检索结果-内蒙古大学图书馆

Efficient distributed algorithms for holistic aggregation functions on random regular graphs

Science China(Information Sciences) 2022年第5期65卷 32-50页

作者： Lin JIA Qiang-Sheng HUA Haoqiang FAN Qiuping WANG Hai JIN National Engineering Research Center for Big Data Technology and System/Services Computing Technology and System Lab/Cluster and Grid Computing Lab School of Computer Science and TechnologyHuazhong University of Science and Technology Institute for Interdisciplinary Information Science Tsinghua University

In this paper, we propose efficient distributed algorithms for three holistic aggregation functions on random regular graphs that are good candidates for network topology in next-generation data *** three holistic aggregation functions include SELECTION（select the k-th largest or smallest element）,DISTINCT（query the count of distinct elements）, MODE（query the most frequent element）. We design three basic techniques — Pre-order Network Partition, Pairwise-independent Random Walk, and Random Permutation Delivery, and devise the algorithms based on the techniques. The round complexity of the distributed SELECTION is Θ（log N） which meets the lower bound where N is the number of nodes and each node holds a numeric element. The round complexity of the distributed DISTINCT and MODE algorithms are O（log3N/log log N） and O（log2N log log N） respectively. All of our results break the lower bounds obtained on general graphs and our distributed algorithms are all based on the CON GE S T model, which restricts each node to send only O（log N） bits on each edge in one round under synchronous communications.

关键词： distributed algorithms holistic aggregation function random regular graph ${\cal C}{\cal O}{\cal N}{\cal G}{\cal E}{\cal S}{\cal T}$ model communication complexity round complexity

来源：评论

学校读者我要写书评

暂无评论

GraphMetaP: Efficient MetaPath Generation for Dynamic Heterogeneous Graph Models

GraphMetaP: Efficient MetaPath Generation for Dynamic Hetero...

引用

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Haiheng He Dan Chen Long Zheng Yu Huang Haifeng Liu Chaoqiang Liu Xiaofei Liao Hai Jin National Engineering Research Center for Big Data Technology and System/Services Computing Technology and System Lab/Clusters and Grid Computing Lab Huazhong University of Science and Technology China Zhejiang Lab Hangzhou China

Metapath-based heterogeneous graph models (MHGM) show excellent performance in learning semantic and structural information in heterogeneous graphs. Metapath matching is an essential processing step in MHGM to find all metapath instances, bringing significant overhead compared to the total model execution time. Even worse, in dynamic heterogeneous graphs, metapath instances require to be rematched while graph updated. In this paper, we observe that only a small fraction of metapath instances change and propose GraphMetaP, an efficient incremental metapath maintenance method in order to eliminate the matching overhead in dynamic heterogeneous graphs. GraphMetaP introduces a novel format for metapath instances to capture the dependencies among the metapath instances. The format incrementally maintains metapath instances based on the graph updates to avoide the rematching metapath overhead for the updated graph. Furthermore, GraphMetaP uses the fold way to simplify the format in order to recover all metapath instances faster. Experiments show that GraphMetaP enables efficient maintenance of metapath instances on dynamic heterogeneous graphs and outperforms 172.4X on average compared to the matching metapath method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Duo: Improving data Sharing of Stateful Serverless Applications by Efficiently Caching Multi-Read data

Duo: Improving Data Sharing of Stateful Serverless Applicati...

引用

International Symposium on Parallel and Distributed Processing (IPDPS)

作者： Zhuo Huang Hao Fan Chaoyi Cheng Song Wu Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China

A growing number of applications are moving to serverless architectures for high elasticity and fine-grained billing. For stateful applications, however, the use of serverless architectures is likely to lead to significant performance degradation, as frequent data sharing between different execution stages involves time-consuming remote storage access. Current platforms leverage memory cache to speed up remote access. However, conventional caching strategies show limited performance improvement. We experimentally find that the reason is that current strategies overlook the stage-dependent access patterns of stateful serverless applications, i.e., data that are read multiple times across stages (denoted as multi-read data) are wrongly evicted by data that are read only once (denoted as read-once data), causing a high cache miss ***, we propose a new caching strategy, Duo, whose design principle is to cache multi-read data as long as possible. Specifically, Duo contains a large cache list and a small cache list, which act as Leader list and Wingman list, respectively. Leader list ignores the data that is read for the first time to prevent itself from being polluted by massive read-once data at each stage. Wingman list inspects the data that are ignored or evicted by Leader list, and pre-fetches the data that will probably be read again based on the observation that multi-read data usually appear periodically in groups. Compared to the state-of-the-art works, Duo improves hit ratio by 1.1×-2.1× and reduces the data sharing overhead by 25%-62%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enabling Efficient Large Recommendation Model Training with Near CXL Memory Processing

Enabling Efficient Large Recommendation Model Training with ...

