Pub Date : 2023-06-26DOI: 10.1177/10943420231179662
Sabra Ossen, Jeremy Musser, Luke Dalessandro, M. Swany
Data volumes are exploding as sensors proliferate and become more capable. Edge computing is envisioned as a path to distribute processing and reduce latency. Many models of Edge computing consider small devices running conventional software. Our model includes a more lightweight execution engine for network microservices and a network scheduling framework to configure network processing elements to process streams and direct the appropriate traffic to them. In this article, we describe INDIANA, a complete framework for in-network microservices. We will describe how the two components-the INDIANA network Processing Element (InPE) and the Flange Network Operating System (NOS)-work together to achieve effective in-network processing to improve performance in edge to cloud environments. Our processing elements provide lightweight compute units optimized for efficient stream processing. These elements are customizable and vary in sophistication and resource consumption. The Flange NOS provides first-class flow based reasoning to drive function placement, network configuration, and load balancing that can respond dynamically to network conditions. We describe design considerations and discuss our approach and implementations. We evaluate the performance of stream processing and examine the performance of several exemplar applications on networks of increasing scale and complexity.
{"title":"INDIANA—In-Network Distributed Infrastructure for Advanced Network Applications","authors":"Sabra Ossen, Jeremy Musser, Luke Dalessandro, M. Swany","doi":"10.1177/10943420231179662","DOIUrl":"https://doi.org/10.1177/10943420231179662","url":null,"abstract":"Data volumes are exploding as sensors proliferate and become more capable. Edge computing is envisioned as a path to distribute processing and reduce latency. Many models of Edge computing consider small devices running conventional software. Our model includes a more lightweight execution engine for network microservices and a network scheduling framework to configure network processing elements to process streams and direct the appropriate traffic to them. In this article, we describe INDIANA, a complete framework for in-network microservices. We will describe how the two components-the INDIANA network Processing Element (InPE) and the Flange Network Operating System (NOS)-work together to achieve effective in-network processing to improve performance in edge to cloud environments. Our processing elements provide lightweight compute units optimized for efficient stream processing. These elements are customizable and vary in sophistication and resource consumption. The Flange NOS provides first-class flow based reasoning to drive function placement, network configuration, and load balancing that can respond dynamically to network conditions. We describe design considerations and discuss our approach and implementations. We evaluate the performance of stream processing and examine the performance of several exemplar applications on networks of increasing scale and complexity.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"442 - 461"},"PeriodicalIF":3.1,"publicationDate":"2023-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43385869","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-22DOI: 10.1177/10943420231178537
J. Vetter, Prasanna Date, F. Fahim, Shruti R. Kulkarni, P. Maksymovych, A. Talin, Marc Gonzalez Tallada, Pruek Vanna-Iampikul, Aaron R. Young, David Brooks, Yu Cao, Wei Gu-Yeon, S. Lim, Frank Liu, Matthew J. Marinella, B. Sumpter, Narasinga Rao Miniskar
The Abisko project aims to develop an energy-efficient spiking neural network (SNN) computing architecture and software system capable of autonomous learning and operation. The SNN architecture explores novel neuromorphic devices that are based on resistive-switching materials, such as memristors and electrochemical RAM. Equally important, Abisko uses a deep codesign approach to pursue this goal by engaging experts from across the entire range of disciplines: materials, devices and circuits, architectures and integration, software, and algorithms. The key objectives of our Abisko project are threefold. First, we are designing an energy-optimized high-performance neuromorphic accelerator based on SNNs. This architecture is being designed as a chiplet that can be deployed in contemporary computer architectures and we are investigating novel neuromorphic materials to improve its design. Second, we are concurrently developing a productive software stack for the neuromorphic accelerator that will also be portable to other architectures, such as field-programmable gate arrays and GPUs. Third, we are creating a new deep codesign methodology and framework for developing clear interfaces, requirements, and metrics between each level of abstraction to enable the system design to be explored and implemented interchangeably with execution, measurement, a model, or simulation. As a motivating application for this codesign effort, we target the use of SNNs for an analog event detector for a high-energy physics sensor.
