检索结果-内蒙古大学图书馆

ieee International symposium on parallel and Distributed Processing Workshops and Phd Forum (IPDPSW)

作者： Eugenio Angriman Fabian Brandt-Tumescheit Leon Franke Alexander van der Grinten Henning Meyerhenke Department of Computer Science Humboldt-Universitat zu Berlin Germany Department of Chemistry Konstanz Research School Chemical Biology University of Konstanz Germany

ISBN: (数字)9781665497473

ISBN: (纸本)9781665497480

Network analysis has been applied in diverse application domains. We consider an application from protein dynamics, specifically residue interaction networks (RINs). While numerous RIN visualization tools exist, there are no solutions that are both easily programmable and as fast as optimized network analysis toolkits. In this work, we use NetworKit - an established package for network analysis - to build a cloud-based environment that enables domain scientists to run their visualization and analysis workflows on large compute servers, without requiring extensive programming and/or system administration knowledge. To demonstrate the versatility of this approach, we use it to build a custom Jupyter-based widget for RIN visualization. In contrast to existing RIN visualization approaches, our widget can easily be customized through simple modifications of Python code, while both supporting a comprehensive feature set and providing near real-time speed. Due to its integration into Jupyter notebooks, our widget can easily interact with other popular packages of the Python ecosystem to build custom analysis pipelines (e.g., pipelines that feed RIN data into downstream machine learning tasks).

关键词： Proteins Pipelines data visualization Network analyzers Machine learning Programming Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Design and Comparative Evaluation of visualization Onboarding Methods 21

Design and Comparative Evaluation of Visualization Onboardin...

引用

14th International symposium on Visual Information Communication and Interaction, VINCI 2021

作者： Stoiber, Christina Walchshofer, Conny Grassinger, Florian Stitz, Holger Streit, Marc Aigner, Wolfgang St. Poelten Uas Austria Jku Linz Austria Datavisyn GmbH Austria

ISBN: (纸本)9781450386470

Comprehending and exploring large and complex data is becoming increasingly important for a diverse population of users in a wide range of application domains. visualization has proven to be well-suited in supporting this endeavor by tapping into the power of human visual perception. However, non-experts in the field of visual analysis often have difficulties in correctly reading and interpreting information from novel visualization idioms. visualization onboarding can support novices in learning how to use new digital technologies. Therefore, we developed an interactive step-by-step guide and applied the method to four visualization techniques-a bar chart, a horizon graph, a change matrix, and a parallel coordinates plot. Results using Amazon Mechanical Turk workers show that there is a need for onboarding, especially for more complex visualization techniques. We further investigated the perception and rating of a scrollytelling and a video tutorial for the most unfamiliar visualization-the parallel coordinates plot. A comparison between the three onboarding methods indicates that participants appreciated the easy-to-understand examples, the precise wording of the onboarding messages in a step-by-step manner, and the introduction of interaction concepts by highlighting the most relevant information over all onboarding methods. The video tutorial supported the introduction of unknown interaction techniques best. © 2021 Owner/Author.

关键词： visualization

来源：评论

学校读者我要写书评

暂无评论

Top-Down Performance Profiling on NVIDIA's GPUs

Top-Down Performance Profiling on NVIDIA's GPUs

引用

International symposium on parallel and Distributed Processing (IPDPS)

作者： Alvaro Saiz Pablo Prieto Pablo Abad Jose Angel Gregorio Valentin Puente Computer Engineering Group Universidad de Cantabria Santander Spain

The rise of data-intensive algorithms, such as Machine Learning ones, has meant a strong diversification of graphics Processing Units (GPU) in fields with intensive data-Level parallelism. This trend, known as general-purpose computing on GPU (GP-GPU), makes the execution process on a GPU (seemingly simple in its architecture) far from trivial when targeting performance for many dissimilar applications. A proof of this is the existence of many profiling tools that help programmers to understand how to maximize hardware utilization. In contrast, this paper proposes a profiling tool focused on microarchitecture analysis under large sets of dissimilar applications. Therefore, the tool has a double objective. On the one hand, to check the suitability of a GPU for diverse sets of application kernels. On the other hand, to identify possible bottlenecks in a given GPU microarchitecture, facilitating the improvement of subsequent designs. For this purpose, using Top-Down methodology proposed by Intel for their CPUs as inspiration, we have defined a hierarchical organization for the execution pipeline of the GPU. The proposal makes use of the available hardware performance counters to identify how each component contributes to performance losses. We demonstrate the feasibility of the proposed methodology, analyzing how different modern NVIDIA architectures behave running relevant benchmarks, assessing in which microarchitecture component performance losses are the most significant.

关键词： Microarchitecture Machine learning algorithms Pipelines graphics processing units Computer architecture Organizations parallel processing

来源：评论

学校读者我要写书评

暂无评论

A Visual Analytics Framework for Reviewing Streaming Performance data 13

A Visual Analytics Framework for Reviewing Streaming Perform...

