检索结果-内蒙古大学图书馆

45th ACM/IEEE Annual International Symposium on Computer Architecture (ISCA)

作者： Riera, Marc Arnau, Jose-Maria Gonzalez, Antonio UPC Dept Comp Architecture Barcelona Spain

ISBN: (纸本)9781538659847

In recent years, Deep Neural Networks (DNNs) have achieved tremendous success for diverse problems such as classification and decision making. Efficient support for DNNs on CPUs, GPUs and accelerators has become a prolific area of research, resulting in a plethora of techniques for energy-efficient DNN inference. However, previous proposals focus on a single execution of a DNN. Popular applications, such as speech recognition or video classification, require multiple back-to-back executions of a DNN to process a sequence of inputs (e.g., audio frames, images). In this paper, we show that consecutive inputs exhibit a high degree of similarity, causing the inputs/outputs of the different layers to be extremely similar for successive frames of speech or images of a video. Based on this observation, we propose a technique to reuse some results of the previous execution, instead of computing the entire DNN. computations related to inputs with negligible changes can be avoided with minor impact on accuracy, saving a large percentage of computations and memory accesses. We propose an implementation of our reuse-based inference scheme on top of a state-of-the-art DNN accelerator. Results show that, on average, more than 60% of the inputs of any neural network layer tested exhibit negligible changes with respect to the previous execution. Avoiding the memory accesses and computations for these inputs results in 63% energy savings on average.

关键词： DNN computation reuse Input Similarity Hardware Accelerator

来源：评论

学校读者我要写书评

暂无评论

computation reuse in Analytics Job Service at Microsoft 18

Computation Reuse in Analytics Job Service at Microsoft

引用

44th ACM SIGMOD International Conference on Management of Data

作者： Jindal, Alekh Qiao, Shi Patel, Hiren Yin, Zhicheng Di, Jieming Bag, Malay Friedman, Marc Lin, Yifung Karanasos, Konstantinos Rao, Sriram Microsoft Redmond WA 98052 USA

ISBN: (纸本)9781450317436

Analytics-as-a-service, or analytics job service, is emerging as a new paradigm for data analytics, be it in a cloud environment or within enterprises. In this setting, users are not required to manage or tune their hardware and software infrastructure, and they pay only for the processing resources consumed per job. However, the shared nature of these job services across several users and teams leads to significant overlaps in partial computations, i.e., parts of the processing are duplicated across multiple jobs, thus generating redundant costs. In this paper, we describe a computation reuse framework, coined CLOUDVIEWS, which we built to address the computation overlap problem in Microsoft's SCOPE job service. We present a detailed analysis from our production workloads to motivate the computation overlap problem and the possible gains from computation reuse. The key aspects of our system are the following: (i) we reuse computations by creating materialized views over recurring workloads, i.e., periodically executing jobs that have the same script templates but process new data each time, (ii) we select the views to materialize using a feedback loop that reconciles the compile-time and run-time statistics and gathers precise measures of the utility and cost of each overlapping computation, and (iii) we create materialized views in an online fashion, without requiring an offline phase to materialize the overlapping computations.

关键词： Materialized Views computation reuse Shared Clouds

来源：评论

学校读者我要写书评

暂无评论

Highly Optimized Code Generation for Stencil Codes with computation reuse for GPUs

引用

Journal of Computer Science & Technology 2016年第6期31卷 1262-1274页

作者： Wen-Jing Ma Kan Gao Guo-Ping Long Laboratory of Parallel Software and Computing Science Institute of Software Chinese Academy of Sciences Beijing 100190 China State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing 100190 China Information Center China Association for Science and Technology Beijing 100863 China

computation reuse is known as an effective optimization technique. However, due to the complexity of modern GPU architectures, there is yet not enough understanding regarding the intriguing implications of the interplay of compu- ration reuse and hardware specifics on application performance. In this paper, we propose an automatic code generator for a class of stencil codes with inherent computation reuse on CPUs. For such applications, the proper reuse of intermediate results, combined with careful register and on-chip local memory usage, has profound implications on performance. Current state of the art does not address this problem in depth, partially due to the lack of a good program representation that can expose all potential computation reuse. In this paper, we leverage the computation overlap graph （COG）, a simple representation of data dependence and data reuse with ＂element view＂, to expose potential reuse opportunities. Using COG, we propose a portable code generation and tuning framework for GPUs. Compared with current state-of-the-art code generators, our experimental results show up to 56.7% performance improvement on modern GPUs such as NVIDIA C2050.

