With the development of Ethernet systems and the growing capacity of modem silicon technology, embedded communication networks are playing an increasingly important role in embedded and safety critical systems. Hardwa...With the development of Ethernet systems and the growing capacity of modem silicon technology, embedded communication networks are playing an increasingly important role in embedded and safety critical systems. Hardware/software co-design is a methodology for solving design problems in processor based embedded systems. In this work, we implemented a new 1-cycle pipeline microprocessor and a fast Ethemet transceiver and established a low cost, high performance embedded network controller, and designed a TCP/IP stack to access the Intemet. We discussed the hardware/software architecture in the forepart, and then the whole system-on-a-chip on Altera Stratix EP1S25F780C6 device. Using the FPGA environment and SmartBit tester, we tested the system＇s throughput. Our simulation results showed that the maximum throughput of Ethemet packets is up to 7 Mbps, that of UDP packets is up to 5.8 Mbps, and that of TCP packets is up to 3.4 Mbps, which showed that this embedded system can easily transmit basic voice and video signals through Ethemet, and that using only one chip can realize that many electronic devices access to the Intemet directly and get high performance.展开更多
Network processors are used in the core node of network to flexibly process packet streams. With the increase of performance, the power of network processor increases fast, and power and cooling become a bottleneck. A...Network processors are used in the core node of network to flexibly process packet streams. With the increase of performance, the power of network processor increases fast, and power and cooling become a bottleneck. Architecture-level power conscious design must go beyond low-level circuit design. Architectural power and performance tradeoff should be considered at the same time. Simulation is an efficient method to design modem network processor before making chip. In order to achieve the tradeoff between performance and power, the processor simulator is used to design the architecture of network processor. Using Netbeneh, Commubench benchmark and processor simulator-SimpleScalar, the performance and power of network processor are quantitatively evaluated. New performance tradeoff evaluation metric is proposed to analyze the architecture of network processor. Based on the high performance lnteI IXP 2800 Network processor eonfignration, optimized instruction fetch width and speed ,instruction issue width, instruction window size are analyzed and selected. Simulation resuits show that the tradeoff design method makes the usage of network processor more effectively. The optimal key parameters of network processor are important in architecture-level design. It is meaningful for the next generation network processor design.展开更多
A novel Bacterial Foraging Algorithm (BFA) based neural network is presented for image compression. To improve the quality of the decompressed images, the concepts of reproduction, elimination and dispersal in BFA are...A novel Bacterial Foraging Algorithm (BFA) based neural network is presented for image compression. To improve the quality of the decompressed images, the concepts of reproduction, elimination and dispersal in BFA are firstly introduced into neural network in the proposed algorithm. Extensive experiments are conducted on standard testing images and the results show that the pro- posed method can improve the quality of the reconstructed images significantly.展开更多
Rough set theory is a new soft computing tool, and has received much attention of researchers around the world. It can deal with incomplete and uncertain information. Now, it has been applied in many areas successfull...Rough set theory is a new soft computing tool, and has received much attention of researchers around the world. It can deal with incomplete and uncertain information. Now, it has been applied in many areas successfully. This paper introduces the basic concepts of rough set and discusses its applications in Web mining. In particular, some applications of rough set theory to intelligent information processing are emphasized.展开更多
We construct efficient quantum logic network for probabilistic cloning the quantum states used in imple mented tasks for which cloning provides some enhancement in performance.
Generalized hypercubes （denoted by Q（d1,d2,... ,dn）） is an important network topology for parallel processing computer systems. Some methods of forming big cycle from small cycles and links have been developed. Ba...Generalized hypercubes （denoted by Q（d1,d2,... ,dn）） is an important network topology for parallel processing computer systems. Some methods of forming big cycle from small cycles and links have been developed. Basing on which, we has proved that in generalized hypercubes, every edge can be contained on a cycle of every length from 3 to IV（G）I inclusive and all kinds of length cycles have been constructed. The edgepanciclieity and node-pancilicity of generalized hypercubes can be applied in the topology design of computer networks to improve the network performance.展开更多
Human information processing depends mainly on billions of neurons which constitute a complex neural network,and the information is transmitted in the form of neural spikes.In this paper,we propose a spiking neural ne...Human information processing depends mainly on billions of neurons which constitute a complex neural network,and the information is transmitted in the form of neural spikes.In this paper,we propose a spiking neural network(SNN),named MD-SNN,with three key features:(1) using receptive field to encode spike trains from images;(2) randomly selecting partial spikes as inputs for each neuron to approach the absolute refractory period of the neuron;(3) using groups of neurons to make decisions.We test MD-SNN on the MNIST data set of handwritten digits,and results demonstrate that:(1) Different sizes of receptive fields influence classification results significantly.(2) Considering the neuronal refractory period in the SNN model,increasing the number of neurons in the learning layer could greatly reduce the training time,effectively reduce the probability of over-fitting,and improve the accuracy by 8.77%.(3) Compared with other SNN methods,MD-SNN achieves a better classification;compared with the convolution neural network,MD-SNN maintains flip and rotation invariance(the accuracy can remain at 90.44% on the test set),and it is more suitable for small sample learning(the accuracy can reach 80.15%for 1000 training samples,which is 7.8 times that of CNN).展开更多
文摘With the development of Ethernet systems and the growing capacity of modem silicon technology, embedded communication networks are playing an increasingly important role in embedded and safety critical systems. Hardware/software co-design is a methodology for solving design problems in processor based embedded systems. In this work, we implemented a new 1-cycle pipeline microprocessor and a fast Ethemet transceiver and established a low cost, high performance embedded network controller, and designed a TCP/IP stack to access the Intemet. We discussed the hardware/software architecture in the forepart, and then the whole system-on-a-chip on Altera Stratix EP1S25F780C6 device. Using the FPGA environment and SmartBit tester, we tested the system＇s throughput. Our simulation results showed that the maximum throughput of Ethemet packets is up to 7 Mbps, that of UDP packets is up to 5.8 Mbps, and that of TCP packets is up to 3.4 Mbps, which showed that this embedded system can easily transmit basic voice and video signals through Ethemet, and that using only one chip can realize that many electronic devices access to the Intemet directly and get high performance.