引用

ieee Pacific visualization symposium (PacificVis)

作者： Kesavan, Suraj P. Fujiwara, Takanori Li, Jianping Kelvin Ross, Caitlin Mubarak, Misbah Carothers, Christopher D. Ross, Robert B. Ma, Kwan-Liu Univ Calif Davis Davis CA 95616 USA Rensselaer Polytech Inst Troy NY 12181 USA Argonne Natl Lab Argonne IL 60439 USA

ISBN: (数字)9781728156972

ISBN: (纸本)9781728156972

Understanding and tuning the performance of extreme-scale parallel computing systems demands a streaming approach due to the computational cost of applying offline algorithms to vast amounts of performance log data. Analyzing large streaming data is challenging because the rate of receiving data and limited time to comprehend data make it difficult for the analysts to sufficiently examine the data without missing important changes or patterns. To support streaming data analysis, we introduce a visual analytic framework comprising of three modules: data management, analysis, and interactive visualization. The data management module collects various computing and communication performance metrics from the monitored system using streaming data processing techniques and feeds the data to the other two modules. The analysis module automatically identifies important changes and patterns at the required latency. In particular, we introduce a set of online and progressive analysis methods for not only controlling the computational costs but also helping analysts better follow the critical aspects of the analysis results. Finally, the interactive visualization module provides the analysts with a coherent view of the changes and patterns in the continuously captured performance data. Through a multi-faceted case study on performance analysis of parallel discrete-event simulation, we demonstrate the effectiveness of our framework for identifying bottlenecks and locating outliers.

关键词： Human-centered computing visualization visualization application domains Visual analytics

来源：评论

学校读者我要写书评

暂无评论

Toward Localized Topological data Structures: Querying the Forest for the Tree

引用

ieee TRANSACTIONS ON visualization AND COMPUTER graphics 2020年第1期26卷 173-183页

作者： Klacansky, Pavol Gyulassy, Attila Bremer, Peer-Timo Pascucci, Valerio Univ Utah Sci Comp & Imaging Inst Salt Lake City UT 84112 USA Lawrence Livermore Natl Lab Livermore CA USA

Topological approaches to data analysis can answer complex questions about the number, connectivity, and scale of intrinsic features in scalar data. However, the global nature of many topological structures makes their computation challenging at scale, and thus often limits the size of data that can be processed. One key quality to achieving scalability and performance on modern architectures is data locality, i.e., a process operates on data that resides in a nearby memory system, avoiding frequent jumps in data access patterns. From this perspective, topological computations are particularly challenging because the implied data structures represent features that can span the entire data set, often requiring a global traversal phase that limits their scalability. Traditionally, expensive preprocessing is considered an acceptable trade-off as it accelerates all subsequent queries. Most published use cases, however, explore only a fraction of all possible queries, most often those returning small, local features. In these cases, much of the global information is not utilized, yet computing it dominates the overall response time. We address this challenge for merge trees, one of the most commonly used topological structures. In particular, we propose an alternative representation, the merge forest a collection of local trees corresponding to regions in a domain decomposition. Local trees are connected by a bridge set that allows us to recover any necessary global information at query time. The resulting system couples (i) a preprocessing that scales linearly in practice with (ii) fast runtime queries that provide the same functionality as traditional queries of a global merge tree. We test the scalability of our approach on a shared-memory parallel computer and demonstrate how data structure locality enables the analysis of large data with an order of magnitude performance improvement over the status quo. Furthermore, a merge forest reduces the memory overhead compare

关键词： Merge tree parallel computation topology

来源：评论

学校读者我要写书评

暂无评论

Efficient parallel PageRank Algorithm for Network Analysis

Efficient Parallel PageRank Algorithm for Network Analysis

引用

ieee International symposium on parallel and Distributed Processing Workshops and Phd Forum (IPDPSW)

作者： Maxence Vandromme Serge G. Petiton Univ. Lille UMR 9189 - CRIStAL CNRS Lille France

ISBN: (数字)9781665497473

ISBN: (纸本)9781665497480

We propose an efficient version of the PageRank algorithm for adjacency matrices, that reduces the complexity by a factor two. This method computes the $A^{T}x$ operation on the transpose matrix $A^{T}$ without having to explicitly normalize and transpose the matrix. We implement the method using standard row-major and column-major matrix storage formats. We perform experiments with parallel implementations in OpenMP, on synthetic data as well as on matrices extracted from large-scale graphs. The experiments are done on two different Intel processors from recent generations. The columnmajor storage format version of our method shows good scaling and outperforms the standard PageRank in a majority of cases, even when not considering the preprocessing burden in the latter.