关键词： GPGPU OpenCL stencil code generation computation reuse

来源：评论

学校读者我要写书评

暂无评论

Isogeometric computation reuse method for complex objects with topology-consistent volumetric parameterization

引用

COMPUTER-AIDED DESIGN 2017年 91卷 1-13页

作者： Xu, Gang Kwok, Tsz-Ho Wang, Charlie C. L. Hangzhou Dianzi Univ Sch Comp Sci & Technol Hangzhou 310018 Zhejiang Peoples R China Minist Educ Key Lab Complex Syst Modeling & Simulat Hangzhou Zhejiang Peoples R China Chinese Univ Hong Kong Dept Mech & Automat Engn Hong Kong Hong Kong Peoples R China Concordia Univ Dept Mech & Ind Engn Montreal PQ Canada Delft Univ Technol Dept Design Engn Delft Netherlands

Volumetric spline parameterization and computational efficiency are two main challenges in isogeometric analysis (IGA). To tackle this problem, we propose a framework of computation reuse in IGA on a set of three-dimensional models with similar semantic features. Given a template domain, B-spline based consistent volumetric parameterization is first constructed for a set of models with similar semantic features. An efficient quadrature-free method is investigated in our framework to compute the entries of stiffness matrix by Bezier extraction and polynomial approximation. In our approach, evaluation on the stiffness matrix and imposition of the boundary conditions can be pre-computed and reused during IGA on a set of CAD models. Examples with complex geometry are presented to show the effectiveness of our methods, and efficiency similar to the computation in linear finite element analysis can be achieved for IGA taken on a set of models. (C) 2017 Elsevier Ltd. All rights reserved.

关键词： computation reuse Isogeometric analysis Consistent volume parameterization

来源：评论

学校读者我要写书评

暂无评论

Optimizing parameter sensitivity analysis of large-scale microscopy image analysis workflows with multilevel computation reuse

引用

CONCURRENCY AND computation-PRACTICE & EXPERIENCE 2020年第2期32卷 e5403-e5403页

作者： Barreiros, Willian, Jr. Moreira, Jeremias Kurc, Tahsin Kong, Jun Melo, Alba C. M. A. Saltz, Joel H. Teodoro, George Univ Brasilia Dept Comp Sci Brasilia DF Brazil SUNY Stony Brook Dept Biomed Informat Stony Brook NY 11794 USA Oak Ridge Natl Lab Sci Data Grp Oak Ridge TN USA Emory Univ Dept Biomed Informat Atlanta GA 30322 USA Emory Univ Dept Comp Sci Atlanta GA 30322 USA Georgia State Univ Dept Math & Stat Atlanta GA 30303 USA Univ Fed Minas Gerais Dept Comp Sci BR-31270901 Belo Horizonte MG Brazil

Parameter sensitivity analysis (SA) is an effective tool to gain knowledge about complex analysis applications and assess the variability in their analysis results. However, it is an expensive process as it requires the execution of the target application multiple times with a large number of different input parameter values. In this work, we propose optimizations to reduce the overall computation cost of SA in the context of analysis applications that segment high-resolution slide tissue images, ie, images with resolutions of 100k x 100k pixels. Two cost-cutting techniques are combined to efficiently execute SA: use of distributed hybrid systems for parallel execution and computation reuse at multiple levels of an analysis pipeline to reduce the amount of computation. These techniques were evaluated using a cancer image analysis workflow on a hybrid cluster with 256 nodes, each with an Intel Phi and a dual socket CPU. Our parallel execution method attained an efficiency of over 90% on 256 nodes. The hybrid execution on the CPU and Intel Phi improved the performance by 2x. Multilevel computation reuse led to performance gains of over 2.9x.

