In many science areas where datasets need to be transferred or shared, rapid growth in dataset size, coupled with much slower increases in wide area data transfer bandwidths, is making it extremely hard for scientists to analyze the data. This paper addresses the current limitations by developing SDQuery DSI, a GridFTP plug-in that supports flexible server-side data subsetting. An existing GridFTP server is able to dynamically load this tool to support new functionality. Different queries types (query over dimensions, coordinates and values) are supported by our tool. A number of optimizations, like parallel indexing, performance model for data subsetting, and parallel streaming are also applied. We compare our SDQuery DSI with GridFTP default File DSI in different network environments, and show that our method can achieve better efficiency in almost all cases.
{"title":"SDQuery DSI: Integrating data management support with a wide area data transfer protocol","authors":"Yunde Su, Yi Wang, G. Agrawal, R. Kettimuthu","doi":"10.1145/2503210.2503270","DOIUrl":"https://doi.org/10.1145/2503210.2503270","url":null,"abstract":"In many science areas where datasets need to be transferred or shared, rapid growth in dataset size, coupled with much slower increases in wide area data transfer bandwidths, is making it extremely hard for scientists to analyze the data. This paper addresses the current limitations by developing SDQuery DSI, a GridFTP plug-in that supports flexible server-side data subsetting. An existing GridFTP server is able to dynamically load this tool to support new functionality. Different queries types (query over dimensions, coordinates and values) are supported by our tool. A number of optimizations, like parallel indexing, performance model for data subsetting, and parallel streaming are also applied. We compare our SDQuery DSI with GridFTP default File DSI in different network environments, and show that our method can achieve better efficiency in almost all cases.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123662739","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Bhatele, K. Mohror, S. Langer, Katherine E. Isaacs
Predictable performance is important for understanding and alleviating application performance issues; quantifying the effects of source code, compiler, or system software changes; estimating the time required for batch jobs; and determining the allocation requests for proposals. Our experiments show that on a Cray XE system, the execution time of a communication-heavy parallel application ranges from 28% faster to 41% slower than the average observed performance. Blue Gene systems, on the other hand, demonstrate no noticeable run-to-run variability. In this paper, we focus on Cray machines and investigate potential causes for performance variability such as OS jitter, shape of the allocated partition, and interference from other jobs sharing the same network links. Reducing such variability could improve overall throughput at a computer center and save energy costs.
{"title":"There goes the neighborhood: Performance degradation due to nearby jobs","authors":"A. Bhatele, K. Mohror, S. Langer, Katherine E. Isaacs","doi":"10.1145/2503210.2503247","DOIUrl":"https://doi.org/10.1145/2503210.2503247","url":null,"abstract":"Predictable performance is important for understanding and alleviating application performance issues; quantifying the effects of source code, compiler, or system software changes; estimating the time required for batch jobs; and determining the allocation requests for proposals. Our experiments show that on a Cray XE system, the execution time of a communication-heavy parallel application ranges from 28% faster to 41% slower than the average observed performance. Blue Gene systems, on the other hand, demonstrate no noticeable run-to-run variability. In this paper, we focus on Cray machines and investigate potential causes for performance variability such as OS jitter, shape of the allocated partition, and interference from other jobs sharing the same network links. Reducing such variability could improve overall throughput at a computer center and save energy costs.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114932355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
P. Boyle, M. Buchoff, N. Christ, T. Izubuchi, C. Jung, T. Luu, R. Mawhinney, C. Schroeder, R. Soltz, P. Vranas, J. Wasem
The origin of mass is one of the deepest mysteries in science. Neutrons and protons, which account for almost all visible mass in the Universe, emerged from a primordial plasma through a cataclysmic phase transition microseconds after the Big Bang. However, most mass in the Universe is invisible. The existence of dark matter, which interacts with our world so weakly that it is essentially undetectable, has been established from its galactic-scale gravitational effects. Here we describe results from the first truly physical calculations of the cosmic phase transition and a groundbreaking first-principles investigation into composite dark matter, studies impossible with previous state-of-the-art methods and resources. By inventing a powerful new algorithm, “DSDR,” and implementing it effectively for contemporary supercomputers, we attain excellent strong scaling, perfect weak scaling to the LLNL BlueGene/Q two million cores, sustained speed of 7.2 petaflops, and time-to-solution speedup of more than 200 over the previous state-of-the-art.