关键词： Distributed processing Instruction sets Scalability Memory management graphics processing units Network analyzers Complexity theory

来源：评论

学校读者我要写书评

暂无评论

VITALflow: Visual Interactive Traffic Analysis with NetFlow

VITALflow: Visual Interactive Traffic Analysis with NetFlow

引用

ieee symposium on Network Operations and Management

作者： Tina Tremel Jochen Kö gel Florian Jauernig Sebastian Meier Dennis Thom Franziska Becker Christoph Mü ller Steffen Koch IsarNet Software Solutions GmbH University of Stuttgart

Traffic analysis in large enterprise networks has become a vital task for network experts, as understanding application and user traffic is the basis for proper network management with respect to planning, formulating intents, or analyzing causes for implausible behavior. In such networks, NetFlow provides input to network monitoring systems that typically show time series visualizations along different data dimensions. We studied tasks and requirements of network experts and derived a visual analytics approach that improves their analytic workflow as it enables for exploration of large time spans quickly in a multidimensional manner. Our approach guides users and improves the scalability of analyses through a novel combination of a clustered time series view and filtering in interactive parallel sets into a coherent visual analysis framework. Clustering reveals typical patterns and deviations from the daily norm and serves as entry point to exploring, filtering and comparing multiple dimensions in the parallel sets view. In addition, we briefly discuss the feedback we received on two case studies with network experts.

关键词： Filtering Visual analytics Scalability Time series analysis data visualization Planning Behavioral sciences

来源：评论

学校读者我要写书评

暂无评论

ETH: An Architecture for Exploring the Design Space of In-situ Scientific visualization 34

ETH: An Architecture for Exploring the Design Space of In-si...

引用

34th ieee International parallel and Distributed Processing symposium (IPDPS)

作者： Abram, Gregory Adhinarayanan, Vignesh Feng, Wu-chun Rogers, David Ahrens, James Univ Texas Austin Texas Adv Comp Ctr Austin TX 78758 USA Virginia Tech Dept Comp Sci Blacksburg VA 24061 USA Los Alamos Natl Lab Comp Computat & Stat Sci Div Los Alamos NM 87545 USA

ISBN: (纸本)9781728168760

As high-performance computing (HPC) moves towards the exascale era, large-scale scientific simulations are generating enormous datasets. Many techniques (e.g., in-situ methods, data sampling, and compression) have been proposed to help visualize these large datasets under various constraints such as storage, power, and energy. However, evaluating these techniques and understanding the trade-offs (e.g., performance, efficiency, and quality) remains a challenging task. To enable exploration of the design space across such tradeoffs, we propose the Exploration Test Harness (ETH), an architecture for the early-stage exploration of visualization and rendering approaches, job layout, and visualization pipelines. ETH covers a broader parameter space than current large-scale visualization applications such as ParaView and VisIt. It also promotes the study of simulation-visualization coupling strategies through a data-centric approach, rather than requiring coupling with a specific scientific simulation code. Furthermore, with experimentation on an extensively instrumented supercomputer, we study more metrics of interest than was previously possible. Importantly, ETH will help to answer important what-if scenarios and trade-off questions in the early stages of pipeline development, helping scientists to make informed choices about how to best couple a simulation code with visualization at extreme scale.

关键词： In-situ Techniques High-Performance Computing Design-space Exploration Raycasting Energy Efficiency

来源：评论

学校读者我要写书评

暂无评论

AxoNN: An asynchronous, message-driven parallel framework for extreme-scale deep learning

AxoNN: An asynchronous, message-driven parallel framework fo...

引用

International symposium on parallel and Distributed Processing (IPDPS)

作者： Siddharth Singh Abhinav Bhatele Department of Computer Science University of Maryland College Park Maryland USA

In the last few years, the memory requirements to train state-of-the-art neural networks have far exceeded the DRAM capacities of modern hardware accelerators. This has necessitated the development of efficient algorithms to train these neural networks in parallel on large-scale GPU-based clusters. Since computation is relatively inexpensive on modern GPUs, designing and implementing extremely efficient communication in these parallel training algorithms is critical for extracting the maximum performance. This paper presents AxoNN, a parallel deep learning framework that exploits asynchrony and message-driven execution to schedule neural network operations on each GPU, thereby reducing GPU idle time and maximizing hardware efficiency. By using the CPU memory as a scratch space for offloading data periodically during training, AxoNN is able to reduce GPU memory consumption by four times. This allows us to increase the number of parameters per GPU by four times, thus reducing the amount of communication and increasing performance by over 13%. When tested against large transformer models with 12–100 billion parameters on 48–384 NVIDIA Tesla V100 GPUs, AxoNN achieves a per-GPU throughput of 49.4–54.78% of theoretical peak and reduces the training time by 22-37 days (15–25% speedup) as compared to the state-of-the-art.

关键词： Training Deep learning Schedules Neural networks Memory management graphics processing units Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Proceedings - 2020 ieee 10th symposium on large data Analysis and visualization, LDAV 2020

Proceedings - 2020 IEEE 10th Symposium on Large Data Analysi...

引用

10th ieee symposium on large data Analysis and visualization, LDAV 2020

ISBN: (纸本)9781728184685

The proceedings contain 7 papers. The topics discussed include: DIVA: a declarative and reactive language for in situ visualization;data parallel hypersweeps for in situ topological analysis;comparing time-to-solution for in situ visualization paradigms at scale;interactive visualization of terascale data in the browser: fact or fiction?;cinema darkroom: a deferred rendering framework for large-scale datasets;spatial partitioning strategies for memory-efficient ray tracing of particles;and a distributed algorithm for force directed edge bundling.

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：