关键词： computation reuse microscopy imaging parameter sensitivity analysis

来源：评论

学校读者我要写书评

暂无评论

CoxNet: A computation reuse Architecture at the Edge

IEEE TRANSACTIONS ON GREEN COMMUNICATIONS AND NETWORKING

引用

IEEE TRANSACTIONS ON GREEN COMMUNICATIONS AND NETWORKING 2021年第2期5卷 765-777页

作者： Bellal, Zouhir Nour, Boubakr Mastorakis, Spyridon Ecole Super Informat LabRI SBA Lab Sidi Bel Abbes 2045 Algeria Beijing Inst Technol Sch Comp Sci Beijing 100081 Peoples R China Univ Nebraska Dept Comp Sci Omaha NE 68182 USA

In recent years, edge computing has emerged as an effective solution to extend cloud computing and satisfy the demand of applications for low latency. However, with today's explosion of innovative applications (e.g., augmented reality, natural language processing, virtual reality), processing services for mobile and smart devices have become computation-intensive, consisting of multiple interconnected computations. This coupled with the need for delay-sensitivity and high quality of service put massive pressure on edge servers. Meanwhile, tasks invoking these services may involve similar inputs that could lead to the same output. In this paper, we present CoxNet, an efficient computation reuse architecture for edge computing. CoxNet enables edge servers to reuse previous computations while scheduling dependent incoming computations. We provide an analytical model for computation reuse joined with dependent task offloading and design a novel computing offloading scheduling scheme. We also evaluate the efficiency and effectiveness of CoxNet via synthetic and real-world datasets. Our results show that CoxNet is able to reduce the task execution time up to 66% based on a synthetic dataset and up to 50% based on a real-world dataset.

关键词： Task analysis Cloud computing Servers Streaming media Edge computing Automobiles Image segmentation Edge computing service offloading computation reuse serverless computing

来源：评论

学校读者我要写书评

暂无评论

Enhancing computation reuse efficiency in ICN-based edge computing by modifying content store table structure

引用

COMPUTING 2024年第9期106卷 2949-2969页

作者： Javaheri, Atiyeh Bohlooli, Ali Jamshidi, Kamal Univ Isfahan Fac Comp Engn Esfahan Iran

In edge computing, repetitive computations are a common occurrence. However, the traditional TCP/IP architecture used in edge computing fails to identify these repetitions, resulting in redundant computations being recomputed by edge resources. To address this issue and enhance the efficiency of edge computing, Information-Centric Networking (ICN)-based edge computing is employed. The ICN architecture leverages its forwarding and naming convention features to recognize repetitive computations and direct them to the appropriate edge resources, thereby promoting "computation reuse". This approach significantly improves the overall effectiveness of edge computing. In the realm of edge computing, dynamically generated computations often experience prolonged response times. To establish and track connections between input requests and the edge, naming conventions become crucial. By incorporating unique IDs within these naming conventions, each computing request with identical input data is treated as distinct, rendering ICN's aggregation feature unusable. In this study, we propose a novel approach that modifies the Content Store (CS) table, treating computing requests with the same input data and unique IDs, resulting in identical outcomes, as equivalent. The benefits of this approach include reducing distance and completion time, and increasing hit ratio, as duplicate computations are no longer routed to edge resources or utilized cache. Through simulations, we demonstrate that our method significantly enhances cache reuse compared to the default method with no reuse, achieving an average improvement of over 57%. Furthermore, the speed up ratio of enhancement amounts to 15%. Notably, our method surpasses previous approaches by exhibiting the lowest average completion time, particularly when dealing with lower request frequencies. These findings highlight the efficacy and potential of our proposed method in optimizing edge computing performance.