{"title":"The origin of mass","authors":"P. Boyle, M. Buchoff, N. Christ, T. Izubuchi, C. Jung, T. Luu, R. Mawhinney, C. Schroeder, R. Soltz, P. Vranas, J. Wasem","doi":"10.1145/2503210.2504561","DOIUrl":"https://doi.org/10.1145/2503210.2504561","url":null,"abstract":"The origin of mass is one of the deepest mysteries in science. Neutrons and protons, which account for almost all visible mass in the Universe, emerged from a primordial plasma through a cataclysmic phase transition microseconds after the Big Bang. However, most mass in the Universe is invisible. The existence of dark matter, which interacts with our world so weakly that it is essentially undetectable, has been established from its galactic-scale gravitational effects. Here we describe results from the first truly physical calculations of the cosmic phase transition and a groundbreaking first-principles investigation into composite dark matter, studies impossible with previous state-of-the-art methods and resources. By inventing a powerful new algorithm, “DSDR,” and implementing it effectively for contemporary supercomputers, we attain excellent strong scaling, perfect weak scaling to the LLNL BlueGene/Q two million cores, sustained speed of 7.2 petaflops, and time-to-solution speedup of more than 200 over the previous state-of-the-art.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123505187","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Long-term execution of scientific applications often leads to dynamic workloads and varying application requirements. When the execution uses resources provisioned from IaaS clouds, and thus consumption-related payment, efficient and online scheduling algorithms must be found. Portfolio scheduling, which selects dynamically a suitable policy from a broad portfolio, may provide a solution to this problem. However, selecting online the right policy from possibly tens of alternatives remains challenging. In this work, we introduce an abstract model to explore this selection problem. Based on the model, we present a comprehensive portfolio scheduler that includes tens of provisioning and allocation policies. We propose an algorithm that can enlarge the chance of selecting the best policy in limited time, possibly online. Through trace-based simulation, we evaluate various aspects of our portfolio scheduler, and find performance improvements from 7% to 100% in comparison with the best constituent policies and high improvement for bursty workloads.
{"title":"Exploring portfolio scheduling for long-term execution of scientific workloads in IaaS clouds","authors":"Kefeng Deng, Junqiang Song, Kaijun Ren, A. Iosup","doi":"10.1145/2503210.2503244","DOIUrl":"https://doi.org/10.1145/2503210.2503244","url":null,"abstract":"Long-term execution of scientific applications often leads to dynamic workloads and varying application requirements. When the execution uses resources provisioned from IaaS clouds, and thus consumption-related payment, efficient and online scheduling algorithms must be found. Portfolio scheduling, which selects dynamically a suitable policy from a broad portfolio, may provide a solution to this problem. However, selecting online the right policy from possibly tens of alternatives remains challenging. In this work, we introduce an abstract model to explore this selection problem. Based on the model, we present a comprehensive portfolio scheduler that includes tens of provisioning and allocation policies. We propose an algorithm that can enlarge the chance of selecting the best policy in limited time, possibly online. Through trace-based simulation, we evaluate various aspects of our portfolio scheduler, and find performance improvements from 7% to 100% in comparison with the best constituent policies and high improvement for bursty workloads.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"120 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124738016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Scalable parallel computing is essential for processing large scale-free (power-law) graphs. The distribution of data across processes becomes important on distributed-memory computers with thousands of cores. It has been shown that two-dimensional layouts (edge partitioning) can have significant advantages over traditional one-dimensional layouts. However, simple 2D block distribution does not use the structure of the graph, and more advanced 2D partitioning methods are too expensive for large graphs. We propose a new two-dimensional partitioning algorithm that combines graph partitioning with 2D block distribution. The computational cost of the algorithm is essentially the same as 1D graph partitioning. We study the performance of sparse matrix-vector multiplication (SpMV) for scale-free graphs from the web and social networks using several different partitioners and both 1D and 2D data layouts. We show that SpMV run time is reduced by exploiting the graph's structure. Contrary to popular belief, we observe that current graph and hypergraph partitioners often yield relatively good partitions on scale-free graphs. We demonstrate that our new 2D partitioning method consistently outperforms the other methods considered, for both SpMV and an eigensolver, on matrices with up to 1.6 billion nonzeros using up to 16,384 cores.