引用

Annual International Symposium on Computer Architecture, ISCA

作者： Haifeng Liu Long Zheng Yu Huang Jingyi Zhou Chaoqiang Liu Runze Wang Xiaofei Liaot Hai Jinf Jingling Xue National Engineering Research Center for Big Data Technology and System/Services Computing Technology and System Lab/Cluster and Grid Computing Lab Huazhong University of Science and Technology China School of Computer Science and Engineering University of New South Wales Australia Zhejiang Lab Hangzhou China

ISBN: (数字)9798350326581

ISBN: (纸本)9798350326598

Personalized recommendation systems have become one of the most important Internet services nowadays. A critical challenge of training and deploying the recommendation models is their high memory capacity and bandwidth demands, with the embedding layers occupying hundreds of GBs to TBs of storage. The advent of memory disaggregation technology and Compute Express Link (CXL) provides a promising solution for memory capacity scaling. However, relocating memory-intensive embedding layers to CXL memory incurs noticeable performance degradation due to its limited transmission bandwidth, which is significantly lower than the host memory bandwidth. To address this, we introduce ReCXL, a CXL memory disaggregation system that utilizes near-memory processing for scalable, efficient recommendation model training. ReCXL features a unified, hardwareefficient NMP architecture that processes the entire embedding training within CXL memory, minimizing data transfers over the bandwidth-limited CXL and enhancing internal bandwidth. To further improve the performance, ReCXL incorporates softwarehardware co-optimizations, including sophisticated dependencyfree prefetching and fine-grained update scheduling, to maximize hardware utilization. Evaluation results show that ReCXL outperforms the CPU-GPU baseline and the naïve CXL memory by $7.1 \times \sim 10.6 \times(9.4 \times$ on average) and $12.7 \times \sim 31.3 \times(22.6 \times$ on average), respectively.

关键词： Training Degradation Scalability Prefetching Memory management Web and internet services Bandwidth

来源：评论

学校读者我要写书评

暂无评论

Scalable Optimal Margin Distribution Machine

arXiv

引用

arXiv 2023年

作者： Wang, Yilin Cao, Nan Zhang, Teng Shi, Xuanhua Jin, Hai National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China

Optimal margin Distribution Machine (ODM) is a newly proposed statistical learning framework rooting in the latest margin theory, which demonstrates better generalization performance than the traditional large margin based counterparts. However, it suffers from the ubiquitous scalability problem regarding both computation time and memory storage as other kernel methods. This paper proposes a scalable ODM, which can achieve nearly ten times speedup compared to the original ODM training method. For nonlinear kernels, we put forward a novel distribution-aware partition method to make the local ODM trained on each partition be close and converge fast to the global one. When linear kernel is applied, we extend a communication efficient SVRG method to accelerate the training further. Extensive empirical studies validate that our proposed method is highly computational efficient and almost never worsen the generalization. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Fast Parallel Recovery for Transactional Stream Processing on Multicores

Fast Parallel Recovery for Transactional Stream Processing o...

引用

International Conference on data engineering

作者： Jianjun Zhao Haikun Liu Shuhao Zhang Zhuohui Duan Xiaofei Liao Hai Jin Yu Zhang National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China School of Computer Science and Engineering Nanyang Technological University Singapore

ISBN: (数字)9798350317152

ISBN: (纸本)9798350317169

Transactional stream processing engines (TSPEs) have gained increasing attention due to their capability of processing real-time stream applications with transactional semantics. However, TSPEs remain susceptible to system failures and power outages. Existing TSPEs mainly focus on performance improvement, but still face a significant challenge to guarantee fault tolerance while offering high-performance services. We revisit commonly-used fault tolerance approaches in stream processing and database systems, and find that these approaches do not work well on TSPEs due to complex data dependencies. In this paper, we propose a novel TSPE called MorphStreamR to achieve fast failure recovery while guaranteeing low performance overhead at runtime. The key idea of MorphStreamR is to record intermediate results of resolved dependencies at runtime, and thus eliminate data dependencies to improve task parallelism during failure recovery. MorphStreamR further mitigates the runtime overhead by selectively tracking data dependencies and incorporating workload-aware log commitment. Experimental results show that MorphStreamR can significantly reduce the recovery time by up to 3.1 x while experiencing much less performance slowdown at runtime, compared with other applicable fault tolerance approaches.