{"title":"Abisko: Deep codesign of an architecture for spiking neural networks using novel neuromorphic materials","authors":"J. Vetter, Prasanna Date, F. Fahim, Shruti R. Kulkarni, P. Maksymovych, A. Talin, Marc Gonzalez Tallada, Pruek Vanna-Iampikul, Aaron R. Young, David Brooks, Yu Cao, Wei Gu-Yeon, S. Lim, Frank Liu, Matthew J. Marinella, B. Sumpter, Narasinga Rao Miniskar","doi":"10.1177/10943420231178537","DOIUrl":"https://doi.org/10.1177/10943420231178537","url":null,"abstract":"The Abisko project aims to develop an energy-efficient spiking neural network (SNN) computing architecture and software system capable of autonomous learning and operation. The SNN architecture explores novel neuromorphic devices that are based on resistive-switching materials, such as memristors and electrochemical RAM. Equally important, Abisko uses a deep codesign approach to pursue this goal by engaging experts from across the entire range of disciplines: materials, devices and circuits, architectures and integration, software, and algorithms. The key objectives of our Abisko project are threefold. First, we are designing an energy-optimized high-performance neuromorphic accelerator based on SNNs. This architecture is being designed as a chiplet that can be deployed in contemporary computer architectures and we are investigating novel neuromorphic materials to improve its design. Second, we are concurrently developing a productive software stack for the neuromorphic accelerator that will also be portable to other architectures, such as field-programmable gate arrays and GPUs. Third, we are creating a new deep codesign methodology and framework for developing clear interfaces, requirements, and metrics between each level of abstraction to enable the system design to be explored and implemented interchangeably with execution, measurement, a model, or simulation. As a motivating application for this codesign effort, we target the use of SNNs for an analog event detector for a high-energy physics sensor.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"351 - 379"},"PeriodicalIF":3.1,"publicationDate":"2023-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46219745","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-07DOI: 10.1177/10943420231179699
A. Tomás, E. S. Quintana‐Ortí, H. Anzt
We investigate the solution of low-rank matrix approximation problems using the truncated singular value decomposition (SVD). For this purpose, we develop and optimize graphics processing unit (GPU) implementations for the randomized SVD and a blocked variant of the Lanczos approach. Our work takes advantage of the fact that the two methods are composed of very similar linear algebra building blocks, which can be assembled using numerical kernels from existing high-performance linear algebra libraries. Furthermore, the experiments with several sparse matrices arising in representative real-world applications and synthetic dense test matrices reveal a performance advantage of the block Lanczos algorithm when targeting the same approximation accuracy.
{"title":"Fast truncated SVD of sparse and dense matrices on graphics processors","authors":"A. Tomás, E. S. Quintana‐Ortí, H. Anzt","doi":"10.1177/10943420231179699","DOIUrl":"https://doi.org/10.1177/10943420231179699","url":null,"abstract":"We investigate the solution of low-rank matrix approximation problems using the truncated singular value decomposition (SVD). For this purpose, we develop and optimize graphics processing unit (GPU) implementations for the randomized SVD and a blocked variant of the Lanczos approach. Our work takes advantage of the fact that the two methods are composed of very similar linear algebra building blocks, which can be assembled using numerical kernels from existing high-performance linear algebra libraries. Furthermore, the experiments with several sparse matrices arising in representative real-world applications and synthetic dense test matrices reveal a performance advantage of the block Lanczos algorithm when targeting the same approximation accuracy.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"380 - 393"},"PeriodicalIF":3.1,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"65398476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-05DOI: 10.1177/10943420231180504
Megan Hickman Fulp, Dakota Fulp, Changfeng Zou, Cooper Sanders, Ayan Biswas, Melissa C. Smith, Jon C. Calhoun
Due to improvements in high-performance computing (HPC) capabilities, many of today’s applications produce petabytes worth of data, causing bottlenecks within the system. Importance-based sampling methods, including our spatio-temporal hybrid data sampling method, are capable of resolving these bottlenecks. While our hybrid method has been shown to outperform existing methods, its effectiveness relies heavily on user parameters, such as histogram bins, error threshold, or number of regions. Moreover, the throughput it demonstrates must be higher to avoid becoming a bottleneck itself. In this article, we resolve both of these issues. First, we assess the effects of several user input parameters and detail techniques to help determine optimal parameters. Next, we detail and implement accelerated versions of our method using OpenMP and CUDA. Upon analyzing our implementations, we find 9.8× to 31.5× throughput improvements. Next, we demonstrate how our method can accept different base sampling algorithms and the effects these different algorithms have. Finally, we compare our sampling methods to the lossy compressor cuSZ in terms of data preservation and data movement.