{"title":"Scalable matrix computations on large scale-free graphs using 2D graph partitioning","authors":"E. Boman, K. Devine, S. Rajamanickam","doi":"10.1145/2503210.2503293","DOIUrl":"https://doi.org/10.1145/2503210.2503293","url":null,"abstract":"Scalable parallel computing is essential for processing large scale-free (power-law) graphs. The distribution of data across processes becomes important on distributed-memory computers with thousands of cores. It has been shown that two-dimensional layouts (edge partitioning) can have significant advantages over traditional one-dimensional layouts. However, simple 2D block distribution does not use the structure of the graph, and more advanced 2D partitioning methods are too expensive for large graphs. We propose a new two-dimensional partitioning algorithm that combines graph partitioning with 2D block distribution. The computational cost of the algorithm is essentially the same as 1D graph partitioning. We study the performance of sparse matrix-vector multiplication (SpMV) for scale-free graphs from the web and social networks using several different partitioners and both 1D and 2D data layouts. We show that SpMV run time is reduced by exploiting the graph's structure. Contrary to popular belief, we observe that current graph and hypergraph partitioners often yield relatively good partitions on scale-free graphs. We demonstrate that our new 2D partitioning method consistently outperforms the other methods considered, for both SpMV and an eigensolver, on matrices with up to 1.6 billion nonzeros using up to 16,384 cores.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125215027","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Thomas Ropars, Tatiana V. Martsinkevich, Amina Guermouche, A. Schiper, F. Cappello
The high failure rate expected for future supercomputers requires the design of new fault tolerant solutions. Most checkpointing protocols are designed to work with any message-passing application but sudder from scalability issues at extreme scale. We take a different approach: We identify a property common to many HPC applications, namely channel-determinism, and introduce a new partial order relation, called always-happens-before relation, between events of such applications. Leveraging these two concepts, we design a protocol that combines an unprecedented set of features. Our protocol called SPBC combines in a hierarchical way coordinated checkpointing and message logging. It is the first protocol that provides failure containment without logging any information reliably apart from process checkpoints, and this, without penalizing recovery performance. Experiments run with a representative set of HPC workloads demonstrate a good performance of our protocol during both, failure-free execution and recovery.
{"title":"SPBC: Leveraging the characteristics of MPI HPC applications for scalable checkpointing","authors":"Thomas Ropars, Tatiana V. Martsinkevich, Amina Guermouche, A. Schiper, F. Cappello","doi":"10.1145/2503210.2503271","DOIUrl":"https://doi.org/10.1145/2503210.2503271","url":null,"abstract":"The high failure rate expected for future supercomputers requires the design of new fault tolerant solutions. Most checkpointing protocols are designed to work with any message-passing application but sudder from scalability issues at extreme scale. We take a different approach: We identify a property common to many HPC applications, namely channel-determinism, and introduce a new partial order relation, called always-happens-before relation, between events of such applications. Leveraging these two concepts, we design a protocol that combines an unprecedented set of features. Our protocol called SPBC combines in a hierarchical way coordinated checkpointing and message logging. It is the first protocol that provides failure containment without logging any information reliably apart from process checkpoints, and this, without penalizing recovery performance. Experiments run with a representative set of HPC workloads demonstrate a good performance of our protocol during both, failure-free execution and recovery.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126274232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Germán Llort, Harald Servat, Juan Gonzalez, Judit Giménez, Jesús Labarta
Understanding the behavior of a parallel application is crucial if we are to tune it to achieve its maximum performance. Yet the behavior the application exhibits may change over time and depend on the actual execution scenario: particular inputs and program settings, the number of processes used, or hardware-specific problems. So beyond the details of a single experiment a far more interesting question arises: how does the application behavior respond to changes in the execution conditions? In this paper, we demonstrate that object tracking concepts from computer vision have huge potential to be applied in the context of performance analysis. We leverage tracking techniques to analyze how the behavior of a parallel application evolves through multiple scenarios where the execution conditions change. This method provides comprehensible insights on the influence of different parameters on the application behavior, enabling us to identify the most relevant code regions and their performance trends.
