Hardware security issues have emerged in recent years as Internet of Things (IoT) devices have rapidly spread. Power analysis is one of the methods to detect anomalous operations, but it is hard to apply it to IoT devices where an operating system and various software programs are running and hence its power waveforms become more complex. In this paper, we propose an anomalous behavior detection method utilizing application-specific power behaviors extracted by steady-state power waveform, which is generated by LSTM (long short-term memory). The proposed method is based on extracting application-specific power behaviors by predicting steady-state power waveforms. At that time, by using LSTM, we can effectively predict steady-state power waveforms, even if they include one or more cycled waveforms and/or they are composed of many complex waveforms. In the experiment, we implement three normal application programs and one anomalous application program on a single board computer and apply the proposed method to it. The experimental results demonstrate that the proposed method successfully detects the anomalous power behavior of an anomalous application program, while the existing method cannot.

Recently, with the spread of Internet of Things (IoT) devices, embedded hardware devices have been used in a variety of everyday electrical items. Due to the increased demand for embedded hardware devices, some of the IC design and manufacturing steps have been outsourced to third-party vendors. Since malicious third-party vendors may insert malicious circuits, called hardware Trojans, into their products, developing an effective hardware Trojan detection method is strongly required. In this paper, we propose 25 hardware-Trojan features based on the structure of trigger circuits for machine-learning-based hardware Trojan detection. Combining the proposed features into 11 existing hardware-Trojan features, we totally utilize 36 hardware-Trojan features for classification. Then we classify the nets in an unknown netlist into a set of normal nets and Trojan nets based on the random-forest classifier. The experimental results demonstrate that the average true positive rate (TPR) becomes 63.6% and the average true negative rate (TNR) becomes 100.0%. They improve the average TPR by 14.7 points while keeping the average TNR compared to existing state-of-the-art methods. In particular, the proposed method successfully finds out Trojan nets in several benchmark circuits, which are not found by the existing method.

Due to the rapid growth in the information and telecommunications industries, an untrusted vendor might compromise the complicated supply chain by inserting hardware Trojans (HTs). Although hardware Trojan detection methods at gate-level netlists employing machine learning have been developed, the training dataset is insufficient. In this paper, we propose a data augmentation method for machine-learning-based hardware Trojan detection. Our proposed method replaces a gate in a hardware Trojan circuit with logically equivalent gates. The experimental results demonstrate that our proposed method successfully enhances the classification performance with all the classifiers in terms of the true positive rates (TPRs).

Vehicle routing problems (VRPs) can be solved as optimization problems. Practical applications of the VRPs are involved in various areas including manufacturing, supply chain, and tourism. Conventional approaches using von Neumann computers obtain good approximate solutions to the optimization problems, but conventional approaches show disadvantages of computation costs in large-scale or complex problems due to the combinatorial explosion. Oppositely, Ising machines or quantum annealing machines are non-von Neumann computers that are designed to solve complex optimization problems. In this paper, we propose an Ising-machine based approach for the vehicle routing problem with balanced pick-up (VRPBP). The development of the VRPBP is motivated by postal items pick-up services in the real-world. Our approach includes various features of VRP variants. We propose a 2-phase approach to solve the VRPBP and key elements in each phase are mapped onto quadratic unconstrained binary optimization (QUBO) forms. Specifically, the first phase belongs to the clustering phase which is an extension to the knapsack problem with additional distance and load balancing concerns. The second phase is mapped to the traveling salesman problem. Experimental results of our approach are evaluated in terms of solution quality and computation time compared with conventional approaches.

SUMMARY Ising machines have attracted attention as they are expected to solve combinatorial optimization problems at high speed with Ising models corresponding to those problems. An induced subgraph isomorphism problem is one of the decision problems, which determines whether a specific graph structure is included in a whole graph or not. The problem can be represented by equality constraints in the words of combinatorial optimization problem. By using the penalty functions corresponding to the equality constraints, we can utilize an Ising machine to the induced subgraph isomorphism problem. The induced subgraph isomorphism problem can be seen in many practical problems, for example, finding out a particular malicious circuit in a device or particular network structure of chemical bonds in a compound. However, due to the limitation of the number of spin variables in the current Ising machines, reducing the number of spin variables is a major concern. Here, we propose an efficient Ising model mapping method to solve the induced subgraph isomorphism problem by Ising machines. Our proposed method theoretically solves the induced subgraph isomorphism problem. Furthermore, the number of spin variables in the Ising model generated by our proposed method is theoretically smaller than that of the conventional method. Experimental results demonstrate that our proposed method can successfully solve the induced subgraph isomorphism problem by using the Ising-model based simulated annealing and a real Ising machine.

Ising machines have attracted attention, which is expected to obtain better solutions of various combinatorial optimization problems at high speed by mapping the problems to natural phenomena. A slot-placement problem is one of the combinatorial optimization problems, regarded as a quadratic assignment problem, which relates to the optimal logic-block placement in a digital circuit as well as optimal delivery planning. Here, we propose a mapping to the Ising model for solving a slot-placement problem with additional constraints, called a constrained slot-placement problem, where several item pairs must be placed within a given distance. Since the behavior of Ising machines is stochastic and we map the problem to the Ising model which uses the penalty method, the obtained solution does not always satisfy the slot-placement constraint, which is different from the conventional methods such as the conventional simulated annealing. To resolve the problem, we propose an interpretation method in which a feasible solution is generated by post-processing procedures. We measured the execution time of an Ising machine and compared the execution time of the simulated annealing in which solutions with almost the same accuracy are obtained. As a result, we found that the Ising machine is faster than the simulated annealing that we implemented.

A pedestrian dead reckoning method, or PDR method in short, is one of the positioning methods in indoor environments, which estimates user's positions by using sensors such as acceleration and angular velocity sensors. When we consider using a smartphone as a PDR sensor device, it has various carrying modes such as holding it directly and carrying it inside a pocket. How to deal with these various carrying modes is the great concern in PDR using a smartphone. In this paper, we propose a PDR method based on a combination of a smartphone and a smartwatch. By synchronizing smartphone and smartwatch sensors effectively, the proposed method can successfully reduce drift errors and thus estimate accurate user's positions, compared to just using a smartphone. Furthermore, even when the user carries his/her smartphone in various carrying modes, the proposed method still realizes accurate PDR. The experimental results demonstrate that the positioning errors are reduced by approximately 87.5% on average compared to the existing method.

In an amusement park, an attraction-visiting route considering the waiting time and traveling time improves visitors' satisfaction and experience. We focus on Ising machines to solve the problem, which are recently expected to solve combinatorial optimization problems at high speed by mapping the problems to Ising models or quadratic unconstrained binary optimization (QUBO) models. We propose a mapping of the visiting-route recommendation problem in amusement parks to a QUBO model for solving it using Ising machines. By using an actual Ising machine, we could obtain feasible solutions 15 times faster with almost the same accuracy as the simulated annealing method for the visiting-route recommendation problem.

Single-board computers have been widely spread and used in a variety of situations, where they may be requested to operate under low-energy conditions or with an unstable power supply. Utilizing non-volatile memory (NVM) retaining data without power must be one of the effective solutions to tackle this issue. However, compared to volatile memory such as SRAM and DRAM, NVM consumes more energy in writing operations. In this paper, we propose an effective energy reduction method for RISC-V architecture, targeting one of NVMs called spin-transfer torque RAMs (STT-RAM). Firstly, we thoroughly investigate the writing bit patterns to registers in RISC-V architecture for various typical application programs and find out that most of them can be classified into three patterns, in which most bits in writing 32-bit data are 0s (zero's). Secondly, we propose an energy-reduced register-writing method utilizing these frequent writing bit patterns. In this method, when a writing data falls into one of the three frequent bit writing patterns above, we just write the bit pattern type into the extra bits and do not write actual data into registers and hence we can reduce the write energy in NVM register writing extremely. Experimental results on RISC-V architecture demonstrate that the energy consumption is reduced by 12.5%-53.8% by using our proposed method compared to the baseline architecture.

Ising machines have recently attracted much attention because they are expected to solve combinatorial optimization problems efficiently. We focus on an Ising machine whose algorithm is based on parallel tempering (PT), and experimentally evaluate the performance of the Ising machine for MIN-CUT problems. Experimental results show that the Ising machine outperforms a famous graph partitioning solver in terms of the quality of solution and the time-to-target-solution.

The differences in performance among binary-integer encodings in an Ising machine, which can solve combinatorial optimization problems, are investigated. Many combinatorial optimization problems can be mapped to find the lowest-energy (ground) state of an Ising model or its equivalent model, the Quadratic Unconstrained Binary Optimization (QUBO). Since the Ising model and QUBO consist of binary variables, they often express integers as binary when using Ising machines. A typical example is the combinatorial optimization problem under inequality constraints. Here, the quadratic knapsack problem is adopted as a prototypical problem with an inequality constraint. It is solved using typical binary-integer encodings: one-hot encoding, binary encoding, and unary encoding. Unary encoding shows the best performance for large-sized problems.

Recently, the great demand for integrated circuits (ICs) drives third parties to be involved in IC design and manufacturing steps. At the same time, the threat of injecting a malicious circuit, called a hardware Trojan, by third parties has been increasing. Machine learning is one of the powerful solutions for detecting hardware Trojans. How-ever, a weakness of such a machine-learning-based classification method against adversarial examples (AEs) has been reported, which causes misclassification by adding perturbation in input samples. This paper firstly proposes a framework generating adversarial examples for hardware-Trojan detection at gate-level netlists utilizing neural networks. The proposed framework replaces hardware Trojan circuits with logically equivalent ones, and makes it difficult to detect them. Secondly, we propose a Trojan-net concealment degree (TCD) and a modification evaluating value (MEV) as measures of the amount of modifications. Finally, based on the MEV, we pick up adversarial modification patterns to apply to the circuits against hardware-Trojan detection. The experimental results using benchmarks demonstrate that the proposed framework successfully decreases the true positive rate (TPR) by a maximum of 30.15 points.

As smartphones and tablets are widely spread and used, route recommendation and guidance services have become commonplace. Conventional services in route recommendation and guidance try to give best routes in terms of route length, time required, and train/bus fares, whereas even different users are given the same route when inputting the same parameters. However, each user has various preferences from the aspect of safety and comfort. It is strongly desirable to reflect the user’s preferences in route recommendation and recommend the most preferable route to every user. Since user’s preferences are extremely vague and complicated, how to evaluate them in route recommendation is one of the key problems there. In this paper, we propose a route recommendation method, called P-UCT method, considering individual user’s preferences utilizing Monte-Carlo tree search. In the proposed method, we firstly ex-tract route features based on the route recommendation history of every user and construct a route evaluator based on Support Vector Machine (SVM). After that, the method generates a random route from a start point to an end point by Monte-Carlo tree search. The route evaluator determines how well every generated route matches the user’s preferences. By repeating the evaluation, the method obtains the route, which must be closest to the user’s preferences. Experimental results demonstrate that the proposed method outperforms the existing method from the viewpoint of the average evaluation scores. They also demonstrate that the proposed method provides the recommended route reflecting the user’s individual preferences even if it learns the recommended route history of areas in different situations.

In IoT (Internet-of-Things) era, the number and variety of hardware devices becomes continuously increasing. Several IoT devices are utilized at infrastructure equipments. How to maintain such IoT devices is a serious concern. Capacitance measurement is one of the powerful ways to detect anomalous states in the structure of the hardware devices. Particularly, measuring capacitance while the hardware device is running is a major challenge but no such researches proposed so far. This paper proposes a capacitance measuring device which measures device capacitance in operation. We firstly combine the AC (alternating current) voltage signal with the DC (direct current) supply voltage signal and generates the fluctuating signal. We supply the fluctuating signal to the target device instead of supplying the DC supply voltage. By effectively filtering the observed current in the target device, the filtered current can be proportional to the capacitance value and thus we can measure the target device capacitance even when it is running. We have implemented the proposed capacitance measuring device on the printed wiring board with the size of 95 mm × 70 mm and evaluated power consumption and accuracy of the capacitance measurement. The experimental results demonstrate that power consumption of the proposed capacitance measuring device is reduced by 65% in low-power mode from measuring mode and proposed device successfully measured capacitance in 0.002 µF resolution.

Cybersecurity has become a serious concern in our daily lives. The malicious functions inserted into hardware devices have been well known as hardware Trojans. In this letter, we propose a hardware-Trojan classification method at gate-level netlists utilizing boundary net structures. We first use a machine-learning-based hardware-Trojan detection method and classify the nets in a given netlist into a set of normal nets and a set of Trojan nets. Based on the classification results, we investigate the net structures around the boundary between normal nets and Trojan nets, and extract the features of the nets mistakenly identified to be normal nets or Trojan nets. Finally, based on the extracted features of the boundary nets, we again classify the nets in a given netlist into a set of normal nets and a set of Trojan nets. The experimental results demonstrate that our proposed method outperforms an existing machine-learning-based hardware-Trojan detection method in terms of its true positive rate.

With the widespread use of Internet of Things (IoT) devices in recent years, we utilize a variety of hardware devices in our daily life. On the other hand, hardware security issues are emerging. Power analysis is one of the methods to detect anomalous operations, but it is hard to apply it to IoT devices where an operating system and various software programs are running. In this paper, we propose an anomalous behavior detection method for an IoT device by extracting application-specific power behaviors. First, we measure a power consumption of an IoT device, and obtain the power waveform. Next, we extract an application-specific power waveform by eliminating a steady factor from the obtained power waveform. Finally, we extract feature values from the application-specific power waveform and detect an anomalous behavior by utilizing the local outlier factor (LOF) method. The experimental results using a single board computer demonstrate that the proposed method successfully detects the anomalous power behavior of an anomalous application program.

Recently, with the spread of Internet of Things (IoT) devices, embedded hardware devices have been used in a variety of everyday electrical items. Due to the increased demand for embedded hardware devices, some of the IC design and manufacturing steps have been outsourced to third-party vendors. Since malicious third-party vendors may insert hardware Trojans into their products, developing an effective hardware Trojan detection method is strongly required. In this paper, we evaluate hardware Trojan detection methods using neural networks and random forests at gate-level intellectual property (IP) cores that contain more than 10,000 nets. First, we extract 11 features for each net in a given netlist, and learn them with neural networks and random forests. Then, we classify the nets in an unknown netlist into a set of normal nets and Trojan nets based on the learned classifiers. The experimental results demonstrate that the average true positive rate becomes 84.6% and the average true negative rate becomes 95.1%, which is sufficiently high accuracy compared to existing evaluations.

In this study, we propose a novel stochastic number generator architecture and prove that the resulting circuit can deliver independent stochastic numbers and improve the accuracy of the calculation results obtained using some recent conventional stochastic computing-based arithmetic circuits. This study is motivated by the increasingly important role of stochastic computing in various fields, such as the digital circuit design, where the stochastic number generators are responsible for a significant share of the hardware cost.