{"title":"Accelerated dynamic data reduction using spatial and temporal properties","authors":"Megan Hickman Fulp, Dakota Fulp, Changfeng Zou, Cooper Sanders, Ayan Biswas, Melissa C. Smith, Jon C. Calhoun","doi":"10.1177/10943420231180504","DOIUrl":"https://doi.org/10.1177/10943420231180504","url":null,"abstract":"Due to improvements in high-performance computing (HPC) capabilities, many of today’s applications produce petabytes worth of data, causing bottlenecks within the system. Importance-based sampling methods, including our spatio-temporal hybrid data sampling method, are capable of resolving these bottlenecks. While our hybrid method has been shown to outperform existing methods, its effectiveness relies heavily on user parameters, such as histogram bins, error threshold, or number of regions. Moreover, the throughput it demonstrates must be higher to avoid becoming a bottleneck itself. In this article, we resolve both of these issues. First, we assess the effects of several user input parameters and detail techniques to help determine optimal parameters. Next, we detail and implement accelerated versions of our method using OpenMP and CUDA. Upon analyzing our implementations, we find 9.8× to 31.5× throughput improvements. Next, we demonstrate how our method can accept different base sampling algorithms and the effects these different algorithms have. Finally, we compare our sampling methods to the lossy compressor cuSZ in terms of data preservation and data movement.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"539 - 559"},"PeriodicalIF":3.1,"publicationDate":"2023-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48055498","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-31DOI: 10.1177/10943420231178552
N. Chalmers, Abhishek Mishra, Damon McDougall, T. Warburton
We present hipBone, an open-source performance-portable proxy application for the Nek5000 (and NekRS) computational fluid dynamics applications. HipBone is a fully GPU-accelerated C++ implementation of the original NekBone CPU proxy application with several novel algorithmic and implementation improvements which optimize its performance on modern fine-grain parallel GPU accelerators. Our optimizations include a conversion to store the degrees of freedom of the problem in assembled form in order to reduce the amount of data moved during the main iteration and a portable implementation of the main Poisson operator kernel. We demonstrate near-roofline performance of the operator kernel on three different modern GPU accelerators from two different vendors. We present a novel algorithm for splitting the application of the Poisson operator on GPUs which aggressively hides MPI communication required for both halo exchange and assembly. Our implementation of nearest-neighbor MPI communication then leverages several different routing algorithms and GPU-Direct RDMA capabilities, when available, which improves scalability of the benchmark. We demonstrate the performance of hipBone on three different clusters housed at Oak Ridge National Laboratory, namely, the Summit supercomputer and the Frontier early-access clusters, Spock and Crusher. Our tests demonstrate both portability across different clusters and very good scaling efficiency, especially on large problems.