关键词： computation reuse Edge computing ICN ICN-based edge computing

来源：评论

学校读者我要写书评

暂无评论

Power-Efficient Accelerator Design for Neural Networks Using computation reuse

IEEE COMPUTER ARCHITECTURE LETTERS

引用

IEEE COMPUTER ARCHITECTURE LETTERS 2017年第1期16卷 72-75页

作者： Yasoubi, Ali Hojabr, Reza Modarressi, Mehdi Univ Tehran Dept Elect & Comp Engn Coll Engn Tehran Iran

Applications of neural networks in various fields of research and technology have expanded widely in recent years. In particular, applications with inherent tolerance to accuracy loss, such as signal processing and multimedia applications, are highly suited to the approximation property of neural networks. This approximation property has been exploited in many existing neural network accelerators to trade-off accuracy for power-efficiency and speed. In addition to the power saving obtained by approximation, we observed that a considerable amount of arithmetic operations in neural networks are repetitive and can be eliminated to further decrease power consumption. Given this observation, we propose CORN, computation reuse-aware Neural network accelerator that allows neurons to share their computation results, effectively eliminating the power usage of redundant computations. We will show that CORN lowers power consumption by 26 percent on average over low-power neural network accelerators.

关键词： Neural network computation reuse hardware accelerator power-efficiency approximation

来源：评论

学校读者我要写书评

暂无评论

A 43.1TOPS/W Energy-Efficient Absolute-Difference-Accumulation Operation Computing-In-Memory With computation reuse

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS 2021年第5期68卷 1605-1609页

作者： Um, Soyeon Kim, Sangyeob Kim, Sangjin Yoo, Hoi-Jun Korea Adv Inst Sci & Technol Sch Elect Engn Daejeon 34141 South Korea

Recently, Computing-In-Memory (CIM) processors have been proposed to achieve high energy-efficiency by reducing data movement and solving memory bottlenecks. Furthermore, a network with high accurate image classification has been introduced through the Absolute-Difference-Accumulation (ADA) operation instead of the multiplication-and-accumulation operation, which is widely used in DNN. ADA operation provides not only opportunities for high energy-efficient DNN accelerating by reducing multiplication but also a chance to reuse computation results. However, the previous CIM processor cannot reuse previous computation results for other computations. In this brief, we propose a high accurate and high energy-efficient ADA-CIM processor that with two key features: 1) computation reuse for low-power, resulting in a 49.5% CIM operation power reduction, and 2) low-cost sign prediction core with 3-bit activation and weight quantization for high utilization. From the two key features, the proposed ADA-CIM processor is simulated in 28 nm CMOS technology and occupies 3.78 mm(2). It consumes 2.77mW and achieves 43.1 TOPS/W energy-efficiency with a high-accuracy of 91.62% at CIFAR-10 (ResNet-20).

关键词： Random access memory Program processors Energy efficiency Common Information Model (computing) Quantization (signal) Computer architecture Circuits and systems Absolute-difference-accumulation (ADA) operation computation reuse computing-in-memory (CIM) energy-efficient low-cost absolute difference sign prediction SRAM

来源：评论

学校读者我要写书评

暂无评论

Service Caching and computation reuse Strategies at the Edge: A Survey

引用

ACM COMPUTING SURVEYS 2024年第2期56卷 1-38页

作者： Barrios, Carlos Kumar, Mohan Rochester Inst Technol GCCIS 20 Lomb Mem Dr New York NY 14623 USA

With the proliferation of connected devices including smartphones, novel network connectivity and management methods are needed to meet user Quality of Experience (QoE) and computational demands of contemporary applications. Ser vice caching and computation reuse techniques are being employed to alleviate challenges due to scalability, interoperability, and mobility, as well as to reduce application latency by enabling caching at the edge. This survey provides a taxonomy for service caching and computation reuse and describes the current state of the research and its challenges. This is the first survey that provides a comprehensive analysis and suggests future research directions on this topic.

关键词： Service caching computation reuse task caching computation deduplication

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：