Existing studies focus on text information while ignoring category and data source information, both of which are verified to be important in interpreting sentiments in travel comments in this paper. Furthermore, the unique linguistic characteristics of Japanese cause difficulty in applying the conventional token-based word segmentation methods to Japanese comments directly. In this paper, we propose a method of stem-based segmentation based on Japanese linguistic characteristics and incorporate it with category and data source information into a hierarchical network model for document-level sentiment classification. Empirical results of our proposed model outperform existing models on a real-world dataset.

The popularization of IoT devices made image processing very common for users. Image formats used in hardware abound since there are varieties of IoT devices. Conversion of image formats in hardware is relatively complicated compared with other calculation. This paper focuses on conversion of image resolution, especially image reduction. By expressing images with stochastic numbers, this paper proposes an image format which can be treated to be in several resolution with one data. From experimental evaluations, we found that the proposed image format enables image reduction by pixel average to be implemented into hardware with lower costs compared with conventional pixel average using binary numbers. Also, image magnification using the proposed image format can restore the original image, while conventional image magnification cannot.

Recently, due to the increase of outsourcing in integrated circuit (IC) design and manufacturing, the threat of injecting a malicious circuit, called a hardware Trojan, by third party has been increasing. Machine learning has been known to produce a powerful model to detect hardware Trojans. But it is recently reported that such a machine learning based detection is weak against adversarial examples (AEs), which cause misclassification by adding perturbation in input data. Referring to the existing studies on adversarial examples, most of which are discussed in the field of image processing, this paper first proposes a framework generating adversarial examples for hardware-Trojan detection for gate-level netlists utilizing neural networks. The proposed framework replaces hardware Trojan circuits with logically equivalent circuits, and makes it difficult to detect them. Second, we define Trojan-net concealment degree (TCD) as a possibility of misclassification, and modification evaluating value (MEV) as a measure of the amount of modifications. Third, judging from MEV, we pick up adversarial modification patterns to apply to the circuits against hardware-Trojan detection. The experimental results using benchmarks demonstrate that the proposed framework successfully decreases true positive rate (TPR) by at most 30.15 points.

Annealing machines have been developed as non-von Neumann computers aimed at solving combinatorial optimization problems efficiently. To use annealing machines for solving combinatorial optimization problems, we have to represent the objective function and constraints by an Ising model, which is a theoretical model in statistical physics. Further, it is necessary to transform the Ising model according to the hardware limitations. In the transformation, the process of effectively reducing the bit-widths of coefficients in the Ising model has hardly been studied so far. Thus, when we consider the Ising model with a large bit-width, a naive method, which means right bit-shift, has to be applied. Since it is expected that obtaining highly accurate solutions is difficult by the naive method, it is necessary to construct a method for efficiently reducing the bit-width. This paper proposes methods for reducing the bit-widths of interaction and external magnetic field coefficients in the Ising model and proves that the reduction gives theoretically the same ground state of the original Ising model. The experimental evaluations also demonstrate the effectiveness of our proposed methods.

We propose a novel type of minor-embedding (ME) in simulated-annealing-based Ising machines. The Ising machines can solve combinatorial optimization problems. Many combinatorial optimization problems are mapped to find the ground (lowest-energy) state of the logical Ising model. When connectivity is restricted on Ising machines, ME is required for mapping from the logical Ising model to a physical Ising model, which corresponds to a specific Ising machine. Herein we discuss the guiding principle of ME design to achieve a high performance in Ising machines. We derive the proposed ME based on a theoretical argument of statistical mechanics. The performance of the proposed ME is compared with two existing types of MEs for different benchmarking problems. Simulated annealing shows that the proposed ME outperforms existing MEs for all benchmarking problems, especially when the distribution of the degree in a logical Ising model has a large standard deviation. This study validates the guiding principle of using statistical mechanics for ME to realize fast and high-precision solvers for combinatorial optimization problems.

Low-density parity-check (LDPC) codes have previously been considered as combinatorial optimization problems (COPs) in respect to its decoding. However, after defining it as such, none have gone so far as to convert the LDPC code into a quadratic unconstrained binary optimization (QUBO) problem. Thus, a new method is created: one that converts the LDPC code to a QUBO problem, inputs the QUBO problem into Ising machines (computers based on the Ising model that are designed to solve the QUBO problem), obtains the QUBO solution and converts it to a LDPC solution. By utilizing an actual Ising machine, LDPC solutions with code length of 256-bits have been obtained with an accuracy of 93.9% by average annealing time 214.0ms. The benefit of this newfound methodology goes beyond its theoretical imprint of obtaining LDPC solutions more accurately. It has only been a few years since the Ising machine has been developed. Therefore, in formulating this method, one expands the currently scope of studies involving Ising machines, helping current and future researchers unlock its full range of capabilities and possibilities.

Ising machines are a new type of non-Neumann computer that specializes in solving combinatorial optimization problems efficiently. The input form of Ising machines is the energy function of the Ising model or quadratic unconstrained binary optimization form, and Ising machines operate to search for a condition to minimize the energy function. We describe the theory of Ising machines and the present status of the Ising machines, software for Ising machines, and applications using Ising machines.

A heuristic algorithm is one of the approaches to solve an NP-hard problem. In order to enhance the capability of the system, heterogeneous computing is often adapted. In this paper, we propose an FPGA-based heterogeneous solver for three-dimensional routing. The proposed system is implemented into multiple FPGA boards and a single-board computer. The experimental results demonstrate that the proposed system outperforms a single FPGA system.

Stochastic computing is a computation method which can implement arithmetic operations by simple logic circuits. Stochastic numbers are used in this method, whose values are defined by their bit streams' appearance rates of 1's. As a nature of stochastic computing, changing the length of the input stochastic numbers will change the whole circuit's accuracy. However, in some implementations with re-convergence paths, the circuit itself will cause errors, i.e., the length of the input stochastic numbers does not change that circuit's accuracy. This paper proposes a stochastic number duplicator whose outputs differ every time and are all independent. This stochastic number duplicator has a scalable structure by changing the numbers of flip-flops for bit re-arrangement. From the experimental evaluations and discussions, we clarify that the proposed stochastic number duplicator enables accuracy-flexible circuits.

To help people smoothly and efficiently make travel decisions, we utilize the advantages of travel comments posted by thousands of other travelers. In this paper, we analyze the feasibility of exploring landmark activity queries and representative examples from Japanese travel comments. Contributions in this paper include a framework for extracting activity concerned keywords and queries, quantifying the relationship between landmark activities and comment contents. An evaluation of activity-example extraction is conducted in two case studies through 18,939 travel comments.

© 2019, Springer Nature Switzerland AG. This paper proposes a personalized landmark recommendation algorithm aiming at exploring new sights into the determinants of landmark satisfaction prediction. We gather 1,219,048 user-generated comments in Tokyo, Shanghai and New York from four travel websites. We find that users have diverse satisfaction on landmarks those findings, we propose an effective algorithm for personalize landmark satisfaction prediction. Our algorithm provides the top-6 landmarks with the highest satisfaction to users for a one-day trip plan our proposed algorithm has better performances than previous studies from the viewpoints of landmark recommendation and landmark satisfaction prediction.

© 2018 IEEE. This paper proposes a personalized landmark recommendation algorithm based on the prediction of users' satisfaction on landmarks. We have accumulated 270,239 user-generated comments from travel websites of Ctrip, Jaran and TripAdvisor for 196 landmarks in Tokyo, Japan. We find that users do have different satisfaction on landmarks depending on their commonly used languages and travel websites. Then we establish a database for landmarks with abundant and accurate landmark type and landmark satisfaction information. Finally, we propose an effective personalized landmark satisfaction prediction algorithm based on users' landmark type, language and travel website preferences. After that, landmarks with the top-6 highest satisfaction are provided to the user for a one-day visit plan in Tokyo. Experimental results demonstrate that the proposed algorithm can recommend landmarks that fit the user's preferences and our algorithm also successfully predicts the user's landmark satisfaction with a low error rate less than 7%, which is superior to other previous studies.

This paper proposes a robust AES (advanced encryption standard) circuit for delay variation. In our proposed AES circuit, suspicious timing error prediction circuits (STEPCs) and their associating gating circuit are incorporated into a normal AES circuit to predict timing errors. STEPCs are inserted between inter-module connections and thus we can monitor almost all of the signal paths between registers and effectively prevent timing errors. The simulation results demonstrate that our AES circuit with STEPCs can be overclocked by up to 1.66X with just 8.05% area overheads.

Fast Fourier transform (FFT) is used in various applications such as signal processings and developing a high-speed FFT processor is quite required. In this paper, we propose a high-speed FFT processor based on selector logics. The selector-based FFT processor is constructed by focusing on the subtract-multiplication operations and partly applying selector logics to them. Furthermore, we implement the selector-based FFT processor on a Xilinx FPGA. Experimental results show that our proposed FFT processor can improve the processing speed by up to 21% and also reduce the number of LUTs by up to 33% compared with a naive FFT processor.

Multiplication with a large number of digits is heavily used when processing data encrypted by a fully homomorphic encryption, which is a bottleneck in computation time. An algorithm utilizing fast number theoretic transform (FNTT) is known as a high-speed multiplication algorithm and the further speeding up is expected by implementing the FNTT process on an FPGA. A high-level synthesis tool enables efficient hardware implementation even for FNTT with a large number of points. In this paper, we propose a methodology for optimizing the loop structure included in a software description of FNTT so that the performance of the synthesized FNTT processor can be maximized. The loop structure optimization is considered in terms of loop flattening and trip count reduction. We implement a 65,536-point FNTT processor with the loop structure optimization on an FPGA, and demonstrate that it can be executed 6.9 times faster than the execution on a CPU.

Cell phones with GPS function as well as GPS loggers are widely used and users' geographic information can be easily obtained. However, still battery consumption in these mobile devices is main concern and then obtaining GPS positioning data so frequently is not allowed. In this paper, a stayed location estimation method for sparse GPS positioning information is proposed. After generating initial clusters from a sequence of measured positions, the e ective radius is set for every cluster based on positioning accuracy and the clusters are merged e ectively using it. After that, short-time clusters are removed temporarily but measured positions included in them are not removed. Then the clusters are merged again, taking all the measured positions into consideration. This process is performed twice, in other words, two-stage short-time cluster removal is performed, and finally accurate stayed location estimation is realized even when the GPS positioning interval is five minutes or more. Experiments demonstrate that the total distance error between the estimated stayed location and the true stayed location is reduced by more than 33% and also the proposed method much improves F1 measure compared to conventional state-of-the-art methods.

Recently, cybersecurity has become a serious concern for us. For example, the threats of hardware Trojans (malfunctions inserted into hardware devices) have appeared. Since hardware vendors often outsource parts of their hardware products to third-party vendors, the risk of hardware-Trojan insertion has been increased. Especially in the hardware design step, malicious vendors have a chance to insert hardware Trojans easily. In this paper, we propose a hardware-Trojan classification method utilizing boundary net structures. To begin with, we use a machine-learning-based hardware-Trojan detection method and classify the nets in a given netlist into a set of normal nets and that of Trojan nets. Based on the classification, we investigate the nets around the boundary between normal nets and Trojan nets and extract the features of the nets identified to be normal nets or Trojan nets mistakenly. Finally, using the classification results of machine-learning-based hardware-Trojan detection and the extracted features of the boundary nets, we classify the nets in a given netlist into a set of normal nets and that of Trojan nets again. The experimental results demonstrate that our method outperforms an existing machine-learning-based hardware-Trojan detection method in terms of true positive rate.

This paper proposes a road-illuminance level inference method based on the naive Bayesian analysis. We investigate quantities and types of road lights and landmarks with a large set of roads in real environments and reorganize them into two safety classes, safe or unsafe, with seven road attributes. Then we carry out data learning using three types of datasets according to different groups of the road attributes. Experimental results demonstrate that the proposed method successfully classifies a set of roads with seven attributes into safe ones and unsafe ones with the accuracy of more than 85%, which is superior to other machine-learning based methods and a manual-based method.

A scan chain is used by scan-path test, one of design-for-test techniques, which can control and observe internal registers in an LSI chip. On the other hand, a scan-based side-channel attack is focused on which can restore secret information by exploiting the scan data obtained from a scan chain inside the crypto chip during cryptographic processing. In this paper, we propose a scan-based attack method against a hash generator circuit called HMAC-SHA- 256. Our proposed method is composed of three steps
Firstly, we isolate 64 bit-transition groups from a scan data using scan signatures based on the property of the HMAC-SHA-256 algorithm. Secondly, we classify these 64 bittransition groups into 32 pairs. Lastly, we find out the correspondence between the scan data and the internal registers in the target HMAC-SHA-256 circuit. Our proposed method restores the secret information by the three steps above, even if the scan chain includes registers other than the target hash generator circuit and hence it becomes too long. Experimental results show that our proposed method successfully restores two secret keys of the HMAC-SHA-256 circuit using up to 425 input messages in 7.5 hours.

Since hardware production become inexpensive and international, hardware vendors often outsource their products to third-party vendors. Due to the situation, malicious vendors can easily insert malfunctions (also known as 'hardware Trojans') to their products. In this paper, we experimentally evaluate a machine-learning-based hardware-Trojan detection method using several hardware Trojans we designed. To begin with, we design three types of hardware Trojans and insert them to simple RS232 transceiver circuits. After that, we learn known netlists, where we know which nets are Trojan ones or normal ones beforehand, using a machine-learning-based hardware-Trojan detection method with a support vector machine (SVM) classifier. Finally, we classify the nets in the designed hardware-Trojan-inserted netlists into a set of Trojan nets and that of normal nets using the learned classifier. The experimental results demonstrate that the hardware-Trojan detection method with the SVM-based approach can detect a part of hardware Trojans we designed.

In this paper, a low cost and high speed CSD-based symmetric transpose block FIR design was proposed for low cost digital signal processing. First, the existing area-efficient CSD-based multiplier was optimized by considering the reusability and the symmetry of coefficients for area reduction. Second, the position of the input register was changed for high speed transpose block FIR processing in which half of the number of required multipliers can be saved. When compared with the existing block FIR designs, the proposed FIR design can increase the data rate from 238.66 MHz to 373.13 MHz while saving 10.89% area and 21.30% energy consumption as well.

In this paper, we propose a floorplan-driven highlevel synthesis algorithm utilizing both volatile and non-volatile registers for hybrid energy-harvesting systems. In our algorithm, we firstly introduce an idea of safety line candidates. Based on them, we perform safety-line (SL) scheduling so that every operation does not cross the safety line candidates and then perform volatile/non-volatile register binding so that all the data crossing the safety line candidates are stored into non-violate registers. We can safely restore all the data and re-start the circuit operation from every safety line candidate, even if the power shut-off occurs while running the circuit. Experimental results show that our algorithm reduces average latency by 30.76% and the average energy consumption by 24.94% compared to the naive algorithm when sufficient energy is given (normal mode). Experimental results also show that our algorithm reduces average latency by 30.58% compared to the naive algorithm by reducing rollback execution if a small amount of energy is given (energy-harvesting mode).