{"title":"HipBone: A performance-portable graphics processing unit-accelerated C++ version of the NekBone benchmark","authors":"N. Chalmers, Abhishek Mishra, Damon McDougall, T. Warburton","doi":"10.1177/10943420231178552","DOIUrl":"https://doi.org/10.1177/10943420231178552","url":null,"abstract":"We present hipBone, an open-source performance-portable proxy application for the Nek5000 (and NekRS) computational fluid dynamics applications. HipBone is a fully GPU-accelerated C++ implementation of the original NekBone CPU proxy application with several novel algorithmic and implementation improvements which optimize its performance on modern fine-grain parallel GPU accelerators. Our optimizations include a conversion to store the degrees of freedom of the problem in assembled form in order to reduce the amount of data moved during the main iteration and a portable implementation of the main Poisson operator kernel. We demonstrate near-roofline performance of the operator kernel on three different modern GPU accelerators from two different vendors. We present a novel algorithm for splitting the application of the Poisson operator on GPUs which aggressively hides MPI communication required for both halo exchange and assembly. Our implementation of nearest-neighbor MPI communication then leverages several different routing algorithms and GPU-Direct RDMA capabilities, when available, which improves scalability of the benchmark. We demonstrate the performance of hipBone on three different clusters housed at Oak Ridge National Laboratory, namely, the Summit supercomputer and the Frontier early-access clusters, Spock and Crusher. Our tests demonstrate both portability across different clusters and very good scaling efficiency, especially on large problems.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"560 - 577"},"PeriodicalIF":3.1,"publicationDate":"2023-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42825171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-30DOI: 10.1177/10943420231172140
Hongwei Jin, Krishnan Raghavan, G. Papadimitriou, Cong Wang, A. Mandal, M. Kiran, E. Deelman, Prasanna Balaprakash
Identifying and addressing anomalies in complex, distributed systems can be challenging for reliable execution of scientific workflows. We model these workflows as directed acyclic graphs (DAGs), where the nodes and edges of the DAGs represent jobs and their dependencies, respectively. We develop graph neural networks (GNNs) to learn patterns in the DAGs and to detect anomalies at the node (job) and graph (workflow) levels. We investigate workflow-specific GNN models that are trained on a particular workflow and workflow-agnostic GNN models that are trained across the workflows. Our GNN models, which incorporate both individual job features and topological information from the workflow, show improved accuracy and efficiency compared to conventional learning methods for detecting anomalies. While joint trained with multiple scientific workflows, our GNN models reached an accuracy more than 80% for workflow level and 75% for job level anomalies. In addition, we illustrate the importance of hyperparameter tuning method in our study that can significantly improve the metric(s) measure of evaluating the GNN models. Finally, we integrate explainable GNN methods to provide insights on job features in the workflow that cause an anomaly.
{"title":"Graph neural networks for detecting anomalies in scientific workflows","authors":"Hongwei Jin, Krishnan Raghavan, G. Papadimitriou, Cong Wang, A. Mandal, M. Kiran, E. Deelman, Prasanna Balaprakash","doi":"10.1177/10943420231172140","DOIUrl":"https://doi.org/10.1177/10943420231172140","url":null,"abstract":"Identifying and addressing anomalies in complex, distributed systems can be challenging for reliable execution of scientific workflows. We model these workflows as directed acyclic graphs (DAGs), where the nodes and edges of the DAGs represent jobs and their dependencies, respectively. We develop graph neural networks (GNNs) to learn patterns in the DAGs and to detect anomalies at the node (job) and graph (workflow) levels. We investigate workflow-specific GNN models that are trained on a particular workflow and workflow-agnostic GNN models that are trained across the workflows. Our GNN models, which incorporate both individual job features and topological information from the workflow, show improved accuracy and efficiency compared to conventional learning methods for detecting anomalies. While joint trained with multiple scientific workflows, our GNN models reached an accuracy more than 80% for workflow level and 75% for job level anomalies. In addition, we illustrate the importance of hyperparameter tuning method in our study that can significantly improve the metric(s) measure of evaluating the GNN models. Finally, we integrate explainable GNN methods to provide insights on job features in the workflow that cause an anomaly.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"394 - 411"},"PeriodicalIF":3.1,"publicationDate":"2023-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46037147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-29DOI: 10.1177/10943420231176527
Iker Martín-Álvarez, J. Aliaga, M. Castillo, Sergio Iserte, R. Mayo
Malleability allows computing facilities to adapt their workloads through resource management systems to maximize the throughput of the facility and the efficiency of the executed jobs. This technique is based on reconfiguring a job to a different resource amount during execution and then continuing with it. One of the stages of malleability is the dynamic spawning of processes in execution time, where different decisions in this stage will affect how the next stage of data redistribution is performed, which is the most time-consuming stage. This paper describes different methods and strategies, defining eight different alternatives to spawn processes dynamically and indicates which one should be used depending on whether a strong or weak scaling application is being used. In addition, it is described for both types of applications which strategies benefit most the application performance or the system productivity. The results show that reducing the number of spawning processes by reusing the older ones can reduce reconfiguration time compared to the classical method by up to 2.6 times for expanding and up to 36 times for shrinking. Furthermore, the asynchronous strategy requires analysing the impact of oversubscription on application performance.