{"title":"On the usefulness of object tracking techniques in performance analysis","authors":"Germán Llort, Harald Servat, Juan Gonzalez, Judit Giménez, Jesús Labarta","doi":"10.1145/2503210.2503267","DOIUrl":"https://doi.org/10.1145/2503210.2503267","url":null,"abstract":"Understanding the behavior of a parallel application is crucial if we are to tune it to achieve its maximum performance. Yet the behavior the application exhibits may change over time and depend on the actual execution scenario: particular inputs and program settings, the number of processes used, or hardware-specific problems. So beyond the details of a single experiment a far more interesting question arises: how does the application behavior respond to changes in the execution conditions? In this paper, we demonstrate that object tracking concepts from computer vision have huge potential to be applied in the context of performance analysis. We leverage tracking techniques to analyze how the behavior of a parallel application evolves through multiple scenarios where the execution conditions change. This method provides comprehensible insights on the influence of different parameters on the application behavior, enabling us to identify the most relevant code regions and their performance trends.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"430 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120875476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Modern interconnects offer remote direct memory access (RDMA) features. Yet, most applications rely on explicit message passing for communications albeit their unwanted overheads. The MPI-3.0 standard defines a programming interface for exploiting RDMA networks directly, however, it's scalability and practicability has to be demonstrated in practice. In this work, we develop scalable bufferless protocols that implement the MPI-3.0 specification. Our protocols support scaling to millions of cores with negligible memory consumption while providing highest performance and minimal overheads. To arm programmers, we provide a spectrum of performance models for all critical functions and demonstrate the usability of our library and models with several application studies with up to half a million processes. We show that our design is comparable to, or better than UPC and Fortran Coarrays in terms of latency, bandwidth, and message rate. We also demonstrate application performance improvements with comparable programming complexity.
{"title":"Enabling highly-scalable remote memory access programming with MPI-3 one sided","authors":"R. Gerstenberger, Maciej Besta, T. Hoefler","doi":"10.1145/2503210.2503286","DOIUrl":"https://doi.org/10.1145/2503210.2503286","url":null,"abstract":"Modern interconnects offer remote direct memory access (RDMA) features. Yet, most applications rely on explicit message passing for communications albeit their unwanted overheads. The MPI-3.0 standard defines a programming interface for exploiting RDMA networks directly, however, it's scalability and practicability has to be demonstrated in practice. In this work, we develop scalable bufferless protocols that implement the MPI-3.0 specification. Our protocols support scaling to millions of cores with negligible memory consumption while providing highest performance and minimal overheads. To arm programmers, we provide a spectrum of performance models for all critical functions and demonstrate the usability of our library and models with several application studies with up to half a million processes. We show that our design is comparable to, or better than UPC and Fortran Coarrays in terms of latency, bandwidth, and message rate. We also demonstrate application performance improvements with comparable programming complexity.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131481407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
It is difficult to manually identify opportunities for enhancing data locality. To address this problem, we extended the HPCToolkit performance tools to support data-centric profiling of scalable parallel programs. Our tool uses hardware counters to directly measure memory access latency and attributes latency metrics to both variables and instructions. Different hardware counters provide insight into different aspects of data locality (or lack thereof). Unlike prior tools for data-centric analysis, our tool employs scalable measurement, analysis, and presentation methods that enable it to analyze the memory access behavior of scalable parallel programs with low runtime and space overhead. We demonstrate the utility of HPCToolkit's new data-centric analysis capabilities with case studies of five well-known benchmarks. In each benchmark, we identify performance bottlenecks caused by poor data locality and demonstrate non-trivial performance optimizations enabled by this guidance.