As semiconductor technology continues scaling down, the reliability issue has become much more critical than ever before. Unlike traditional hard-errors caused by permanent physical damage which can't be recovered in field, soft errors are caused by radiation or voltage/current fluctuations that lead to transient changes on internal node states, thus they can be viewed as temporary errors. However, due to the unpredictable occurrence of soft errors, it is desirable to develop soft error tolerant designs. For this reason, soft error tolerant design techniques have gained great research interest. In this paper, we will explain the soft error mechanism and then review the existing soft error tolerant design techniques with particular emphasis on SEH family because they can achieve low power consumption and small performance overhead as well.

Indoor positioning without GPS is one of the most important problems in indoor pedestrian navigation. In this paper, we propose an accurate indoor positioning algorithm using a particle filter based on a floormap, where we use the proximity of the Bluetooth beacons as well as acceleration and geomagnetic sensors. In designing the likelihood function in the particle filter, we effectively use the proximity of the Bluetooth beacons, which just gives rough distance to the target beacon but more stable than conventional RSSI-based distance estimation. In addition to that, by effectively utilizing a floormap, the accumulated positioning errors due to the acceleration and geomagnetic sensors are much reduced. Moreover, when the radio waves from the Bluetooth beacons are blocked by obstacles, we can also take it into account in designing the likelihood function in the particle filter. Experimental results demonstrate that our algorithm can reduce the indoor positioning errors by up to 79% compared to several conventional algorithms.

Cell phones with GPS function as well as GPS loggers are widely used and we can easily obtain users' geographic information. However, still battery consumption in these mobile devices is main concern and then we are not allowed to obtain GPS positioning data so frequently. In this paper, we propose a stayed location estimation method for sparse GPS positioning data. After generating initial clusters from a sequence of measured positions, we set the effective radius for every cluster based on positioning accuracy and merge the clusters effectively using it. After that, we temporarily remove short-time clusters but do not remove measured positions included in them. Then we merge the clusters again, taking all the measured positions into consideration. We perform this process twice, i.e, we perform two-stage short-time cluster removal, and finally realize accurate stayed location estimation even when the GPS positioning interval is five minutes or more. Experiments demonstrate that the total distance error between the estimated stayed location and the true stayed location is reduced by more than 50% compared to a conventional state-of-the-art method.

Travel route recommendation can strongly influence users' satisfaction and the success of touristic businesses. This paper proposes a personalized travel recommendation algorithm with time planning. We use landmark categorization and region clustering to obtain effective elements. Then we build a travel map to generate all possible travel routes. Our proposed algorithm has higher precision in landmark recommendation and time planning than thoes in previous algorithms.

A scan-based side-channel attack is still a real threat against a crypto circuit as well as a hash generator circuit, which can restore secret information by exploiting the scan data obtained from scan chains inside the chip during its processing. In this paper, we propose a scan-based attack method against a hash generator circuit called HMAC-SHA-256. Our proposed method restores the secret information by finding out the correspondence between the scan data obtained from a scan chain and the internal registers in the target HMAC-SHA-256 circuit, even if the scan chain includes registers other than the target hash generator circuit and an attacker does not know well the hash generation timing. Experimental results show that our proposed method successfully restores two secret keys of the HMAC-SHA-256 circuit in at most 6 hours.

As application hardware designs and implementations in a short term are required, high-level synthesis is more and more essential EDA technique nowadays. In deep-submicron era, interconnection delays are not negligible even in high-level synthesis thus distributed-register and - controller architectures (DR architectures) have been proposed in order to cope with this problem. It is also profitable to take data-bitwidth into account in high-level synthesis. In this paper, we propose a bitwidth-aware high-level synthesis algorithm using operation chainings targeting Tiled-DR architectures. Our proposed algorithm optimizes bitwidths of functional units and utilizes the vacant tiles by adding some extra functional units to realize effective operation chainings to generate high performance circuits without increasing the total area. Experimental results show that our proposed algorithm reduces the overall latency by up to 47% comparedtothe conventional approach without area overheads by eliminating unnecessary bitwidths and adding efficient extra FUs for Tiled-DR architectures.

In this paper, we propose a safe and comprehensive route finding algorithm for pedestrians based on lighting and landmark conditions. Safety and comprehensiveness can be predicted by the five possible indicators: (1) lighting conditions, (2) landmark visibility, (3) landmark effectiveness, (4) turning counts along a route, and (5) road widths. We first investigate impacts of these five indicators on pedestrians' perceptions on safety and comprehensiveness during route findings. After that, a route finding algorithm is proposed for pedestrians. In the algorithm, we design the score based on the indicators (1), (2), (3), and (5) above and also introduce a turning count reduction strategy for the indicator (4). Thus we find out a safe and comprehensive route through them. In particular, we design daytime score and nighttime score differently and find out an appropriate route depending on the time periods. Experimental simulation results demonstrate that the proposed algorithm obtains higher scores compared to several existing algorithms. We also demonstrate that the proposed algorithm is able to find out safe and comprehensive routes for pedestrians in real environments in accordance with questionnaire results.

Recently, the requirement for non-volatile memory on embedded systems has increased because they can be applied with normally-off and power gating technologies to. However, they have a lower endurance than volatile memories. When data is encoded as a write-reduction code appropriately, the endurance of non-volatile memory can be enhanced by writing the encoded data into the memory. We propose a highly effective write-reduction method for a multi-level cell (MLC) non-volatile memory focusing on the write-reduction code (WRC) as the optimal bit-write reduction method. The WRC can be applied only to single-level cell non-volatile memory. The proposed method generates a cell-write reduction code based on the WRC
the cell has multiple bits as the holdable data. Our proposed method achieves a cell-write reduction by 31.6% compared to the conventional method.

Recently, due to the increase of outsourcing in IC design, it has been reported that malicious third-party vendors often insert hardware Trojans into their ICs. How to detect them is a strong concern in IC design process. The features of hardware-Trojan infected nets (or Trojan nets) in ICs often differ from those of normal nets. To classify all the nets in netlists designed by third-party vendors into Trojan ones and normal ones, we have to extract effective Trojan features from Trojan nets. In this paper, we first propose 51 Trojan features which describe Trojan nets from netlists. Based on the importance values obtained from the random forest classifier, we extract the best set of 11 Trojan features out of the 51 features which can effectively detect Trojan nets, maximizing the F-measures. By using the 11 Trojan features extracted, the machine-learning based hardware Trojan classifier has achieved at most 100% true positive rate as well as 100% true negative rate in several TrustHUB benchmarks and obtained the average F-measure of 74.6%, which realizes the best values among existing machine-learning-based hardware-Trojan detection methods.

Recently, due to the increase of outsourcing in IC design, it has been reported that malicious third-party vendors often insert hardware Trojans into their ICs. How to detect them is a strong concern in IC design process. The features of hardware-Trojan infected nets (or Trojan nets) in ICs often differ from those of normal nets. To classify all the nets in netlists designed by third-party vendors into Trojan ones and normal ones, we have to extract effective Trojan features from Trojan nets. In this paper, we first propose 51 Trojan features which describe Trojan nets from netlists. Based on the importance values obtained from the random forest classifier, we extract the best set of 11 Trojan features out of the 51 features which can effectively detect Trojan nets, maximizing the F-measures. By using the 11 Trojan features extracted, the machine-learning based hardware Trojan classifier has achieved at most 100% true positive rate as well as 100% true negative rate in several TrustHUB benchmarks and obtained the average F-measure of 74.6%, which realizes the best values among existing machine-learning-based hardware-Trojan detection methods.

Recently, due to the increase of outsourcing in IC design and manufacturing, it has been reported that malicious third-party IC vendors often insert hardware Trojans into their products. Especially in IC design step, it is strongly required to detect hardware Trojans because malicious third-party vendors can easily insert hardware Trojans in their products. In this paper, we propose a machine-learning-based hardware-Trojan detection method for gate-level netlists using multi-layer neural networks. First, we extract 11 Trojan-net feature values for each net in a netlist. After that, we classify the nets in an unknown netlist into a set of Trojan nets and that of normal nets using multi-layer neural networks. We obtained at most 100% true positive rate with our proposed method.

In this paper, we propose a logic-testing based HT detection and classification method utilizing steady state learning. We first observe that HTs are hidden while applying random test patterns in a short time but most of them can be activated in a very long-term random circuit operation. Hence it is very natural that we learn steady signal-transition states of every suspicious Trojan net in a netlist by performing short-term random simulation. After that, we simulate or emulate the netlist in a very long time by giving random test patterns and obtain a set of signal-transition states. By discovering correlation between them, our method detects HTs and finds out its behavior. Experimental results demonstrate that our method can successfully identify all the real Trojan nets to be Trojan nets and all the normal nets to be normal nets, while other existing logic-testing HT detection methods cannot detect some of them.

In this paper, we propose a floorplan aware high-level synthesis algorithm with body biasing for delay variation compensation, which minimizes the average leakage energy of manufactured chips. In order to realize floorplan-aware high-level synthesis, we utilize huddle-based distributed register architecture (HDR architecture). HDR architecture divides the chip area into small partitions called a huddle and we can control a body bias voltage for every huddle. During high-level synthesis, we iteratively obtain expected leakage energy for every huddle when applying a body bias voltage. A huddle with smaller expected leakage energy contributes to reducing expected leakage energy of the entire circuit more but can increase the latency. We assign control-data flow graph (CDFG) nodes in non-critical paths to the huddles with larger expected leakage energy and those in critical paths to the huddles with smaller expected leakage energy. We expect to minimize the entire leakage energy in a manufactured chip without increasing its latency. Experimental results show that our algorithm reduces the average leakage energy by up to 39.7% without latency and yield degradation compared with typical-case design with body biasing.

Due to the increase of outsourcing by IC vendors, we face a serious risk that malicious third-party vendors insert hardware Trojans very easily into their IC products. However, detecting hardware Trojans is very difficult because today's ICs are huge and complex. In this paper, we propose a hardware-Trojan classification method for gate-level netlists to identify hardware-Trojan infected nets (or Trojan nets) using a support vector machine (SVM) or a neural network (NN). At first, we extract the five hardware-Trojan features from each net in a netlist. These feature values are complicated so that we cannot give the simple and fixed threshold values to them. Hence we secondly represent them to be a five-dimensional vector and learn them by using SVM or NN. Finally, we can successfully classify all the nets in an unknown netlist into Trojan ones and normal ones based on the learned classifiers. We have applied our machine-learning based hardware-Trojan classification method to Trust-HUB benchmarks. The results demonstrate that our method increases the true positive rate compared to the existing state-of-the-art results in most of the cases. In some cases, our method can achieve the true positive rate of 100%, which shows that all the Trojan nets in an unknown netlist are completely detected by our method.

As seen in stream data processing, it is necessary to extract a particular data field from bulk data, where we can use a field-data extractor. Particularly, an (M, N)-field-data extractor reads out any consecutive N bytes from an M-byte register by connecting its input/output using multiplexers (MUXs). However, the number of required MUXs increases too much as the input/output byte widths increase. It is known that partitioning a MUX network leads to reducing the number of MUXs. In this paper, we firstly pick up a multi-layered MUX network, which is generated by repeatedly partitioning a MUX network into a collection of single layered MUX networks. We show that the multi-layered MUX network is equivalent to the barrel shifter from which redundant MUXs and wires are removed, and we prove that the number of required MUXs becomes the smallest among MUX-network-partitioning based field-data extractors. Next, we propose a rotator-based MUX network for a field-data extractor, which is based on reading out a particular data in an input register to a rotator. The byte width of the rotator is the same as its output register and hence we no longer require any extra wires nor MUXs. By rotating the input data appropriately, we can finally have a right-ordered data into an output register. Experimental results show that a multi-layered MUX network reduces the number of required gates to construct a field-data extractor by up to 97.0% compared with the one using a naive approach and its delay becomes 1.8 ns-2.3 ns. A rotator-based MUX network with a control circuit also reduces the number of required gates to construct a field-data extractor by up to 97.3% compared with the one using a naive approach and its delay becomes 2.1 ns-2.9 ns.

Non-volatile memories are paid attention to as a promising alternative to memory design. Data stored in them still may be destructed due to crosstalk and radiation. We can restore the data by using errorcorrecting codes which require extra bits to correct bit errors. Further, nonvolatile memories consume ten to hundred times more energy than normal memories in bit-writing. When we configure them using error-correcting codes, it is quite necessary to reduce writing bits. In this paper, we propose a method to generate a bit-write-reducing code with error-correcting ability. We first pick up an error-correcting code which can correct t-bit errors. We cluster its codeswords and generate a cluster graph satisfying the S-bit flip conditions. We assign a data to be written to each cluster. In other words, we generate one-to-many mapping from each data to the codewords in the cluster. We prove that, if the cluster graph is a complete graph, every data in a memory cell can be re-written into another data by flipping at most S bits keeping error-correcting ability to t bits. We further propose an efficient method to cluster error-correcting codewords. Experimental results show that the bit-write-reducing and error-correcting codes generated by our proposed method efficiently reduce energy consumption. This paper proposes the world-first theoretically near-optimal bit-write-reducing code with error-correcting ability based on the efficient coding theories.

Power analysis for IoT devices is strongly required to protect attacks from malicious attackers. It is also very important to reduce power consumption itself of IoT devices. In this paper, we propose a highly-adaptable and small-sized in-field power analyzer for low-power IoT devices. The proposed power analyzer has the following advantages: (A) The proposed power analyzer realizes signal-averaging noise reduction with synchronization signal lines and thus it can reduce wide frequency range of noises; (B) The proposed power analyzer partitions a long-term power analysis process into several analysis segments and measures voltages and currents of each analysis segment by using small amount of data memories. By combining these analysis segments, we can obtain long-term analysis results; (C) The proposed power analyzer has two amplifiers that amplify current signals adaptively depending on their magnitude. Hence maximum readable current can be increased with keeping minimum readable current small enough. Since all of (A), (B) and (C) do not require complicated mechanisms nor circuits, the proposed power analyzer is implemented on just a 2.5 cm x 3.3 cm board, which is the smallest size among the other existing power analyzers for IoT devices. We have measured power and energy consumption of the AES encryption process on the IoT device and demonstrated that the proposed power analyzer has only up to 1.17% measurement errors compared to a high-precision oscilloscope.

Since digital ICs are often designed and fabricated by third parties at any phases today, we must eliminate risks that malicious attackers may implement Hardware Trojans (HTs) on them. In particular, they can easily insert HTs during design phase. This paper proposes an HT rank which is a new quantitative analysis criterion against HTs at gate-level netlists. We have carefully analyzed all the gate-level netlists in Trust-HUB benchmark suite and found out several Trojan net features in them. Then we design the three types of Trojan points: feature point, count point, and location point. By assigning these points to every net and summing up them, we have the maximum Trojan point in a gate-level netlist. This point gives our HT rank. The HT rank can be calculated just by net features and we do not perform any logic simulation nor random test. When all the gate-level netlists in Trust-HUB, ISCAS85, ISCAS89 and ITC99 benchmark suites as well as several OpenCores designs, HT-free and HT-inserted AES netlists are ranked by our HT rank, we can completely distinguish HT-inserted ones (which HT rank is ten or more) from HT-free ones (which HT rank is nine or less). The HT rank is the world-first quantitative criterion which distinguishes HT-inserted netlists from HT-free ones in all the gate-level netlists in Trust-HUB, ISCAS85, ISCAS89, and ITC99.

