检索结果-内蒙古大学图书馆

Logic Design of an 8-bit RSFQ Microprocessor

学校读者我要写书评

暂无评论

Logic Design of an 8-bit RSFQ Microprocessor

IEEE International Superconductive Electronics Conference (ISEC)

作者： Jia-Hong Yang Guang-Ming Tang Pei-Yao Qu Xiao-Chun Ye Dong-Rui Fan Zhi-Min Zhang Ning-Hui Sun University of Chinese Academy of Sciences Beijing China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China

An 8-bit bit-parallel RSFQ microprocessor, named HUTU, is proposed. It can execute 28 different instructions. Each instruction consists of eight bits. Harvard-type architecture is adopted for parallel processing between the control unit and the datapath. The control unit uses an asynchronous timing method to avoid pipeline flushing and to reduce the area. Concurrent-flow clocking is adopted in the datapath for high performance. The simulation result shows that the elements of HUTU run correctly.

关键词：

Efficient dynamic evolution of service composition

学校读者我要写书评

暂无评论

Efficient dynamic evolution of service composition

作者： Lv, Chen Jiang, Wei Hu, Songlin Wang, Jiye Lu, Guoliang Liu, Zhiyong School of Information Science and Engineering Shandong Normal University China Greatwall Drilling Company R&D Academy of Well Logging CNPC Beijing China Institute of Information Engineering Chinese Academy of Sciences Beijing China Department of Information Technology State Grid Corporation of China Beijing China School of Mechanical Engineering Shandong University China Institute of Computing Technology Chinese Academy of Sciences Beijing China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China

Current QoS-aware automatic service composition queries over a network of web services are often one-time in nature. After a network of web services is built, such queries are issued once, and answers are found from the scratch. The underlying assumption is that the participating web services are rather static, which means that their functional and non-functional parameters seldom change. However, such an assumption is often baseless. New services come and go, service APIs change gradually, and QoS values fluctuate in practice. Therefore, a support for efficiently adapting service composition is desired. In this paper, we propose an event driven continuous query algorithm to intelligently cope with different types of dynamic services, and thus enable evolution of service composition. Evaluation using both real QoS data and synthetic web service data shows its superior performance, compared with the state-of-the-art solution which won the performance championship of Web Service Challenge in 2009 and 2010. Moreover, we generalize a new graph problem: dynamic single-source optimal-Directed Acyclic Graphs (DAG) problem based on the above work. It is argued that API recommendation for framework evolution, as another typical application, could also be modeled and addressed efficiently using the proposed new graph approach. © 2017 IEEE.

关键词： Web services

Automatic Implementation of Multi-partitioning Using Global Tiling

学校读者我要写书评

暂无评论

Automatic Implementation of Multi-partitioning Using Global ...

International Conference on Parallel and Distributed systems (ICPADS)

作者： Lei Liu Dingfei Zhang Hengjie Li Li Chen Graduate School of Chinese Academy of Sciences Beijing China Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy and Sciences Beijing China

Strategies for partitioning an application¿s data and computation play fundamental role in determining the efficiency of parallelization. This paper describes a sophisticated strategy for partitioning data and computation known as multi-partitioning, which can support the best parallelization for some applications such as the line sweep computations. However, the implementation of multi-partitioning is very difficult and, as we know, there is none automatic parallelizing compiler supports such partitioning strategy. Though the dHPF compiler implemented multi-partitioning as a special extension for block style HPF partitioning, it still needs the programmer¿s participation to analyze the application and decide the data distribution scheme. In this paper, we present a global tiling transformation algorithm and a tile-to-processors mapping strategy called hyper-diagonal modular mapping, to implement the multi-partitioning strategy. The experimentation with NPB2.3-serial SP shows that the code generated by the compiler achieves scalable performance.

关键词： Concurrent computing Distributed computing computer applications Program processors Partitioning algorithms Tiles Laboratories computer architecture Partial differential equations Navier-Stokes equations

Logic Design of a 16-bit Bit-Slice Shifter for 64-bit RSFQ Microprocessors

学校读者我要写书评

暂无评论

Logic Design of a 16-bit Bit-Slice Shifter for 64-bit RSFQ M...

IEEE International Superconductive Electronics Conference (ISEC)

作者： Wei Xuan Guang-Ming Tang Pei-Yao Qu Zhi-Min Tang Xiao-Chun Ye Dong-Rui Fan Zhi-Min Zhang Ning-Hui Sun University of Chinese Academy of Sciences Beijing China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China

Logic design of a 16-bit bit-slice shifter for 64-bit superconducting rapid single-flux-quantum (RSFQ) microprocessors is proposed. The shifter supports three types of shift operations including logic shift, arithmetic shift and rotating shift. Each of 64-bit shift input operands is divided into four slices of 16-bit each. In order to simulate the digital function and timing of the proposed 16-bit bit-slice shifter, we design a logic-level simulation model based on the Open Dataset of CONNECT Cell Library for AIST ADP2. As the results of simulation, the information of RSFQ circuits, such as the number of Josephson junctions, area and latency of the 16-bit bit slice shifter can be obtained. The simulation results show that the proposed 16-bit bit-slice shifter can work correctly.

