included identification of problems, an improved understanding of feature sets, enhanced usage documentation, and insights for future tool-development. Results are provided from a large number and variety of performance analysis runs with the target codes, together with instructions for how to make use of the tools with the codes.
97 MATHEMATICS AND COMPUTING↗
Performance Analysis of Traditional and Data-Parallel Primitive Implementations of Visualization and Analysis Kernels
Measurements of absolute runtime are useful as a summary of performance when studying parallel visualization and analysis methods on computational platforms of increasing concurrency and complexity. We can obtain even more insights by measuring and examining more detailed measures from hardware performance counters, such as the number of instructions executed by an algorithm implemented in a particular way, the amount of data moved to/from memory, memory hierarchy utilization levels via cache hit/miss ratios, and so forth. This work focuses on performance analysis on modern multi-core platforms of three different visualization and analysis kernels that are implemented in different ways: one is "traditional", using combinations of C++ and VTK, and the other uses a data-parallel approach using VTK-m. Our performance study consists of measurement and reporting of several different hardware performance counters on two different multi-core CPU platforms. The results reveal interesting performance differences between these two different approaches for implementing these kernels, results that would not be apparent using runtime as the only metric.
97 MATHEMATICS AND COMPUTING↗
New machine protection system at the Spallation Neutron Source – design process and performance analysis
A New Machine Protection System (MPS) at the Spallation Neutron Source (SNS) was developed and implemented on µTCA-based hardware platforms. The system monitors more than 2500 field inputs and shuts off the beam within 10 µs if adverse events occur. We will present system level design process of various firmware and software components as well as the system integration into EPICS environment. The performance analysis of the MPS after two SNS run cycles will also be presented.
Bobrek, Miljko [ORNL] (ORCID:0000000332763451)↗