Multi-scenario high-level synthesis for distributed register/controller architecture has been proposed targeting static delay variation. In this paper, we extend it and propose a floorplan-driven high-level synthesis algorithm which can be applied to dynamic delay variation by effectively using an error prediction technique, where pre-error registers are introduced to local registers in every circuit block. Experimental results show that the proposed algorithm using two and three scenarios on an FPGA chip reduces the average number of required control steps by 17.6% and 25.5% on average compared to worst-case high-level synthesis at the expense of increasing lookup-tables and flip-flops. Moreover, we implement a multi-scenario elliptic-wave-filter (EWF) circuit with three scenarios synthesized by our proposed algorithm onto an FPGA chip and run it under the environment with varying supply voltages which causes dynamic delay variation. The FPGA implementation experiments also demonstrate that the EWF circuit effectively runs on the real FPGA chip. As far as we know, this is the world-first experiment where a multi-scenario circuit runs under real dynamic delay variation environment.

An (M,N)-field-data extractor reads out any consecutive N bytes from an M-byte register by connecting its input/output using a multiplexer (MUX) network. It is used in packet analysis and/or stream data processing for video/audio data. In this letter, we propose an efficient MUX network for an (M,N)-field-data extractor. By bi-partitioning a simple MUX network into an upper one and a lower one, we can theoretically reduce the number of required MUXs without increasing the MUX network depth. Experimental results show that we can reduce the gate count by up to 92% compared to a naive approach.

Recently, high-level synthesis techniques for FPGA designs (FPGA-HLS techniques) are strongly required in various applications. Both interconnection delays and clock skews have a large impact on circuit performance implemented onto FPGA, which indicates the need for floorplan-driven FPGA-HLS algorithms considering them. To appropriately estimate interconnection delays and clock skews at HLS phase, a reasonable model to estimate them becomes essential. In this paper, we demonstrate several experiments to characterize interconnection delays and clock skews in FPGA and propose novel estimate models called "IDEF" and "CSEF". In order to evaluate our models, we integrate them into a conventional floorplan-driven FPGA-HLS algorithm. Experimental results demonstrate that our algorithm can realize FPGA designs which reduce the latency by up to 22% compared with conventional approaches.

In order to tackle a process-variation problem, we can define several scenarios, each of which corresponds to a particular LSI behavior, such as a typical-case scenario and a worst-case scenario. By designing a single LSI chip which realizes multiple scenarios simultaneously, we can have a process-variation-tolerant LSI chip. In this paper, we propose a multi-scenario high-level synthesis algorithm for variation-tolerant floorplan-driven design targeting new distributed-register architectures, called HDR architectures. We assume two scenarios, a typical-case scenario and a worst-case scenario, and realize them onto a single chip. We first schedule/bind each of the scenarios independently. After that, we commonize the scheduling/binding results for the typical-case and worstcase scenarios and thus generate a commonized area-minimized floorplan result. At that time, we can explicitly take into account interconnection delays by using distributed-register architectures. Experimental results show that our algorithm reduces the latency of the typical-case scenario by up to 50% without increasing the latency of the worst-case scenario, compared with several existing methods.

Indoor areas such as office buildings and railway stations have almost no landmarks and how to navigate pedestrians to their destination points there is one of the big challenges. In this paper, we propose an indoor navigation system based on real-time direction information generation using wearable glasses. The proposed system effectively calculates a pedestrian's direction in a real-time manner using sensors and superimposes the direction to proceed on wearable glasses. Hence it navigates pedestrians to their right direction without using landmarks. Experiments demonstrate the effectiveness of the proposed navigation system.

High-level synthesis for FPGA designs (FPGA-HLS) is recently required in various applications. Since wire delays are becoming a design bottleneck in FPGA, we need to handle interconnection delays and clock skews in FPGA-HLS flow. In this paper, we propose an FPGA-HLS algorithm optimizing critical path with interconnection-delay and clock-skew consideration. By utilizing HDR architecture, we floorplan circuit modules in HLS flow and, based on the result, estimate interconnection delays and clock skews. To reduce the critical-path delay(s) of a circuit, we propose two novel methods for FPGA-HLS. Experimental results demonstrate that our algorithm can improve circuit performance by up to 24% compared with conventional approaches.

As seen in stream data processing, it is necessary to extract a particular data field from bulk data, where we can use a field-data extractor. Particularly, an (M; N)-field-data extractor reads out any consecutive N bytes from an M -byte register by connecting its input/output using multiplexers (MUXs). However, the number of required MUXs increases too much as the input/output byte lengths increase. It is known that partitioning an MUX network leads to reducing the number of MUXs. In this paper, we firstly pick up a multi-layered MUX network, which is generated by repeatedly partitioning a MUX network into a collection of single-layered MUX networks. We show that the multi-layered MUX network is equivalent to the barrel shifter from which redundant MUXs and wires are removed, and we prove that the number of its required MUXs becomes the smallest among MUX-network-partitioning based field-data extractors. Next, we propose a rotator-based MUX network for a field-data extractor, which reads out a particular data in an input register to a rotator. The size of the rotator is the same as its output register and hence we no longer require any extra wires nor MUXs. By rotating the input data appropriately, we can finally have a right-ordered data into an output register. Experimental results show that our rotator-based MUX network reduces the required number of gates to implement a field-data extractor by up to 33% compared with the one using a multi-layered MUX network.

Due to the fact that we do not know who will create hardware Trojans (HTs), and when and where they would be inserted, it is very difficult to correctly and completely detect all the real HTs in untrusted ICs, and thus it is desired to incorporate in-situ HT invalidating functions into untrusted ICs as a countermeasure against HTs. This paper proposes an in situ Trojan authentication technique for gate-level netlists to avoid security leakage. In the proposed approach, an untrusted IC operates in authentication mode and normal mode. In the authentication mode, an embedded Trojan authentication circuit monitors the bit-flipping count of a suspicious Trojan net within the pre-defined constant clock cycles and identify whether it is a real Trojan or not. If the authentication condition is satisfied, the suspicious Trojan net is validated. Otherwise, it is invalidated and HT functions are masked. By doing this, even untrusted netlists with HTs can still be used in the normal mode without security leakage. By setting the appropriate authentication condition using training sets from Trust-HUB gate-level benchmarks, the proposed technique invalidates successfully only HTs in the training sets. Furthermore, by embedding the in-situ Trojan authentication circuit into a Trojan-inserted AES crypto netlist, it can run securely and correctly even if HTs exist where its area overhead is just 1.5% with no delay overhead.

In this paper, we propose a delay variation and floorplan aware high-level synthesis algorithm with body biasing, which minimizes the average leakage energy of manufactured chips. To realize a floorplan-oriented high-level synthesis, we utilize a huddle-based distributed register architecture (HDR architecture), one of the DR architectures. HDR architecture divides the chip area into small partitions called a huddle and we can control a body bias voltage for every huddle. During high-level synthesis, we iteratively obtain expected leakage energy for every huddle when applying a body bias voltage. A huddle with smaller expected leakage energy contributes to reducing expected leakage energy of the entire circuit but can increase the latency. We assign CDFG nodes in critical paths to the huddles with larger expected leakage energy and those in non-critical paths to the huddles with smaller expected leakage energy. We expect to minimize the entire leakage energy in a manufactured chip without increasing its latency. Experimental results show that our algorithm reduces the average leakage energy by up to 38.9% without latency and yield degradation compared with typical-case design with body biasing.

This paper proposes a high-performance circuit design algorithm using input data dependent approximation. In our algorithm, STEPCs (Suspicious Timing Error Prediction Circuits) are utilized for identifying the paths to be optimized inside a circuit efficiently. Experimental results targeting a set of basic adders show that our algorithm can achieve performance increase by up to 11.1% within the error rate of 2.1% compared to a conventional design technique.

Recently, content centric networking (CCN) attracts attention as a next generation network on which every router forwards a packet to another router and also functions as a server. A CCN router has a forwarding table called FIB (Forwarding Information Base) but its table look-up can become a bottleneck. In this paper, we propose FIB data structure for CCN routers which can reduce the number of comparisons in its look-up table. Our proposed FIB is composed of a bloom filter and a hash table and each hash entry is connected to a balanced binary search tree. By using our FIB, the number of comparisons cannot much increase even if hash collisions occur. Experimental results demonstrate the effectiveness of the proposed FIB over the several existing methods.

Power analysis for IoT devices is strongly required to reduce power consumption and realize secure communications. In this paper, we propose a scalable and small-sized power analyzer with signal-averaging noise reduction for low-power IoT devices. The proposed power analyzer reduces a wide frequency range of noises by using a signal averaging method and is implemented on just a 2cm x 3cm board, which is the smallest size among the other existing power analyzers for IoT devices. It further has the following advantages: (a) It has a two-level amplifier that amplifies current signals adaptively depending on their magnitude. Hence maximum readable current can be increased with keeping minimum readable current small enough. (b) If long-time analysis is required, it can be partitioned into several analysis segments. The proposed power analyzer can measure currents and voltages of each analysis segment by using a small amount of data memories. After that, by combining these analysis segments using a timer module, we can obtain long-time analysis results. We have analyzed power and energy consumption of encryption processes of AES block cipher on the IoT device and demonstrated that the proposed power analyzer has only 1.8% measurement error compared with a high-precision oscilloscope.

This paper proposes a redesign technique which designs from untrusted netlists to trusted netlists. Our approach consists of two phases, detection phase and invalidation phase. The detection phase picks up suspicious hardware Trojans (HTs) by pattern matching. The invalidation phase modifies the suspicious HTs in order not to activate them. In the invalidation phase, three invalidation techniques are selected by analyzing location of suspicious malicious nets. Applying appropriately the invalidation technique to the nets can correctly invalidate HTs. In our results, the proposed technique can successfully invalidate HTs on several Trust-HUB benchmarks without HT activations. The results clearly demonstrate that our redesign technique is very effective to remove HT risks.

Recently, we face a serious risk that malicious third-party vendors can very easily insert hardware Trojans into their IC products but it is very difficult to analyze huge and complex ICs. In this paper, we propose a hardware-Trojan classification method to identify hardware-Trojan infected nets (or Trojan nets) using a support vector machine (SVM). Firstly, we extract the five hardware-Trojan features in each net in a netlist. Secondly, since we cannot effectively give the simple and fixed threshold values to them to detect hardware Trojans, we represent them to be a five-dimensional vector and learn them by using SVM. Finally, we can successfully classify a set of all the nets in an unknown netlist into Trojan ones and normal ones based on the learned SVM classifier. We have applied our SVM-based hardware-Trojan classification method to Trust-HUB benchmarks and the results demonstrate that our method can much increase the true positive rate compared to the existing state-of-the-art results in most of the cases. In some cases, our method can achieve the true positive rate of 100%, which shows that all the Trojan nets in a netlist are completely detected by our method.

In this paper, we propose a pedestrian navigation system based on landmark recognition. Our proposed system utilizes a glass-type wearable device and gives a correspondence between a map and a real-world landscape. By recognizing a landmark position effectively, a pedestrian can easily know where to turn at each turning position and hence he/she can reach his/her goal without losing his/her way. Experimental results demonstrate that the proposed system can increase the landmarks which pedestrians can recognize and thus gives comprehensive navigation effectively.

Recently, wristwatch-type wearable devices are developed and geographic information services have become widely available on them. In this paper, we propose a comprehensive deformed map generation algorithm for wristwatch-type wearable devices. Our algorithm first normalizes a pedestrian route to 0 degrees, 45 degrees, or 90 degrees so that the pedestrian can see the route not tilting the wristwatch-type wearable device on his/her wrist. Second, our algorithm partitions the normalized map so that several landmarks are overlapped in the partitioned sub-maps. Hence the sub-maps can be largely displayed on wristwatch-type wearable devices and the pedestrian can recognize his/her location even when the sub-maps displayed are changed. Experiments demonstrate the effectiveness of our deformed map generation algorithm on wristwatch-type wearable devices.

This paper proposes a safe and comprehensive route finding method for pedestrians. We evaluate five factors that do relieve pedestrians' fear of darkness. Based upon the evaluation, we propose a comprehensive route finding method by taking road width and reduction on turning points into consideration. The experimental results on real outdoor environments under different lighting situations confirm that the proposed method can obtain safety and comprehensive routes for pedestrians.

Scan-path test, which is one of design-for-test techniques using a scan chain, can control and observe internal registers in an LSI chip. However, attackers can also use it. to retrieve secret information from cipher circuits. Recently, scan-based attacks using a scan chain inside an LSI chip is reported which can restore secret information by analyzing the scan data during cryptographic processing. In this paper, we pick up a scan-based attack method against a Trivium cipher, one of synchronous stream ciphers, and evaluate it using the FPGA platform called SASEBO-GII We implement the Trivium cipher on the FPGA chip and perform the scan-based attack against it. We demonstrate that the scan-based attack can successfully restore the secret information in the FPGA chip within several minutes, even if the FPGA chip contains several circuits other than the Trivium cipher circuit, which reveals that the scan-based attack against the Trivium cipher is not only a simulation threat but a real threat.

Camellia is a block cipher jointly developed by Mitsubishi and NTT of Japan. It is designed suitable for both software and hardware implementations. One of the design-for-test techniques using scan chains is called scan-path test, in which testers can observe and control the registers inside the LSI chip directly in order to check if the LSI chip correctly operates or not. Recently, a scan-based side-channel attack is reported which retrieves the secret information from the cryptosystem using scan chains. In this paper, we propose a scan-based attack method on the Camellia cipher using scan signatures. Our proposed method is based on the equivalent transformation of the Camellia algorithm and the possible key candidate reduction in order to retrieve the secret key. Experimental results show that our proposed method sucessfully retrieved its 128-bit secret key using 960 plaintexts even if the scan chain includes the Camellia cipher and other circuits and also sucessfully retrieves its secret key on the SASEBO-GII board, which is a side-channel attack standard evaluation board.

Outsourcing IC design and fabrication is one of the effective solutions to reduce design cost but it may cause severe security risks. Particularly, malicious outside vendors may implement Hardware Trojans (HTs) on ICs. When we focus on IC design phase, we cannot assume an HT-free netlist or a Golden netlist and it is too difficult to identify whether a given netlist is HT-free or not. In this paper, we propose a score-based hardware-trojans identifying method at gate-level netlists without using a Golden netlist. Our proposed method does not directly detect HTs themselves in a gate-level netlist but it detects a net included in HTs, which is called Trojan net, instead. Firstly, we observe Trojan nets from several HT-inserted benchmarks and extract several their features. Secondly, we give scores to extracted Trojan net features and sum up them for each net in benchmarks. Then we can find out a score threshold to classify HT-free and HT-inserted netlists. Based on these scores, we can successfully classify HT-free and HT-inserted netlists in all the Trust-HUB gate-level benchmarks and ISCAS85 benchmarks as well as HT-free and HT-inserted AES gate-level netlists. Experimental results demonstrate that our method successfully identify all the HT-inserted gate-level benchmarks to be "HT-inserted" and all the HT-free gate-level benchmarks to be "HT-free" in approximately three hours for each benchmark.