基金Sponsored by the National Defence Research Foundation of China(Grant No.413460303).
文摘Network processors are used in the core node of network to flexibly process packet streams. With the increase of performance, the power of network processor increases fast, and power and cooling become a bottleneck. Architecture-level power conscious design must go beyond low-level circuit design. Architectural power and performance tradeoff should be considered at the same time. Simulation is an efficient method to design modem network processor before making chip. In order to achieve the tradeoff between performance and power, the processor simulator is used to design the architecture of network processor. Using Netbeneh, Commubench benchmark and processor simulator-SimpleScalar, the performance and power of network processor are quantitatively evaluated. New performance tradeoff evaluation metric is proposed to analyze the architecture of network processor. Based on the high performance lnteI IXP 2800 Network processor eonfignration, optimized instruction fetch width and speed ,instruction issue width, instruction window size are analyzed and selected. Simulation resuits show that the tradeoff design method makes the usage of network processor more effectively. The optimal key parameters of network processor are important in architecture-level design. It is meaningful for the next generation network processor design.
基金Supported by the National Natural Science Foundation of China (No.60572100)by the Royal Society (U.K.) International Joint Projects 2006/R3-Cost Share with NSFC (No.60711130233)
文摘A novel Bacterial Foraging Algorithm (BFA) based neural network is presented for image compression. To improve the quality of the decompressed images, the concepts of reproduction, elimination and dispersal in BFA are firstly introduced into neural network in the proposed algorithm. Extensive experiments are conducted on standard testing images and the results show that the pro- posed method can improve the quality of the reconstructed images significantly.
文摘Rough set theory is a new soft computing tool, and has received much attention of researchers around the world. It can deal with incomplete and uncertain information. Now, it has been applied in many areas successfully. This paper introduces the basic concepts of rough set and discusses its applications in Web mining. In particular, some applications of rough set theory to intelligent information processing are emphasized.
文摘We construct efficient quantum logic network for probabilistic cloning the quantum states used in imple mented tasks for which cloning provides some enhancement in performance.
基金This project is supported by National Natural Science Foundation of China （10671081）
文摘Generalized hypercubes （denoted by Q（d1,d2,... ,dn）） is an important network topology for parallel processing computer systems. Some methods of forming big cycle from small cycles and links have been developed. Basing on which, we has proved that in generalized hypercubes, every edge can be contained on a cycle of every length from 3 to IV（G）I inclusive and all kinds of length cycles have been constructed. The edgepanciclieity and node-pancilicity of generalized hypercubes can be applied in the topology design of computer networks to improve the network performance.
基金supported by the National Natural Science Foundation of China(Nos.61773312,61773307,and L1522023)the China Postdoctoral Science Foundation(No.2016M590949)the National Basic Research Program(973)of China(No.2015CB351703)
文摘Human information processing depends mainly on billions of neurons which constitute a complex neural network,and the information is transmitted in the form of neural spikes.In this paper,we propose a spiking neural network(SNN),named MD-SNN,with three key features:(1) using receptive field to encode spike trains from images;(2) randomly selecting partial spikes as inputs for each neuron to approach the absolute refractory period of the neuron;(3) using groups of neurons to make decisions.We test MD-SNN on the MNIST data set of handwritten digits,and results demonstrate that:(1) Different sizes of receptive fields influence classification results significantly.(2) Considering the neuronal refractory period in the SNN model,increasing the number of neurons in the learning layer could greatly reduce the training time,effectively reduce the probability of over-fitting,and improve the accuracy by 8.77%.(3) Compared with other SNN methods,MD-SNN achieves a better classification;compared with the convolution neural network,MD-SNN maintains flip and rotation invariance(the accuracy can remain at 90.44% on the test set),and it is more suitable for small sample learning(the accuracy can reach 80.15%for 1000 training samples,which is 7.8 times that of CNN).