关键词：

Auto-tuning neural network quantization framework for collaborative inference between the cloud and edge

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Li, Guangli Liu, Lei Wang, Xueying Dong, Xiao Zhao, Peng Feng, Xiaobing State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

Recently, deep neural networks (DNNs) have been widely applied in mobile intelligent applications. The inference for the DNNs is usually performed in the cloud. However, it leads to a large overhead of transmitting data via wireless network. In this paper, we demonstrate the advantages of the cloud-edge collaborative inference with quantization. By analyzing the characteristics of layers in DNNs, an auto-tuning neural network quantization framework for collaborative inference is proposed. We study the effectiveness of mixed-precision collaborative inference of state-of-the-art DNNs by using ImageNet dataset. The experimental results show that our framework can generate reasonable network partitions and reduce the storage on mobile devices with trivial loss of accuracy. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks

Utilizing the instability in weakly supervised object detection

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Gao, Yan Liu, Boxiao Guo, Nan Ye, Xiaochun Wan, Fang You, Haihang Fan, Dongrui State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

Weakly supervised object detection (WSOD) focuses on training object detector with only image-level annotations, and is challenging due to the gap between the supervision and the objective. Most of existing approaches model WSOD as a multiple instance learning (MIL) problem. However, we observe that the result of MIL based detector is unstable, i.e., the most confident bounding boxes change significantly when using different initializations. We quantitatively demonstrate the instability by introducing a metric to measure it, and empirically analyze the reason of instability. Although the instability seems harmful for detection task, we argue that it can be utilized to improve the performance by fusing the results of differently initialized detectors. To implement this idea, we propose an end-to-end framework with multiple detection branches, and introduce a simple fusion strategy. We further propose an orthogonal initialization method to increase the difference between detection branches. By utilizing the instability, we achieve 52.6% and 48.0% mAP on the challenging PASCAL VOC 2007 and 2012 datasets, which are both the new state-of-the-arts. Copyright © 2019, The Authors. All rights reserved.

关键词： Stability

Design of Datapath Circuits for a Bit-Parallel 8-bit RSFQ Microprocessor

学校读者我要写书评

暂无评论

Design of Datapath Circuits for a Bit-Parallel 8-bit RSFQ Mi...

IEEE International Superconductive Electronics Conference (ISEC)

作者： Pei-Yao Qu Guang-Ming Tang Xiao-Chun Ye Dong-Rui Fan Zhi-Min Zhang Ning-Hui Sun University of Chinese Academy of Sciences Beijing China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China

Rapid single-flux-quantum (RSFQ) is expected to be the next generation integrated circuit technology because of its ultra-high-speed with ultra-low-power consumption. We propose datapath circuits for an 8-bit bit-parallel RSFQ microprocessor. The proposed datapath circuits process 8-bit data each clock cycle. Seven instructions are executed in the datapath, including ADD, ADDI, IN, OUT, LOADI, SRL and MOV. The datapath circuits consist of eight input ports, eight output ports, five multiplexers (MUXs), two 8-bit data registers and one 8-bit bit-parallel arithmetic logic unit (ALU). The datapath circuits contain 12 pipeline stages and 2993 JJs based on the Open Dataset of CONNECT Cell Library for AIST ADP2 without considering wiring cells. We perform digital simulation of the proposed datapath circuits. The simulation results show correct operation with the assumed frequency of 20 GHz.

关键词：

Pixel-semantic revise of position: Learning a one-stage object detector with a shared encoder-decoder

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Li, Qian Guo, Nan Ye, Xiaochun Fan, Dongrui Tang, Zhimin Chen, Honggang Li, Wenming State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

We analyze that different methods based channel or position attention mechanism give rise to different performance on scale, and some of state-of-the-art detectors applying feature pyramid are integrated with various variants convolutions with many mechanisms to enhance information, resulting in increasing runtime. This work addresses the difficulty by constructing an anchor-free detector with shared module consisting of encoder and decoder with attention mechanism. First, we consider different level features from backbone (e.g., ResNet-50) as the base features. Second, we feed the feature into a simple block, rather than various complex operations. Then, location and classification tasks are obtained by the detector head and classifier, respectively. At the same time, we use the semantic information to revise geometry locations. Additionally, we show that the detector is a pixel-semantic revise of position, universal, effective and simple to detect, especially large-scale objects. More importantly, this work compares different feature processing (e.g., mean, maximum or minimum) performance across channel. Finally, we present that our method improves detection AP of 3.8%, compared with state-of-the-art MNC based ResNet-101 on the standard MSCOCO baseline. Copyright © 2020, The Authors. All rights reserved.

关键词： Pixels

Video face recognition system: Retinaface-mnet-faster and secondary search

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Li, Qian Guo, Nan Ye, Xiaochun Fan, Dongrui Tang, Zhimin State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

Face recognition is widely used in the scene. However, different visual environments require different methods, and face recognition has a difficulty in complex environments. Therefore, this paper mainly experiments complex faces in the video. First, we design an image pre-processing module for fuzzy scene or under-exposed faces to enhance images. Our experimental results demonstrate that effective images pre-processing improves the accuracy of 0.11%, 0.2% and 1.4% on LFW, WIDER FACE and our datasets, respectively. Second, we propose RetinacFace-mnet-faster for detection and a confidence threshold specification for face recognition, reducing the lost rate. Our experimental results show that our RetinaFace-mnet-faster for 640×480 resolution on the Tesla P40 and single-thread improve speed of 16.7% and 70.2%, respectively. Finally, we design secondary search mechanism with HNSW to improve performance. Ours is suitable for large-scale datasets, and experimental results show that our method is 82% faster than the violent retrieval for the single-frame detection. Copyright © 2020, The Authors. All rights reserved.

关键词： Face recognition