Non-volatile memory has many advantages such as high density and low leakage power but it consumes larger writing energy than SRAM. It is quite necessary to reduce writing energy in non-volatile memory design. In this paper, we propose write-reduction codes based on error correcting codes and reduce writing energy in non-volatile memory by decreasing the number of writing bits. When a data is written into a memory cell, we do not write it directly but encode it into a codeword. In our write-reduction codes, every data corresponds to an information vector in an error-correcting code and an information vector corresponds not to a single codeword but a set of write-reduction codewords. Given a writing data and current memory bits, we can deterministically select a particular write-reduction codeword corresponding to the data to be written, where the maximum number of flipped bits are theoretically minimized. Then the number of writing bits into memory cells will also be minimized. Experimental results demonstrate that we have achieved writing-bits reduction by an average of 51% and energy reduction by an average of 33% compared to non-encoded memory.

Data stored in non-volatile memories may be destructed due to crosstalk and radiation but we can restore their data by using error-correcting codes. However, non-volatile memories consume a large amount of energy in writing. How to reduce maximum writing bits even using error-correcting codes is one of the challenges in non-volatile memory design. In this paper, we first propose Doughnut code which is based on state encoding limiting maximum and minimum Hamming distances. After that, we propose a code expansion method, which improves maximum and minimum Hamming distances. When we apply our code expansion method to Doughnut code, we can obtain a code which reduces maximum-flipped bits and has error-correcting ability equal to Hamming code. Experimental results show that the proposed code efficiently reduces the number of maximum-writing bits.

As process technologies advance, interconnection delays are not negligible even in high-level synthesis and regular-distributed-register (RDR) architecture has been proposed to cope with this problem. In this paper, we propose a floorplan-driven high-level synthesis algorithm using multiple-operation chainings composed of two or more operations, and reduce the overall latency targeting RDR architecture. Our algorithm enumerates multiple-operation-chaining path candidates before performing scheduling/binding. Based on them, we find out optimal ones taking into account RDR floorplan information. Experimental results show that our algorithm successfully reduces the latency by up to 30.4% compared to the conventional approaches.

As process technologies advance, timing-error correction techniques have become important as well. A suspicious timing-error prediction (STEP) technique has been proposed recently, which predicts timing errors by monitoring themiddle points, or check points of several speed-paths in a circuit. However, if we insert STEP circuits (STEPCs) in the middle points of all the paths from primary inputs to primary outputs, we need many STEPCs and thus require too much area overhead. How to determine these check points is very important. In this paper, we propose an effective STEPC insertion algorithm minimizing area overhead. Our proposed algorithm moves the STEPC insertion positions to minimize inserted STEPC counts. We apply a max-flow and min-cut approach to determine the optimal positions of inserted STEPCs and reduce the required number of STEPCs to 1/10-1/80 and their area to 1/5-1/8 compared with a naive algorithm. Furthermore, our algorithm realizes 1.12X-1.5X overclocking compared with just inserting STEPCs into several speed-paths.

Recently, high-level synthesis (HLS) techniques for FPGA designs are required in various applications such as computerized stock tradings and reconfigurable network processings. In HLS for FPGA designs, we need to consider module floorplan and reduce multiplexer's cost concurrently. In this paper, we propose a floorplan-driven HLS algorithm for multiplexer reduction targeting FPGA designs. By utilizing distributed-register architectures called HDR, we can easily consider module floorplan in HLS. In order to reduce multiplexer's cost, we propose two novel binding methods called datapath-oriented scheduling/FU binding and datapath-oriented register binding. Experimental results demonstrate that our algorithm can realize FPGA designs which reduce the number of slices by up to 47% and latency by up to 22% compared with conventional approaches while the number of required control steps is almost the same.

In this paper, we first propose an HDR-mcd architecture, which integrates periodically all-in-phase based multiple clock domains and multi-cycle interconnect communication into high-level synthesis. In HDR-mcd, an entire chip is divided into several huddles. Huddles can realize synchronization between different clock domains in which interconnection delay should be considered during high-level synthesis. Next, we propose a high-level synthesis algorithm for HDR-mcd, which can reduce energy consumption by optimizing configuration and placement of huddles. Experimental results show that the proposed method achieves 32.5% energy-saving compared with the existing single clock domain based methods.

In deep-submicron era, interconnection delays are not negligible even in high-level synthesis and regular-distributed-register architectures (RDR architectures) have been proposed to cope with this problem. In this paper, we propose a high-level synthesis algorithm using operation chainings which reduces the overall latency targeting RDR architectures. Our algorithm consists of three steps: The first step enumerates candidate operations for chaining. The second step introduces maximal chaining distance (MCD), which gives the maximal allowable inter-island distance on RDR architecture between chaining candidate operations. The last step performs list-scheduling and binding simultaneously based on the results of the two preceding steps. Our algorithm enumerates feasible chaining candidates and selects the best ones for RDR architecture. Experimental results show that our proposed algorithm reduces the latency by up to 40.0% compared to the original approach, and by up to 25.0% compared to a conventional approach. Our algorithm also reduces the number of registers and the number of multiplexers compared to the conventional approaches in some cases.

With the miniaturization and high performance of current and future LSIs, demand for portable devices has much more increased. Especially the problems of battery runtime and device overheating have occurred. In addition, with the downsize of the LSI design process, the ratio of an interconnection delay to a gate delay has continued to increase. High-level synthesis to estimate the interconnection delays and reduce energy consumption is essential. In this paper, we propose a high-level synthesis algorithm based on HDR architectures (huddle-based distributed register architectures) utilizing multi-stage clock gating. By increasing the number of clock gating stages in each huddle, we increase the number of the control steps at which we can apply the clock gating to registers. We can determine the configuration of the clock gating with optimized energy consumption. The experimental results demonstrate that our proposed algorithm reduced energy consumption by up to 27.7% compared with conventional algorithms.

Lithography is a technology to make circuit patterns on a wafer. UV light diffracted by a photomask forms optical images on a photoresist. Then, a photoresist is melt by an amount of exposed UV light exceeding the threshold. The UV light diffracted by a photomask through lens exposes the photoresist on the wafer. Its lightness and darkness generate patterns on the photoresist. As the technology node advances, the feature sizes on photoresist becomes much smaller. Diffracted UV light is dispersed on the wafer, and then exposing photoresists has become more difficult. Exposure source optimization, SO in short, techniques for optimizing illumination shape have been studied. Although exposure source has hundreds of grid-points, all of previous works deal with them one by one. Then they consume too much running time and that increases design time extremely. How to reduce the parameters to be optimized in SO is the key to decrease source optimization time. In this paper, we propose a variation-resilient and high-speed cluster-based exposure source optimization algorithm. We focus on image log slope (ILS) and use it for generating clusters. When an optical image formed by a source shape has a small ILS value at an EPE (Edge placement error) evaluation point, dose/focus variation much affects the EPE values. When an optical image formed by a source shape has a large ILS value at an evaluation point, dose/focus variation less affects the EPE value. In our algorithm, we cluster several grid-points with similar ILS values and reduce the number of parameters to be simultaneously optimized in SO. Our clustering algorithm is composed of two STEPs: In STEP 1, we cluster grid-points into four groups based on ILS values of grid-points at each evaluation point. In STEP 2, we generate super clusters from the clusters generated in STEP 1. We consider a set of grid-points in each cluster to be a single light source element. As a result, we can optimize the SO problem very fast. Experimental results demonstrate that our algorithm runs speed-up compared to a conventional algorithm with keeping the EPE values.

For realizing better trade-off between performance and yield rate in recent LSI designs, it is required to deal with increasing the ratios of interconnect delay as well as delay variation. In this paper, a novel floorplan-aware high-level synthesis technique with delay-variation tolerance is proposed. By utilizing floorplan-driven architectures, interconnect delays can be estimated and then handled even in high-level synthesis. Applying our technique enables to realize two scheduling/binding results (one is a non-delayed result and the other is a delayed result) simultaneously on a chip with small area/performance overhead, and either one of them can be selected according to the post-silicon delay variation. Experimental results demonstrate that our technique can reduce delayed scheduling/binding latency by up to 32.3% compared with conventional approaches.

There are a number of studies on a side-channel attack which uses information exploited from the physical implementation of a cryptosystem. A scan-based side-channel attack utilizes scan chains, one of design-for-test techniques and retrieves the secret information inside the cryptosystem. In this paper, scan based side-channel attack methods against symmetric key ciphers such as block ciphers and stream ciphers using scan signatures are presented to show the risk of scan-based attacks.

As seen in packet analysis of TCP/IP offload engine and stream data processing for video/audio data, it is necessary to extract a particular data field from bulk data, where we can use a field-data extractor. Particularly, an (M, N)-field-data extractor reads out any consecutive N bytes from an M-byte register by connecting its input/ output using multiplexers. However, the number of required multiplexers increases too much as the input/ output byte lengths increase. How to reduce the number of its required multiplexers is a major challenge. In this paper, we propose an efficient multiplexer network synthesis method for an (M, N)-field-data extractor. Our method is based on inserting an (N + B - 1)-byte virtual intermediate register into a multiplexer network and partitioning it into an upper network and a lower network. Our method theoretically reduces the number of required multiplexers without increasing the multiplexer network depth. We also propose how to determine the size of the virtual intermediate register that minimizes the number of required multiplexers. Experimental results show that our method reduces the required number of gates to implement a field-data extractor by up to 92% compared with the one using a naive multiplexer network.

In order to tackle a process-variation problem, we can define several scenarios, each of which corresponds to a particular LSI behavior, such as a typical-case scenario and a worst-case scenario. By designing a single LSI chip which realizes multiple scenarios simultaneously, we can have a process-variation-tolerant LSI chip. In this paper, we propose a processvariation- aware low-latency and multi-scenario high-level synthesis algorithm targeting new distributed-register architectures, called HDR architectures. We assume two scenarios, a typical-case scenario and a worst-case scenario, and realize them onto a single chip. We first schedule/bind each of the scenarios independently. After that, we commonize the scheduling/binding results for the typical-case and worst-case scenarios and thus generate a commonized area-minimized floorplan result. Experimental results show that our algorithm reduces the latency of the typical-case scenario by up to 50% without increasing the latency of the worst-case scenario, compared with several existing methods.

As process technologies advance, interconnection delays are not negligible even in high-level synthesis and regular-distributed-register (RDR) architecture has been proposed to cope with this problem. In this paper, we propose a floorplan-driven high-level synthesis algorithm using multiple-operation chainings composed of two or more operations, and reduce the overall latency targeting RDR architecture. Our algorithm enumerates multiple-operation-chaining path candidates before performing scheduling/ binding. Based on them, we find out optimal ones taking into account RDR floorplan information. Experimental results show that our algorithm successfully reduces the latency by up to 30.4% compared to the conventional approaches.

Non-volatile memories are paid attention to as a promising alternative to memory design. Data stored in them still may be destructed due to crosstalk and radiation. We can restore the data by using error-correcting codes which require extra bits to correct bit errors. Further, non-volatile memories consume ten to hundred times more energy than normal memories in bit-writing. When we configure them using error-correcting codes, it is quite necessary to reduce writing bits. In this paper, we propose a method to generate a bit-write-reducing code with error-correcting ability. We first pick up an error-correcting code which can correct t-bit errors. We cluster its codeswords and generate a cluster graph satisfying the S-bit flip conditions. We assign a data to be written to each cluster. In other words, we generate one-to-many mapping from each data to the codewords in the cluster. We prove that, if the cluster graph is a complete graph, every data in a memory cell can be re-written into another data by flipping at most S bits keeping error-correcting ability to t bits. We further propose an efficient method to cluster error-correcting codewords. Experimental results demonstrate that, when we apply our bit-write-reducing code to MediaBench applications, it can reduce writing-bit counts by up to 28.2% and also energy consumption of non-volatile memory cells by up to 27.9% compared to existing error-correcting codes keeping the same error-correcting ability. This paper proposes the world-first theoretically near-optimal bit-write-reducing code with error-correcting ability based on the efficient coding theories.

The recent development of network technologies that offer centralized control of explicit routes opens the door to the online optimization of explicit routing. For this kind of Traffic Engineering optimization, raising the calculation speeds by using multi-core processors with effective parallel algorithms is a key goal. This paper proposes an effective parallel algorithm for General purpose Programming on Graphic Processing Unit (GPGPU); its massively parallel style promises strong acceleration of calculation speed. The proposed algorithm parallelizes not only the search method of the Genetic Algorithm, but also its fitness functions, which calculate the network congestion ratio, so as to fully utilize the power of modern GPGPUs. Concurrently, each execution is designed for thread-block execution on the GPU with consideration of thread occupancy, local resources, and SIMT execution to maximize GPU performance. Evaluations show that the proposed algorithm offers, on average, a nine fold speedup compared to the conventional CPU approach.

This paper proposes a landmark-based route recommendation method for enjoyable walking atmosphere strategies by accumulating and analyzing geographical information. We utilize landmark categorization and region clustering to obtain effective elements. Experimental results demonstrate that our proposed method improves walking environment quality and confirm that it is applicable in both urban and rural areas.

Although many GPS-based pedestrian navigations are released, their instructions at decision points are not sufficient. This is mainly due to the lack of landmark informations. They may cause pedestrians to pass decision points or misunderstand when to turn. This paper proposes a visible corner-landmark based route finding algorithm. The proposed algorithm is based on visibility edges for landmarks and can obtain a pedestrian route that has visible landmarks on its corner points. Experiments demonstrate that the proposed algorithm can maximize the visible corner landmarks included in the generated routes.

Recently, digital ICs are often designed by outside vendors to reduce design costs in semiconductor industry, which may introduce severe risks that malicious attackers implement Hardware Trojans (HTs) on them. Since IC design phase generates only a single design result, an RT-level or gate-level netlist for example, we cannot assume an HT-free netlist or a Golden netlist and then it is too difficult to identify whether a generated netlist is HT-free or HT-inserted. In this paper, we propose a score-based classification method for identifying HT-free or HT-inserted gate-level netlists without using a Golden netlist. Our proposed method does not directly detect HTs themselves in a gate-level netlist but a net included in HTs, which is called Trojan net, instead. Firstly, we observe Trojan nets from several HT-inserted benchmarks and extract several their features. Secondly, we give scores to extracted Trojan net features and sum up them for each net in benchmarks. Then we can find out a score threshold to classify HT-free and HT-inserted netlists. Based on these scores, we can successfully classify HT-free and HT-inserted netlists in all the Trust-HUB gate-level benchmarks. Experimental results demonstrate that our method successfully identify all the HT-inserted gate-level benchmarks to be "HT-inserted" and all the HT-free gate-level benchmarks to be "HT-free" in approximately three hours for each benchmark.