{"title":"Dynamic spawning of MPI processes applied to malleability","authors":"Iker Martín-Álvarez, J. Aliaga, M. Castillo, Sergio Iserte, R. Mayo","doi":"10.1177/10943420231176527","DOIUrl":"https://doi.org/10.1177/10943420231176527","url":null,"abstract":"Malleability allows computing facilities to adapt their workloads through resource management systems to maximize the throughput of the facility and the efficiency of the executed jobs. This technique is based on reconfiguring a job to a different resource amount during execution and then continuing with it. One of the stages of malleability is the dynamic spawning of processes in execution time, where different decisions in this stage will affect how the next stage of data redistribution is performed, which is the most time-consuming stage. This paper describes different methods and strategies, defining eight different alternatives to spawn processes dynamically and indicates which one should be used depending on whether a strong or weak scaling application is being used. In addition, it is described for both types of applications which strategies benefit most the application performance or the system productivity. The results show that reducing the number of spawning processes by reusing the older ones can reduce reconfiguration time compared to the classical method by up to 2.6 times for expanding and up to 36 times for shrinking. Furthermore, the asynchronous strategy requires analysing the impact of oversubscription on application performance.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2023-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48600798","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-24DOI: 10.1177/10943420231172978
Vladimir Ostapenco, L. Lefèvre, Anne-Cécile Orgerie, Benjamin Fichel
Data centers are very energy-intensive facilities that can generate various environmental impacts. Numerous energy, power, and environmental leverages exist and can help cloud providers and data center managers to reduce some of these impacts. But dealing with such heterogeneous leverages can be a challenging task that requires some support from a dedicated framework. This article presents a new approach for modeling, evaluating, and orchestrating a large set of technological and logistical leverages. Our framework is based on leverages modeling and Gantt chart leverages mapping. First experimental results based on selected scenarios show the pertinence of the proposed approach in terms of management facilities and potential impacts reduction.
{"title":"Modeling, evaluating, and orchestrating heterogeneous environmental leverages for large-scale data center management","authors":"Vladimir Ostapenco, L. Lefèvre, Anne-Cécile Orgerie, Benjamin Fichel","doi":"10.1177/10943420231172978","DOIUrl":"https://doi.org/10.1177/10943420231172978","url":null,"abstract":"Data centers are very energy-intensive facilities that can generate various environmental impacts. Numerous energy, power, and environmental leverages exist and can help cloud providers and data center managers to reduce some of these impacts. But dealing with such heterogeneous leverages can be a challenging task that requires some support from a dedicated framework. This article presents a new approach for modeling, evaluating, and orchestrating a large set of technological and logistical leverages. Our framework is based on leverages modeling and Gantt chart leverages mapping. First experimental results based on selected scenarios show the pertinence of the proposed approach in terms of management facilities and potential impacts reduction.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"328 - 350"},"PeriodicalIF":3.1,"publicationDate":"2023-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44246339","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-23DOI: 10.1177/10943420231175687
Olga Pearce, S. Brink
In this work, we develop novel data science methodologies for ensemble performance data that have the potential to uncover orders of magnitude of performance that is unknowingly being left on the table. Building on years of successful performance tool design and tool integration into million-line codes at Lawrence Livermore National Laboratory (Caliper (Boehme et al. 2016), Hatchet (Bhatele et al. 2019; Brink et al. 2020))—successes highlighted as key deliverables in meeting LLNL’s L1 and L2 milestones (Rieben and Weiss 2020)—we design a data science methodology for integrating multi-dimensional, multi-scale, multi-architecture, and multi-tool performance data, and provide data analytics and interactive visualization capabilities for further analysis and exploration of the data. Our work provides developers with a comprehensive multi-dimensional performance landscape, enabling enhanced capabilities for pinpointing performance bottlenecks on emerging hardware platforms composed of heterogeneous elements.