{"title":"A data-centric profiler for parallel programs","authors":"Xu Liu, J. Mellor-Crummey","doi":"10.1145/2503210.2503297","DOIUrl":"https://doi.org/10.1145/2503210.2503297","url":null,"abstract":"It is difficult to manually identify opportunities for enhancing data locality. To address this problem, we extended the HPCToolkit performance tools to support data-centric profiling of scalable parallel programs. Our tool uses hardware counters to directly measure memory access latency and attributes latency metrics to both variables and instructions. Different hardware counters provide insight into different aspects of data locality (or lack thereof). Unlike prior tools for data-centric analysis, our tool employs scalable measurement, analysis, and presentation methods that enable it to analyze the memory access behavior of scalable parallel programs with low runtime and space overhead. We demonstrate the utility of HPCToolkit's new data-centric analysis capabilities with case studies of five well-known benchmarks. In each benchmark, we identify performance bottlenecks caused by poor data locality and demonstrate non-trivial performance optimizations enabled by this guidance.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"100 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131640436","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Marc Gamell, I. Rodero, M. Parashar, Janine Bennett, H. Kolla, Jacqueline H. Chen, P. Bremer, Aaditya G. Landge, A. Gyulassy, P. McCormick, S. Pakin, Valerio Pascucci, S. Klasky
As scientific applications target exascale, challenges related to data and energy are becoming dominating concerns. For example, coupled simulation workflows are increasingly adopting in-situ data processing and analysis techniques to address costs and overheads due to data movement and I/O. However it is also critical to understand these overheads and associated trade-offs from an energy perspective. The goal of this paper is exploring data-related energy/performance trade-offs for end-to-end simulation workflows running at scale on current high-end computing systems. Specifically, this paper presents: (1) an analysis of the data-related behaviors of a combustion simulation workflow with an insitu data analytics pipeline, running on the Titan system at ORNL; (2) a power model based on system power and data exchange patterns, which is empirically validated; and (3) the use of the model to characterize the energy behavior of the workflow and to explore energy/performance tradeoffs on current as well as emerging systems.
{"title":"Exploring power behaviors and trade-offs of in-situ data analytics","authors":"Marc Gamell, I. Rodero, M. Parashar, Janine Bennett, H. Kolla, Jacqueline H. Chen, P. Bremer, Aaditya G. Landge, A. Gyulassy, P. McCormick, S. Pakin, Valerio Pascucci, S. Klasky","doi":"10.1145/2503210.2503303","DOIUrl":"https://doi.org/10.1145/2503210.2503303","url":null,"abstract":"As scientific applications target exascale, challenges related to data and energy are becoming dominating concerns. For example, coupled simulation workflows are increasingly adopting in-situ data processing and analysis techniques to address costs and overheads due to data movement and I/O. However it is also critical to understand these overheads and associated trade-offs from an energy perspective. The goal of this paper is exploring data-related energy/performance trade-offs for end-to-end simulation workflows running at scale on current high-end computing systems. Specifically, this paper presents: (1) an analysis of the data-related behaviors of a combustion simulation workflow with an insitu data analytics pipeline, running on the Titan system at ORNL; (2) a power model based on system power and data exchange patterns, which is empirically validated; and (3) the use of the model to characterize the energy behavior of the workflow and to explore energy/performance tradeoffs on current as well as emerging systems.","PeriodicalId":371074,"journal":{"name":"2013 SC - International Conference for High Performance Computing, Networking, Storage and Analysis (SC)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129485720","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","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：