Non-volatile memory has many advantages over SRAM. However, one of its largest problems is that it consumes a large amount of energy in writing. In this paper, we propose a bit-write reduction method based on error correcting codes for non-volatile memories. When a data is written into a memory cell, we do not write it directly but encode it into a codeword. We focus on error-correcting codes and generate new codes called write-reduction codes. In our write-reduction codes, each data corresponds to an information vector in an error-correcting code and an information vector corresponds not to a single codeword but a set of write-reduction codewords. Given a writing data and current memory bits, we can deterministically select a particular write-reduction codeword corresponding to a data to be written, where the maximum number of flipped bits are theoretically minimized. Then the number of writing bits into memory cells will also be minimized. We perform several experimental evaluations and demonstrate up to 72% energy reduction.

As process technology is scaling down, timing speculation techniques such as Razor and STEP are emerged as alternative solutions to reduce required margins due to various variation effects. Unlike Razor, STEP is a prediction-based timing speculation method to predict suspicious timing errors before they really appear, and thus it can result in more performance improvement. Therefore, an improved monitoring-path selection algorithm for STEP-based timing speculation is proposed in this paper, in which candidate monitoring-paths are selected based on short path removement and path length estimation. Experimental results show that the proposed algorithm realizes an average of 1.71X overclocking compared with worst-case based designs.

As process technology continues scaling, low power and reliability of integrated circuits are becoming more critical than ever before. Particularly, due to the reduction of node capacitance and operating voltage for low power consumption, it makes the circuits more sensitive to high-energy particles induced soft errors. In this paper, a soft-error tolerant latch called TSPC-SEH is proposed for soft error tolerance with low power consumption. The simulation results show that the proposed TSPC-SEH latch can achieve up to 42% power consumption reduction and 54% delay improvement compared to the existing soft error tolerant SEH and DICE designs.

Power analysis for IoT devices is strongly required to reduce power consumption and realize secure communications, where we need a small-sized power analyzer that can reduce a wide frequency range of noises is needed. In this paper, we propose a small-sized and noise-reduced power analyzer for IoT devices. We utilize a signal averaging method to reduce a wide frequency range of noises. At that time, how to implement a synchronous process between a power analyzer and a target IoT device becomes the key problem. We solve this problem by ( a) using synchronization signals generated by a general-purpose I/O interface of a microprocessor and ( b) introducing a data-order correction process. We analyze power/energy consumption of the encryption process of LED block cipher on the IoT device and obtain an average power of 146.3mW and energy of 3.84mJ. The proposed power analyzer is just implemented on a 5cmx5cm board but these results only have 5% errors compared to a highprecision oscilloscope.

Alpha blending is one of image synthesis techniques, which synthesizes a new image by summing up weighted input images and realizes transparent effect. In this paper, we focus on alpha blending using selector logics and implement it on an FPGA board. By applying selector logics to the alpha blending operation, its total product terms are decreased and thus a circuit size and circuit delay are improved simultaneously. In our implementation, original pixel values are stored into a memory on the FPGA board and then a new pixel value is synthesized based on input transmittance factors. We realize approximately 23% speed-up and 8% area reduction simultaneously using selector-logic based alpha blending.

Recently, high-level synthesis (HLS) techniques for FPGA designs are required in various applications. Clock network in FPGA has already been built before implementing any circuits, which may lead a large impact of clock skews and then degrade operation frequency. In this paper, we formulate a clock skew estimate model for FPGA-HLS (CSEF). CSEF is an accurate model to estimate clock skews in HLS flow. CSEF is then integrated into a floorplan-aware HLS algorithm targeting FPGA designs. Experimental results demonstrate that our HLS algorithm can realize FPGA designs which reduce the latency by up to 19% compared with conventional approaches.

LED (Light Encryption Device) block cipher, one of lightweight block ciphers, is very compact in hardware. Its encryption process is composed of AES-like rounds. Recently, a scan-based side-channel attack is reported which retrieves the secret information inside the cryptosystem utilizing scan chains, one of design-for-test techniques. In this paper, a scan-based attack method on the LED block cipher using scan signatures is proposed. In our proposed method, we focus on a particular 16-bit position in scanned data obtained from an LED LSI chip and retrieve its secret key using scan signatures. Experimental results show that our proposed method successfully retrieves its 64-bit secret key using 36 plaintexts on average if the scan chain is only connected to the LED block cipher. These experimental results also show the key is successfully retrieved even if the scan chain includes additional 130,000 1-bit data.

Trivium is a synchronous stream cipher using three shift registers. It is designed to have a simple structure and runs at high speed. A scan-based side-channel attack retrieves secret information using scan chains, one of design-for-test techniques. In this paper, a scan-based side-channel attack method against Trivium using scan signatures is proposed. In our method, we reconstruct a previous internal state in Trivium one by one from the internal state just when a ciphertext is generated. When we retrieve the internal state, we focus on a particular 1-bit position in a collection of scan chains and then we can attack Trivium even if the scan chain includes other registers than internal state registers in Trivium. Experimental results show that our proposed method successfully retrieves a plaintext from a ciphertext generated by Trivium.

As device feature size drops, interconnection delays often exceed gate delays. We have to incorporate interconnection delays even in high-level synthesis. Using RDR architectures is one of the effective solutions to this problem. At the same time, process and delay variation also becomes a serious problem which may result in several timing errors. How to deal with this problem is another key issue in high-level synthesis. In this paper, we propose a delay-variation-aware high-level synthesis algorithm for RDR architectures. We first obtain a non-delayed scheduling/binding result and, based on it, we also obtain a delayed scheduling/binding result. By adding several extra functional units to vacant RDR islands, we can have a delayed scheduling/binding result so that its latency is not much increased compared with the non-delayed one. After that, we similarize the two scheduling/binding results by repeatedly modifying their results. We can finally realize non-delayed and delayed scheduling/binding results simultaneously on RDR architecture with almost no area/performance overheads and we can select either one of them depending on post-silicon delay variation. Experimental results show that our algorithm successfully reduces delayed scheduling/binding latency by up to 42.9% compared with the conventional approach.

LED (Light Encryption Device) block cipher, one of lightweight block ciphers, is very compact in hardware. Its encryption process is composed of AES-like rounds. Recently, a scan-based side-channel attack is reported which retrieves the secret information inside the cryptosystem utilizing scan chains, one of design-for-test techniques. In this paper, a scan-based attack method on the LED block cipher using scan signatures is proposed. In our proposed method, we focus on a particular 16-bit position in scanned data obtained from an LED LSI chip and retrieve its secret key using scan signatures. Experimental results show that our proposed method successfully retrieves its 64-bit secret key using 73 plaintexts on average if the scan chain is only connected to the LED block cipher. These experimental results also show the key is successfully retrieved even if the scan chain includes additional some 4000 1-bit registers. © 2014 IEEE.

Interpolation is a technique that presumes a value between existing data, which is often used for image scaling and correction of distortion. Linear interpolation is one of the interpolation techniques which interpolates inbetween values by linearly connecting two known values. Also, bi-linear interpolation is one of interpolation techniques, which interpolates a value linearly from its four circumferences. Both of them are used practically in many cases. In this paper, we propose high-speed and small-sized linear and bi-linear interpolation circuits based on selector logics. The proposed linear and bi-linear interpolation circuits reduce carry propagation delays by using selector logics and then realize fast and small-sized circuits. We have implemented our linear interpolation circuit and bi-linear interpolation circuits in several ways and evaluated each of them. We can find out that a selector-based bi-linear interpolation circuit where its partial products are summed up by using the arithmetic operator saves its area by up to 42% and reduces its delay by up to 18% compared with a conventional design. © 2014 IEEE.

In this paper, a throughput-driven design technique is proposed, in which a suspicious timing error prediction circuit is inserted to monitor the signal transitions at some selected check points. Unlike previous works where timing errors are detected after their occurrence, the proposed method tries to use the real intermediate signal transitions for timing error prediction. The check point selection will affect both the maximal operation frequency and the suspicious timing error overestimation rate, both of which have an effect on the overall throughput, thus an analysis on the check point selection is also given. In our work, the circuit can be overclocked by a factor of 2 or more with ignorable area overhead while guarantees the always-correct output.

LED (Light Encryption Device) block cipher, one of lightweight block ciphers, is very compact in hardware. Its encryption process is composed of AES-like rounds. Recently, a scan-based side-channel attack is reported which retrieves the secret information inside the cryptosystem utilizing scan chains, one of design-for-test techniques. In this paper, a scan-based attack method on the LED block cipher using scan signatures is proposed. In our proposed method, we focus on a particular 16-bit position in scanned data obtained from an LED LSI chip and retrieve its secret key using scan signatures. Experimental results show that our proposed method successfully retrieves its 64-bit secret key using 73 plaintexts on average if the scan chain is only connected to the LED block cipher. These experimental results also show the key is successfully retrieved even if the scan chain includes additional some 4000 1-bit registers.

Interpolation is a technique that presumes a value between existing data, which is often used for image scaling and correction of distortion. Linear interpolation is one of the interpolation techniques which interpolates inbetween values by linearly connecting two known values. Also, bi-linear interpolation is one of interpolation techniques, which interpolates a value linearly from its four circumferences. Both of them are used practically in many cases. In this paper, we propose high-speed and small-sized linear and bi-linear interpolation circuits based on selector logics. The proposed linear and bi-linear interpolation circuits reduce carry propagation delays by using selector logics and then realize fast and small-sized circuits. We have implemented our linear interpolation circuit and bi-linear interpolation circuits in several ways and evaluated each of them. We can find out that a selector-based bi-linear interpolation circuit where its partial products are summed up by using the arithmetic operator saves its area by up to 42% and reduces its delay by up to 18% compared with a conventional design.

With technology scaling, process, voltage, and temperature (PVT) variations pose great challenges on integrated circuit designs. Conventionally, LSI circuits are designed by adding pessimistic timing margin to guarantee "always correct" operations even under worst-case conditions. However, due to the increasing PVT variations, unacceptable larger design guard band should be reserved to avoid timing errors on critical paths of circuits, which will therefore lead to very inefficient designs in terms of power and performance. For this reason, in-situ timing monitoring technique has gained great research interest. In this paper, we will review existing variation-resilient design techniques with particular emphasis on in-situ timing monitoring techniques including both detection and prediction-based methods. The effectiveness of in-situ timing monitoring techniques will be discussed. Finally, we show an example of in-situ timing monitoring technique called STEP with applications to general pipeline designs.

Scan test using scan chains is one of the most important DFT techniques. However, scan-based attacks are reported which can retrieve the secret key in crypto circuits by using scan chains. Secure scan architecture is strongly required to protect scan chains from scan-based attacks. This paper proposes an improved version of random order as a secure scan architecture. In improved random order, a scan chain is partitioned into multiple sub-chains. The structure of the scan chain changes dynamically by selecting a subchain to scan out. Testability and security of the proposed improved random order are also discussed in the paper, and the implementation results demonstrate the effectiveness of the proposed method.

Data stored in non-volatile memories may be destructed due to crosstalk and radiation but we can restore their data by using error-correcting codes. However, non-volatile memories consume a large amount of energy in writing. How to reduce writing bits even using error-correcting codes is one of the challenges in non-volatile memory design. In this paper, we propose a new write-reducing and error-correcting code, called Doughnut code. Doughnut code is based on state encoding limiting maximum and minimum Hamming distances. After that, we propose a code expansion method, which improves minimum and maximum Hamming distances by expanding a write-reducing code. When we apply our code expansion method to Doughnut code, we can obtain a write-reducing code whose error-correcting ability is equal to Hamming code. Experimental results show that the proposed write-reducing code reduces the number of writing bits by up to 36% compared to Hamming code.

As process technologies advance, the importance of timing error correction techniques is increasing as well. In this paper, We propose an area-overhead-oriented monitoring-path selection algorithm for suspicious timing error prediction circuits (STEPCs). STEPC predicts timing errors by monitoring the middle points of several speed-paths in a circuit. However, we need many STEPCs with a high area overhead to predict timing errors in an overall circuit. Our proposed method moves the STEPC insertion positions to minimize the number of inserted STEPCs. We apply a max-flow and min-cut approach to determine the optimal positions of inserted STEPCs. Our proposed algorithm reduces the required number of STEPCs to 1/19 and their area to 1/5 compared with a naive algorithm. Furthermore, our algorithm realizes 2.25X overclocking compared with just inserting STEPCs into several speed-paths.

Camellia, a block cipher jointly developed by Mitsubishi and NTT of Japan, is suitable for both software and hardware implementations and more secure than AES cipher. One of design-for-test techniques using scan chains is called scan-path test, in which testers can observe and control registers inside the LSI chip directly. Recently, scan-based side-channel attack is reported which retrieves the secret information from the cryptosystem using scan chains. In this paper, we propose a scan-based attack method on Camellia cipher using scan signatures. Our proposed method is based on equivalent transformation of the Camellia algorithm and key pattern reduction in order to retrieve the secret key. Experimental results show that our proposed method sucessfully retrieves its 128-bit secret key using 960 plaintexts if the scan chain is only connected to the Camellia cipher and also sucessfully retrieves its key on SASEBO-GII, which is a side-channel attack standard evaluation board.

In deep-submicron era, interconnection delays are not negligible even in high-level synthesis and RDR (Regular-Distributed-Register) architecture has been proposed to cope with this problem. In this paper, we propose a high-level synthesis algorithm using operation chainings which reduces the overall latency targeting RDR architectures. Our algorithm consists of three steps: The first step enumerates candidates for chaining. The second step introduces maximal chaining distance (MCD), which gives the maximum allowable distance on RDR architecture between chaining candidate operations. The last step performs list-scheduling and binding simultaneously using the results of two preceding steps. Our algorithm enumerates feasible chaining candidates and selects the best ones for RDR architecture. Experimental results show that our algorithm reduces the latency by up to 28.6%, the number of registers by up to 37.5%, the number of multiplexers by up to 25.0%, compared to the conventional approaches.

Recently, high-level synthesis (HLS) techniques for FPGA designs are required in various applications such as computerized stock tradings and reconfigurable network processings. In HLS for FPGA designs, we need to consider module floorplan and reduce multiplexer's cost concurrently. In this paper, we propose a floorplan-aware HLS algorithm for multiplexer reduction targeting FPGA designs. By utilizing distirbuted-register architectures called HDR, we can easily consider module floorplan in HLS. In order to reduce multiplexer's cost, we propose two novel binding methods called datapath-oriented scheduling/FU binding and datapath-oriented register binding. Experimental results demonstrate that our algorithm can realize FPGA designs which reduces the number of slices by up to 47% and circuit delay by up to 16% compared with the conventional approach.