{"title":"Finding the forest in the trees: Enabling performance optimization on heterogeneous architectures through data science analysis of ensemble performance data","authors":"Olga Pearce, S. Brink","doi":"10.1177/10943420231175687","DOIUrl":"https://doi.org/10.1177/10943420231175687","url":null,"abstract":"In this work, we develop novel data science methodologies for ensemble performance data that have the potential to uncover orders of magnitude of performance that is unknowingly being left on the table. Building on years of successful performance tool design and tool integration into million-line codes at Lawrence Livermore National Laboratory (Caliper (Boehme et al. 2016), Hatchet (Bhatele et al. 2019; Brink et al. 2020))—successes highlighted as key deliverables in meeting LLNL’s L1 and L2 milestones (Rieben and Weiss 2020)—we design a data science methodology for integrating multi-dimensional, multi-scale, multi-architecture, and multi-tool performance data, and provide data analytics and interactive visualization capabilities for further analysis and exploration of the data. Our work provides developers with a comprehensive multi-dimensional performance landscape, enabling enhanced capabilities for pinpointing performance bottlenecks on emerging hardware platforms composed of heterogeneous elements.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"434 - 441"},"PeriodicalIF":3.1,"publicationDate":"2023-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47280730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-15DOI: 10.1177/10943420231175854
E. Jeannot, Guillaume Pallez, Nicolas Vidal
In high performance, computing concurrent applications are sharing the same file system. However, the bandwidth which provides access to the storage is limited. Therefore, too many I/O operations performed at the same time lead to conflicts and performance loss due to contention. This scenario will become more common as applications become more data intensive. To avoid congestion, job-schedulers have to play an important role in selecting which application run concurrently. However I/O-aware mapping strategies need to be simple, robust and fast. Hence, in this article, we discuss two plain and practical strategies to mitigate I/O congestion. They are based on the idea of scheduling I/O access so as not to exceed some prescribed I/O bandwidth. More precisely, we compare two approaches: one grouping applications into packs that will be run independently (i.e., pack-scheduling), the other one scheduling greedily applications using a predefined order (i.e. list-scheduling). Results show that performances depend heavily on the I/O load and the homogeneity of the underlying workload. Finally, we introduce the notion of characteristic time that represents information on the average time between consecutive I/O transfers. We show that it could be important to the design of schedulers and that we expect it to be easily obtained by analysis tools.
{"title":"IO-aware Job-Scheduling: Exploiting the Impacts of Workload Characterizations to select the Mapping Strategy","authors":"E. Jeannot, Guillaume Pallez, Nicolas Vidal","doi":"10.1177/10943420231175854","DOIUrl":"https://doi.org/10.1177/10943420231175854","url":null,"abstract":"In high performance, computing concurrent applications are sharing the same file system. However, the bandwidth which provides access to the storage is limited. Therefore, too many I/O operations performed at the same time lead to conflicts and performance loss due to contention. This scenario will become more common as applications become more data intensive. To avoid congestion, job-schedulers have to play an important role in selecting which application run concurrently. However I/O-aware mapping strategies need to be simple, robust and fast. Hence, in this article, we discuss two plain and practical strategies to mitigate I/O congestion. They are based on the idea of scheduling I/O access so as not to exceed some prescribed I/O bandwidth. More precisely, we compare two approaches: one grouping applications into packs that will be run independently (i.e., pack-scheduling), the other one scheduling greedily applications using a predefined order (i.e. list-scheduling). Results show that performances depend heavily on the I/O load and the homogeneity of the underlying workload. Finally, we introduce the notion of characteristic time that represents information on the average time between consecutive I/O transfers. We show that it could be important to the design of schedulers and that we expect it to be easily obtained by analysis tools.","PeriodicalId":54957,"journal":{"name":"International Journal of High Performance Computing Applications","volume":"37 1","pages":"213 - 228"},"PeriodicalIF":3.1,"publicationDate":"2023-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46118985","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：