关键词： Fault tolerance Runtime Fault tolerant systems Semantics Parallel processing Real-time systems Power system reliability

来源：评论

学校读者我要写书评

暂无评论

On Pipelined GCN with Communication-Efficient Sampling and Inclusion-Aware Caching

On Pipelined GCN with Communication-Efficient Sampling and I...

引用

IEEE Annual Joint Conference: INFOCOM, IEEE Computer and Communications Societies

作者： Shulin Wang Qiang Yu Xiong Wang Yuqing Li Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China School of Cyber Science and Engineering Wuhan University Wuhan China

ISBN: (数字)9798350383508

ISBN: (纸本)9798350383515

Graph convolutional network (GCN) has achieved enormous success in learning structural information from unstructured data. As graphs become increasingly large, distributed training for GCNs is severely prolonged by frequent cross-worker communications. Existing efforts to improve the training efficiency often come at the expense of GCN performance, while the communication overhead persists. In this paper, we propose PSC-GCN, a holistic pipelined framework for distributed GCN training with communication-efficient sampling and inclusion-aware caching, to address the communication bottleneck while ensuring satisfactory model performance. Specifically, we devise an asynchronous pre-fetching scheme to retrieve stale statistics (features, embedding, gradient) of boundary nodes in advance, such that the embedding aggregation and model update are pipelined with statistics transmission. To alleviate communication volume and staleness effect, we introduce a variance-reduction based sampling policy, which prioritizes inner nodes over boundary ones for reducing the access frequency to remote neighbors, thus mitigating cross-worker statistics exchange. Complementing graph sampling, a feature caching module is co-designed to buffer hot nodes with high inclusion probability, ensuring that frequently sampled nodes will be available in local memory. Extensive evaluations on real-world datasets show the superiority of PSC-GCN over state-of-the-art methods, where we can reduce training time by 72%-80% without sacrificing model accuracy.

关键词： Training Estimation error Accuracy Graph convolutional networks Computational modeling Probability Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

Reveal training performance mystery between Tensor Flow and PyTorch in the single GPU environment

引用

Science China(Information Sciences) 2022年第1期65卷 147-163页

作者： Hulin DAI Xuan PENG Xuanhua SHI Ligang HE Qian XIONG Hai JIN National Engineering Research Center for Big Data Technology and System Service Computing Technology and System LabSchool of Computer Science and Technology Huazhong University of Science and Technology Department of Computer Science University of Warwick

Deep learning has gained tremendous success in various fields while training deep neural networks(DNNs) is very compute-intensive, which results in numerous deep learning frameworks that aim to offer better usability and higher performance to deep learning practitioners. Tensor Flow and Py Torch are the two most popular frameworks. Tensor Flow is more promising within the industry context, while Py Torch is more appealing in academia. However, these two frameworks differ much owing to the opposite design philosophy:static vs dynamic computation graph. Tensor Flow is regarded as being more performance-friendly as it has more opportunities to perform optimizations with the full view of the computation graph. However, there are also claims that Py Torch is faster than Tensor Flow sometimes, which confuses the end-users on the choice between them. In this paper, we carry out the analytical and experimental analysis to unravel the mystery of comparison in training speed on single-GPU between Tensor Flow and Py Torch. To ensure that our investigation is as comprehensive as possible, we carefully select seven popular neural networks, which cover computer vision, speech recognition, and natural language processing(NLP). The contributions of this work are two-fold. First, we conduct the detailed benchmarking experiments on Tensor Flow and Py Torch and analyze the reasons for their performance difference. This work provides the guidance for the end-users to choose between these two frameworks. Second, we identify some key factors that affect the performance,which can direct the end-users to write their models more efficiently.