As device feature size drops, interconnection delays often exceed gate delays. We have to incorporate interconnection delays even in high-level synthesis. Using RDR architectures is one of the effective solutions to this problem. At the same time, process and delay variation also becomes a serious problem which may result in several timing errors. How to deal with this problem is another key issue in high-level synthesis. In this paper, we propose a delay-variation-aware high-level synthesis algorithm for RDR architectures. We first obtain a non-delayed scheduling/binding result and, based on it, we also obtain a delayed scheduling/binding result. By adding several extra functional units to vacant RDR islands, we can have a delayed scheduling/binding result so that its latency is not much increased compared with the non-delayed one. After that, we similarize the two scheduling/binding results by repeatedly modifying their results. We can finally realize non-delayed and delayed scheduling/binding results simultaneously on RDR architecture with almost no area/performance overheads and we can select either one of them depending on post-silicon delay variation. Experimental results show that our algorithm successfully reduces delayed scheduling/binding latency by up to 42.9% compared with the conventional approach.

With the miniaturization and high performance of current and future LSIs, demand for portable devices has much more increased. Especially the problems of battery runtime and device overheating have occurred. In addition, with the downsize of the LSI design process, the ratio of an interconnection delay to a gate delay has continued to increase. High-level synthesis to estimate the interconnection delays and reduce energy consumption is essential. In this paper, we propose a high-level synthesis algorithm based on HDR architectures (huddle-based distributed register architectures) utilizing multi-stage clock gating. By increasing the number of clock gating stages in each huddle, we increase the number of the control steps at which we can apply the clock gating to registers. We can determine the configuration of the clock gating with optimized energy consumption. The experimental results demonstrate that our proposed algorithm reduced energy consumption by up to 27.7% compared with conventional algorithms.

In this paper, we propose an adaptive voltage huddle-based distributed-register architecture (AVHDR architecture), which integrates dynamic multiple supply voltages and interconnection delay into high-level synthesis. In AVHDR architecture, voltages can be dynamically assigned for energy reduction. In other words, low supply voltages are assigned to non-critical operations, and leakage power is cut off by turning off the power supply to the sleeping functional units. Next, an AVHDR-based high-level synthesis algorithm is proposed. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, the modules in each huddle can be placed close to each other and the corresponding AVHDR architecture can be generated and optimized with floorplanning information. Experimental results show that on average our algorithm achieves 43.9% energy-saving compared with conventional algorithms.

Recently, multi-core processors are used in embedded systems very often. Since application programs is much limited running on embedded systems, there must exists an optimal cache memory configuration in terms of power and area. Simulating application programs on various cache configurations is one of the best options to determine the optimal one. Multi-core cache configuration simulation, however, is much more complicated and takes much more time than single-core cache configuration simulation. In this paper, we propose a very fast dual-core L1 cache configuration simulation algorithm. We first propose a new data structure where just a single data structure represents two or more multi-core cache configurations with different cache associativities. After that, we propose a new multi-core cache configuration simulation algorithm using our new data structure associated with new theorems. Experimental results demonstrate that our algorithm obtains exact simulation results but runs 20 times faster than a conventional approach.

A scan-path test is one of the useful design-for-test techniques, in which testers can observe and control registers inside the target LSI chip directly. On the other hand, the risk of side-channel attacks against cryptographic LSIs and modules has been pointed out. In particular, scan-based attacks which retrieve secret keys by analyzing scan data obtained from scan chains have been attracting attention. In this paper, we propose two scan-based attack methods against DES and Triple DES using scan signatures. Our proposed methods are based on focusing on particular bit-column-data in a set of scan data and observing their changes when giving several plaintexts. Based on this property, we introduce the idea of a scan signature first and apply it to DES cryptosystems. In DES cryptosystems, we can retrieve secret keys by partitioning the S-BOX process into eight independent sub-processes and reducing the number of the round key candidates from 2⁴⁸ to 2⁶ × 8 = 512. In Triple DES cryptosystems, three secret keys are used to encrypt plaintexts. Then we retrieve them one by one, using the similar technique as in DES cryptosystems. Although some problems occur when retrieving the second/third secret key, our proposed method effectively resolves them. Our proposed methods can retrieve secret keys even if a scan chain includes registers except a crypto module and attackers do not know when the encryption is really done in the crypto module. Experimental results demonstrate that we successfully retrieve the secret keys of a DES cryptosystem using at most 32 plaintexts and that of a Triple DES cryptosystem using at most 36 plaintexts.

With the miniaturization of LSIs and its increasing performance, demand for high-functional portable devices has grown significantly. At the same time, battery lifetime and device overheating are leading to major design problems hampering further LSI integration. On the other hand, the ratio of an interconnection delay to a gate delay has continued to increase as device feature size decreases. We have to estimate interconnection delays and reduce energy consumption even in a high-level synthesis stage. In this paper, we propose a high-level synthesis algorithm for huddle-based distributed-register architectures (HDR architectures) with clock gatings based on concurrency-oriented scheduling/functional unit binding. We assume coarse-grained clock gatings to huddles and we focus on the number of control steps, or gating steps, at which we can apply the clock gating to registers in every huddle. We propose two methods to increase gating steps: One is that we try to schedule and bind operations to be performed at the same timing. By adjusting the clock gating timings in a high-level synthesis stage, we expect that we can enhance the effect of clock gatings more than applying clock gatings after logic synthesis. The other is that we try to synthesize huddles such that each of the synthesized huddles includes registers which have similar or the same clock gating timings. At this time, we determine the clock gating timings to minimize all energy consumption including clock tree energy. The experimental results show that our proposed algorithm reduces energy consumption by a maximum of 23.8% compared with several conventional algorithms.

In this paper, a concurrent faulty clock detection method is proposed for crypto circuits against clock glitch based differential fault analysis (DFA). In the proposed method, a nonlogic buffer-based delay chain is inserted, and then by monitoring the delay along the delay chain, a possible clock glitch based DFA can be detected. Experimental results on an AES circuit show that the proposed method can successfully detect clock glitch based attacks, and the required area overhead is only 0.47% that is much smaller than previous works. © 2013 IEEE.

In this paper, we propose an adaptive voltage huddle-based distributed-register architecture (AVHDR architecture) that integrates dynamic multiple supply voltages and interconnection delays into high-level synthesis. Next, we propose a high-level synthesis algorithm for AVHDR architectures. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, huddles, each of which abstracts modules placed close to each other, are naturally generated using floorplanning. Low-supply voltages are assigned to non-critical operations, and leakage power is cut off by turning off the power supply to the sleeping functional units. Experimental results show that our algorithm achieves 50% energy-saving compared with conventional algorithms.

In this paper, we propose a high-level synthesis algorithm with post-silicon delay tuning for RDR architectures. We first obtain a non-delayed scheduling/binding result and a delayed scheduling/binding result. By adding several extra functional units to vacant RDR islands, we have a delayed scheduling/binding result so that its latency cannot be increased compared with the non-delayed one. After that, we similarize the two scheduling/binding results by repeatedly modifying their results. We can finally realize non-delayed and delayed scheduling/binding results simultaneously on RDR architecture with almost no area/performance overheads and we can select either one of them depending on post-silicon delay variation. Experimental results show that our algorithm successfully reduces delayed scheduling/binding latency by up to 42.9% compared with the conventional approach.

In this paper, we propose an adaptive voltage huddle-based distributed-register architecture (AVHDR architecture) that integrates dynamic multiple supply voltages and interconnection delays into high-level synthesis. Next, we propose a high-level synthesis algorithm for AVHDR architectures. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, huddles, each of which abstracts modules placed close to each other, are naturally generated using floorplanning. Low-supply voltages are assigned to non-critical operations, and leakage power is cut off by turning off the power supply to the sleeping functional units. Experimental results show that our algorithm achieves 50% energy-saving compared with conventional algorithms.

Scan test is a powerful test technique which can control and observe the internal states of the circuit under test through scan chains. However, it has been reported that it's possible to retrieve secret keys from cryptographic LSIs through scan chains. Therefore new secure test methods are required to satisfy both testability and security requirements. In this paper, a secure scan design is proposed to achieve adequate security requirement as a countermeasure against scan-based attacks, while still maintain high testability like normal scan testing. In our method, the internal scan chain is divided into several sub chains, and the connection order of sub chains can be dynamically changed. In addition, how to decide the connection order of those sub chains so that it can't be identified by an attacker is also proposed in this paper. The proposed method is implemented on an AES circuit to show its effectiveness, and a security analysis is also given to show how the proposed approach can be used as a countermeasure against those known scan-based attacks. © 2013 IEEE.

Conventionally, circuits are designed to add pessimistic timing margin to solve delay variation problems, which guarantees "always correct" operations. However, due to the fact that such a worst-case condition occurs rarely, the traditional pessimistic design method is therefore becoming one of the main obstacles for designers to achieve higher performance and/or ultra-low power consumption. By monitoring timing error occurrence during circuit operation, adaptive timing error detection and recovery methods have gained wide interests recently as a promising solution. As an extension of existing research, in this paper, we propose a suspicious timing error prediction method for performance or energy efficiency improvement in pipeline designs. Experimental results show that with when compared with typical margin designs, the proposed method can 1) achieve up to 1.41X throughput improvement with in-situ timing error prediction ability; and 2) allow the design to be overclocked by up to 1.88X with "always correct" outputs.

In this paper, we propose a partial redundant fault-secure high-level synthesis algorithm for RDR architectures, where we duplicate a part of the original CDFG and maximize its reliability under a timing constraint. Firstly, our algorithm allocates some new additional functional units to vacant spaces on RDR islands for recomputation and increases the number of duplicated operation nodes. Secondly, it minimizes the number of inserted comparator nodes through re-scheduling/re-binding the recomputation CDFG's nodes. As a result, we will obtain a scheduled/bound recomputation CDFG and renewed functional unit allocation with high reliability. Experimental results demonstrate that our algorithm improves reliability by up to 52% compared with the conventional approach. © 2013 IEEE.

In this paper, a concurrent faulty clock detection method is proposed for crypto circuits against clock glitch based differential fault analysis (DFA). In the proposed method, a non-logic buffer-based delay chain is inserted, and then by monitoring the delay along the delay chain, a possible clock glitch based DFA can be detected. Experimental results on an AES circuit show that the proposed method can successfully detect clock glitch based attacks, and the required area overhead is only 0.47% that is much smaller than previous works.

As leakage power of traditional SRAM becomes larger, a ratio of static energy in total energy of memory architecture becomes also larger. Non-volatile memory (NVM) has many advantages over SRAM, such as high density, low leakage power, and non-volatility, but consumes too much write energy. In this paper, we evaluate energy consumption of two-level cache using NVM in part on mobile processors and confirm that it effectively reduces energy consumption.

Trivium is a synchronous stream cipher using three shift registers running at high speed with simple structure. A scan-based side-channel attack retrieves secret information using scan chains, one of design-for-test techniques. In this paper, a scan-based side-channel attack method against Trivium using scan signatures is proposed. In our method, we focus on a particular I-bit position in a collection of scan chains and then we can attack Trivium even if the scan chain includes other registers than internal state registers in Trivium. Experimental results show that our proposed method successfully retrieves a plaintext from a ciphertext.

A scan-path test is one of the useful design-for-test techniques, in which testers can observe and control registers inside the target LSI chip directly. On the other hand, the risk of side-channel attacks against cryptographic LSIs and modules has been pointed out. In particular, scan-based attacks which retrieve secret keys by analyzing scan data obtained from scan chains have been attracting attention. In this paper, we propose two scan-based attack methods against DES and Triple DES using scan signatures. Our proposed methods are based on focusing on particular bit-column-data in a set of scan data and observing their changes when giving several plaintexts. Based on this property, we introduce the idea of a scan signature first and apply it to DES cryptosystems. In DES cryptosystems, we can retrieve secret keys by partitioning the S-BOX process into eight independent sub-processes and reducing the number of the round key candidates from 248 to 26 × 8 = 512. In Triple DES cryptosystems, three secret keys are used to encrypt plaintexts. Then we retrieve them one by one, using the similar technique as in DES cryptosystems. Although some problems occur when retrieving the second/third secret key, our proposed method effectively resolves them. Our proposed methods can retrieve secret keys even if a scan chain includes registers except a crypto module and attackers do not know when the encryption is really done in the crypto module. Experimental results demonstrate that we successfully retrieve the secret keys of a DES cryptosystem using at most 32 plaintexts and that of a Triple DES cryptosystem using at most 36 plaintexts. © 2013 Information Processing Society of Japan.

With the miniaturization of LSIs and its increasing performance, demand for high-functional portable devices has grown significantly. At the same time, battery lifetime and device overheating are leading to major design problems hampering further LSI integration. On the other hand, the ratio of an interconnection delay to a gate delay has continued to increase as device feature size decreases. We have to estimate interconnection delays and reduce energy consumption even in a high-level synthesis stage. In this paper, we propose a high-level synthesis algorithm for huddle-based distributed-register architectures (HDR architectures) with clock gatings based on concurrency-oriented scheduling/functional unit binding. We assume coarse-grained clock gatings to huddles and we focus on the number of control steps, or gating steps, at which we can apply the clock gating to registers in every huddle. We propose two methods to increase gating steps: One is that we try to schedule and bind operations to be performed at the same timing. By adjusting the clock gating timings in a high-level synthesis stage, we expect that we can enhance the effect of clock gatings more than applying clock gatings after logic synthesis. The other is that we try to synthesize huddles such that each of the synthesized huddles includes registers which have similar or the same clock gating timings. At this time, we determine the clock gating timings to minimize all energy consumption including clock tree energy. The experimental results show that our proposed algorithm reduces energy consumption by a maximum of 23.8% compared with several conventional algorithms.

With process technology scaling, a heat problem in ICs is becoming a serious issue. Since high temperature adversely impacts on reliability, design costs, and leakage power, it is necessary to incorporate thermal-aware synthesis into IC design flows. In particular, hot spots are serious concerns where a chip is locally too much heated and reducing the peak temperature inside a chip is very important. On the other hand, increasing the average interconnect delays is also becoming a serious issue. By using RDR architectures (Regular-Distributed-Register architectures), the interconnect delays can be easily estimated and their influence can be much reduced even in high-level synthesis. In this paper, we propose a thermal-aware high-level synthesis algorithm for RDR architectures. The RDR architecture divides the entire chip into islands and each island has uniform area. Our algorithm balances the energy consumption among islands through re-binding to functional units. By allocating some new additional functional units to vacant areas on islands, our algorithm further balances the energy consumption among islands and thus reduces the peak temperature. Experimental results demonstrate that our algorithm reduces the peak temperature by up to 9.1% compared with the conventional approach. Copyright © 2013 The Institute of Electronics, Information and Communication Engineers.

Scan-based side channel attack on hardware implementations of cryptographic algorithms has shown its great security threat. Unlike existing scan-based attacks, in our work we observed that instead of the secret-related-registers, some non-secret registers also carry the potential of being misused to help a hacker to retrieve secret keys. In this paper, we first present a scan-based side channel attack method on AES by making use of the round counter registers, which are not paid attention to in previous works, to show the potential security threat in designs with scan chains. And then we discussed the issues of secure DFT requirements and proposed a secure scan scheme to preserve all the advantages and simplicities of traditional scan test, while significantly improve the security with ignorable design overhead, for crypto hardware implementations.