关键词： deep learning performance comparison TensorFlow PyTorch

来源：评论

学校读者我要写书评

暂无评论

Securely Fine-tuning Pre-trained Encoders Against Adversarial Examples

Securely Fine-tuning Pre-trained Encoders Against Adversaria...

引用

IEEE Symposium on Security and Privacy

作者： Ziqi Zhou Minghui Li Wei Liu Shengshan Hu Yechao Zhang Wei Wan Lulu Xue Leo Yu Zhang Dezhong Yao Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology School of Software Engineering Huazhong University of Science and Technology Hubei Engineering Research Center on Big Data Security Hubei Key Laboratory of Distributed System Security School of Cyber Science and Engineering Huazhong University of Science and Technology School of Information and Communication Technology Griffith University

ISBN: (数字)9798350331301

ISBN: (纸本)9798350331318

With the evolution of self-supervised learning, the pre-training paradigm has emerged as a predominant solution within the deep learning landscape. Model providers furnish pre-trained encoders designed to function as versatile feature extractors, enabling downstream users to harness the benefits of expansive models with minimal effort through fine-tuning. Nevertheless, recent works have exposed a vulnerability in pre-trained encoders, highlighting their susceptibility to downstream-agnostic adversarial examples (DAEs) meticulously crafted by attackers. The lingering question pertains to the feasibility of fortifying the robustness of downstream models against DAEs, particularly in scenarios where the pre-trained encoders are publicly accessible to the *** this paper, we initially delve into existing defensive mechanisms against adversarial examples within the pre-training paradigm. Our findings reveal that the failure of current defenses stems from the domain shift between pre-training data and downstream tasks, as well as the sensitivity of encoder parameters. In response to these challenges, we propose Genetic Evolution-Nurtured Adversarial Fine-tuning (Gen-AF), a two-stage adversarial fine-tuning approach aimed at enhancing the robustness of downstream models. Gen-AF employs a genetic-directed dual-track adversarial fine-tuning strategy in its first stage to effectively inherit the pre-trained encoder. This involves optimizing the pre-trained encoder and classifier separately while incorporating genetic regularization to preserve the model’s topology. In the second stage, Gen-AF assesses the robust sensitivity of each layer and creates a dictionary, based on which the top-k robust redundant layers are selected with the remaining layers held fixed. Upon this foundation, we conduct evolutionary adaptability fine-tuning to further enhance the model’s generalizability. Our extensive experiments, conducted across ten self-supervised training methods and six d

关键词： Training Adaptation models Sensitivity Accuracy Self-supervised learning Genetics Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A Survey of Non-Volatile Main Memory Technologies:State-of-the-Arts,Practices,and Future Directions

引用

Journal of Computer Science & technology 2021年第1期36卷 4-32页

作者： Hai-Kun Liu Di Chen Hai Jin Xiao-Fei Liao Binsheng He Kan Hu Yu Zhang National Engineering Research Center for Big Data Technology and System Huazhong University of Science and TechnologyWuhan 430074China Services Computing Technology and System Laboratory Huazhong University of Science and Technology Wuhan 430074China Cluster and Grid Computing Laboratory Huazhong University of Science and TechnologyWuhan 430074China School of Computing Science and Technology Huazhong University of Science and TechnologyWuhan 430074China School of Computing National University of SingaporeSingapore 117418Singapore

Non-Volatile Main Memories (NVMMs) have recently emerged as a promising technology for future memory systems. Generally, NVMMs have many desirable properties such as high density, byte-addressability, non-volatility, low cost, and energy efficiency, at the expense of high write latency, high write power consumption, and limited write endurance. NVMMs have become a competitive alternative of Dynamic Random Access Memory (DRAM), and will fundamentally change the landscape of memory systems. They bring many research opportunities as well as challenges on system architectural designs, memory management in operating systems (OSes), and programming models for hybrid memory systems. In this article, we revisit the landscape of emerging NVMM technologies, and then survey the state-of-the-art studies of NVMM technologies. We classify those studies with a taxonomy according to different dimensions such as memory architectures, data persistence, performance improvement, energy saving, and wear leveling. Second, to demonstrate the best practices in building NVMM systems, we introduce our recent work of hybrid memory system designs from the dimensions of architectures, systems, and applications. At last, we present our vision of future research directions of NVMMs and shed some light on design challenges and opportunities.

关键词： non-volatile memory persistent memory hybrid memory systems memory hierarchy

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：