Network-on-chip (NoC) architectures have emerged as a promising solution to the lack of scalability in multi-processor systems-on-chips (MPSoCs). With the explosive growth in the usage of multimedia applications, it is expected that NoC serves as a multimedia server supporting multi-class services. In this paper, we propose a configuration algorithm for a hybrid bus-NoC architecture together with simulation results. Our target architecture is a hybrid bus-NoC architecture, called busmesh NoC, which is a generalized version of a hybrid NoC with local buses. In our BMNoC configuration algorithm, cores which have a heavy communication volume between them are mapped in a cluster node (CN) and connected by a local bus. CNs can have communication with each other via edge switches (ESes) and mesh routers (MRs). With this hierarchical communication network, our proposed algorithm can improve the latency as compared with conventional methods. Several realistic applications applied to our algorithm illustrate the better performance than earlier studies and feasibility of our proposed algorithm.

As battery runtime and overheating problems for portable devices become unignorable, energy-aware LSI design is strongly required. Moreover, an interconnection delay should be explicitly considered there because it exceeds a gate delay as the semiconductor devices are downsized. We must take account of energy efficiency and interconnection delays even in high-level synthesis. In this paper, we first propose a huddle-based distributed-register architecture (HDR architecture), an island-based distributed-register architecture for multi-cycle interconnect communications where we can develop several energy-saving techniques. Next, we propose an energy-efficient high-level synthesis algorithm for HDR architectures focusing on multiple supply voltages. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, a huddle, which is composed of functional units, registers, controller, and level converters, are very naturally generated using floorplanning results. By assigning high supply voltage to critical huddles and low supply voltage to non-critical huddles, we can finally have energy-efficient floorplan-aware high-level synthesis. Experimental results show that our algorithm achieves 45% energy-saving compared with the conventional distributed-register architectures and conventional algorithms.

Scan test which is one of the useful design for testability techniques is effective for LSIs including cryptographic circuit. It can observe and control the internal states of the circuit under test by using scan chain. However, scan chain presents a significant security risk of information leakage for scan-based attacks which retrieves secret keys of cryptographic LSIs. In this paper, a secure scan architecture against scan-based attack which still has high testability is proposed. In our method, scan data is dynamically changed by adding the latch to any FFs in the scan chain. We show that by using proposed method, neither the secret key nor the testability of an RSA circuit implementation is compromised, and the effectiveness of the proposed method.

With the miniaturization of LSIs and its increasing performance, demand for high-functional portable devices has grown significantly. At the same time, the problems for battery runtime and device overheating have occurred. On the other hand, the ratio of an interconnection delay to a gate delay has continued to increase as device feature size decreases. We have to estimate the interconnection delay and reduce energy consumption even in a high-level synthesis stage. Recently, an HDR architecture and its associated power-optimized high-level synthesis algorithm have been proposed which can effectively estimate the interconnection delays by introducing the idea of "huddles" into an LSI chip. It utilize multiple supply voltages and achieves power-optimized LSI synthesis but does not take into account the clock gatings. In this paper, we propose a high-level synthesis algorithm based on HDR architectures utilizing clock gatings. Firstly we focus on the number of the control steps at which we can apply the clock gating to registers. Secondly, we synthesize the huddles such that each of the synthesized huddles includes registers which have similar or exactly the same clock gating timings. The experimental results show that our proposed algorithm reduces energy consumption by a maximum of 14.9% compared with the conventional algorithm.

Network-on-chip (NoC) architectures are emerged as a promising solution to the lack of scalability in multi-processor systems-on-chips (MPSoCs). In this paper, we propose a novel BMNoC configuration algorithm together with simulation results. Our BMNoC configuration algorithm analyses the data traffic of the target application and determines which core is the right one to put into the certain cluster with its communication volume and locality. Furthermore, the simulation results illustrate the better latency than earlier studies and feasibility of BMNoC.

In this paper, we first propose a huddle-based distributed-register architecture (HDR architecture), an island-based distributed-register architecture for multi-cycle interconnect communications where we can develop several energy-saving techniques. Next, we propose an energy-efficient high-level synthesis algorithm for HDR architectures focusing on multiple supply voltages. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, huddles, each of which is composed of functional units, registers, controller, and level converters, are very naturally generated using floorplanning results. By assigning high supply voltage to critical huddles and low supply voltage to non-critical huddles, we can finally have energy-efficient floorplan-aware high-level synthesis. Experimental results show that our algorithm achieves 45% energy-saving compared with the conventional distributed-register architectures and conventional algorithms.

Scan test is one of the useful design for testability techniques, which can detect circuit failure efficiently. However, it has been reported that it's possible to retrieve secret keys from cryptographic LSIs through scan chains. Therefore testability and security contradicted to each other, and there is a need to an efficient design for testability circuit so as to satisfy both testability and security requirement. In this paper, a secure scan architecture against scan-based attack is proposed to achieve high security without compromising the testability. In our method, scan structure is dynamically changed by adding the latch to any FFs in the scan chain. We made an analysis on an RSA circuit implementation to show the effectiveness of the proposed method and discussed how our approach is resistant to scan-based attack.

Super-resolution is a technique to remove the noise of observed images and restore its high frequencies. We focus on reconstruction-based super-resolution. Reconstruction requires large computation cost since it requires many images. In this paper, we propose a fast weighted adder for reconstruction-based super-resolution. From the viewpoint of reducing partial products, we propose two approaches to speed up a weighted adder. First, we use selector logics to halve its partial products. Second, we propose a weights-range limit method utilizing negative term. By applying our proposed approaches to a weighted adder, we can reduce carry propagations and our weighted adder can be designed by a fast circuit as compared to conventional ones. Experimental evaluations demonstrate that our weighted adder improves the performance by a maximum of 29.9% and reduces a maximum of 592 LUTs, compared to conventional implementations.

With the high integration of LSI in recent years, the importance of design-for-techniques has been increasing. A scan-path test is one of the useful design-for-test techniques, in which testers can observe and control registers inside the target LSI chip directly. On the other hand, the risk of side-channel attacks against cryptographic LSIs and modules has been pointed out. In particular, scan-based attacks which retrieve secret keys by analyzing scan data obtained from scan chains has been attracting attention. In this paper, we propose a scan-based attack method against DES using scan signatures. Our proposed method are based on focusing on particular bit-column-data in a set of scan data and observing their changes when given several plaintexts. We can retrieve secret keys by partitioning the S-BOX process into eight independent sub-processes and reducing the number of the round key candidates from 2(48) to 2(6) x 8 = 512. Our proposed methods can retrieve secret keys even if a scan chain includes registers except a crypto module and attackers do not know when the encryption is really done in the crypto module. Experimental results demonstrate that we successfully retrieve the secret keys of a DES cryptosystem using at most 32 plaintexts.

Network-on-chip architectures have emerged as a promising solution to the lack of scalability in multi-processor systems-on-chips (MPSoCs). With the explosive growth in the usage of multimedia applications, it is expected that NoC serves as a multimedia server supporting multi-class services. Recently, a busmesh NoC (BMNoC) has been proposed. The BMNoC architecture, which analyses the data traffic and makes aware of localities between cores, improves the system performance in terms of latency as compared with conventional NoCs. In this paper, we propose a novel BMNoC utilizing packet transmission priority control methods. Our proposed BMNoC is a generalized and simplified version of a hybrid NoC which is composed of local buses and global mesh routers. Several realistic applications applied to our algorithm illustrate the better performance than previous studies and feasibility of our proposed architecture.

Scan technology carries the potential risk of being misused as a "side channel" to leak out the secrets of crypto cores. The existing scan-based attacks could be viewed as one kind of differential cryptanalysis, which takes advantages of scan chains to observe the bit changes between pairs of chosen plaintexts so as to identify the secret keys. To address such a design/test challenge, this paper proposes a robust secure scan structure design for crypto cores as a countermeasure against scan-based attacks to maintain high security without compromising the testability.

As battery runtime and overheating problems for portable devices become unignorable, energy-aware LSI design is strongly required. Moreover, an interconnection delay should be explicitly considered there because it exceeds a gate delay as the semiconductor devices are downsized. We must take account of energy efficiency and interconnection delays even in high-level synthesis. In this paper, we first propose a huddle-based distributed-register architecture (HDR architecture), an island-based distributed-register architecture for multi-cycle interconnect communications where we can develop several energy-saving techniques. Next, we propose an energy-efficient high-level synthesis algorithm for HDR architectures focusing on multiple supply voltages. Our algorithm is based on iterative improvement of scheduling/binding and floorplanning. In the iteration process, a huddle, which is composed of functional units, registers, controller, and level converters, are very naturally generated using floorplanning results. By assigning high supply voltage to critical huddles and low supply voltage to non-critical huddles, we can finally have energy-efficient floorplan-aware high-level synthesis. Experimental results show that our algorithm achieves 45% energy-saving compared with the conventional distributed-register architectures and conventional algorithms. © 2012 Information Processing Society of Japan.

In recent years, it is quite necessary to convert conventional low-resolution images to high-resolution ones at low cost. Super-resolution is a technique to remove the noise of observed images and restore its high frequencies. We focus on reconstruction-based super-resolution. Reconstruction requires large computation cost since it requires many images. In this paper, we propose a fast weighted adder for reconstruction-based super-resolution. From the viewpoint of reducing partial products, we propose two approaches to speed up a weighted adder. First, we use selector logics to halve its partial products. Second, we propose a weights-range limit method utilizing negative term. By applying our proposed approaches to a weighted adder, we can reduce carry propagations and our weighted adder can be designed by a fast circuit as compared to conventional ones. Experimental evaluations demonstrate that our weighted adder reduces its delay time by a maximum of 25.29% and its area to a maximum of 1/3, compared to conventional implementations. © 2012 Information Processing Society of Japan.

In this paper, we propose multiple-supply-voltages aware high-level synthesis algorithm for HDR architectures which realizes high-speed and high-efficient circuits. We propose three new techniques: virtual area estimation, virtual area adaptation, and floorplanning-directed huddling, and integrate them into our HDR architecture synthesis algorithm. Virtual area estimation/adaptation effectively estimates a huddle area by gradually reducing it during iterations, which improves the convergence of our algorithm. Floorplanning-directed huddling determines huddle composition very effectively by performing floorplanning and functional unit assignment inside huddles simultaneously. Experimental results show that our algorithm achieves about 29% run-time-saving compared with the conventional algorithms, and obtains a solution which cannot be obtained by our original algorithm even if a very tight clock constraint is given.

With the progress of process technology in recent years, low voltage power supplies have become quite predominant. With this, the voltage margin has decreased and therefore the on-chip decoupling capacitance optimization that satisfies the voltage drop constraint becomes more important. In addition, the reduction of the on-chip decoupling capacitance area will reduce the chip area and, therefore, manufacturing costs. Hence, we propose an algorithm that satisfies the voltage drop constraint and at the same time, minimizes the total on-chip decoupling capacitance area. The proposed algorithm uses the idea of the network algorithm where the path which has the most influence on voltage drop is found. Voltage drop is improved by adding the on-chip capacitance to the node on the path. The proposed algorithm is efficient and effectively adds the on-chip capacitance to the greatest influence on the voltage drop. Experimental results demonstrate that, with the proposed algorithm, real size power/ground network could be optimized in just a few minutes which are quite practical. Compared with the conventional algorithm, we confirmed that the total on-chip decoupling capacitance area of the power/ground network was reducible by about 40 similar to 50%.

The number of sets, block size, and associativity determine processor's cache configurations. Particularly in embedded systems, their cache configuration can be optimized since their target applications are much limited. Recently, the CRCB method has been proposed for LRU-based (Least Recently Used-based) cache configuration simulation, which can calculate cache hit/miss counts accurately and very fast changing the three parameters. However many recent processors use FIFO-based (First-In-First-Out-based) caches instead of LRU-based caches due to the viewpoints of their hardware costs. In this paper, we propose a speeding-up cache configuration simulation method for embedded applications that uses FIFO as a cache replacement policy. We first prove several properties for FIFO-based caches and then propose a simulation method that can process two or more FIFO-based cache configurations with different cache associativities simultaneously. Experimental results show that our proposed method can obtain accurate cache hits/misses and runs up to 32% faster than the conventional simulators.

With the process technological progress in recent years, low voltage power supplies have become quite predominant. With this, the voltage margin has decreased and therefore the power/ground design that satisfies the voltage drop constraint becomes more important. In addition, the reduction of the power/ground total wiring area and the number of layers will reduce manufacturing and designing costs. So, we propose an algorithm that satisfies the voltage drop constraint and at the same time, minimizes the power/ground total wiring area. The proposed algorithm uses the idea of a network algorithm [I] where the edge which has the most influence on voltage drop is found. Voltage drop is improved by changing the resistance of the edge. The proposed algorithm is efficient and effectively updates the edge with the greatest influence on the voltage drop. From experimental results, compared with the conventional algorithm, we confirmed that the total wiring area of the power/ground was reducible by about 1/3. Also, the experimental data shows that the proposed algorithm satisfies the voltage drop constraint in the data whereas the conventional algorithm cannot.

Large-scale network and multimedia application LSIs include application specific arithmetic units. A multiply-accumulator unit or a MAC unit which is one of these optimized units arranges partial products and decreases carry propagations. However, there is no method similar to MAC to execute “subtract-multiplication”. In this paper, we propose a high-speed subtract-multiplication unit that decreases latency of a subtract operation by bit-level transformation using selector logics. By using bit-level transformation, its partial products are calculated directly. The proposed subtract-multiplication units can be applied to any types of systems using subtract-multiplications and a butterfly operation in FFT is one of their suitable applications. We apply them effectively to Radix-2 butterfly units and Radix-4 butterfly units. Experimental results show that our proposed operation units using selector logics improves the performance by up to 13.92%, compared to a conventional approach.

Since target applications running on an embedded processor are much limited in embedded systems, we can optimize its cache configuration based on the number of sets, block size, and associativities. An extremely fast cache configuration simulation method, CRCB (Configuration Reduction approach by the Cache Behavior), has been recently proposed which can calculate cache hit/miss counts accurately for possible cache configurations when the three parameters above are changed. The CRCB method assumes LRU-based (Least Recently Used-based) cache but many recent processors use FIFO-based (First In First Out-based) cache or PLRU-based (Pseudo LRU-based) cache due to its hardware cost. In this paper, we propose exact and fast L1 cache configuration simulation algorithms for embedded applications that use PLRU or FIFO as a cache replacement policy. Firstly, we prove that the CRCB method can be applied not only to LRU but also to other cache replacement policies including FIFO and PLRU. Secondly, we prove several properties for FIFO- and PLRU-based caches and we propose associated cache simulation algorithms which can simulate simultaneously more than one cache configurations with different cache associativities accurately for FIFO or PLRU. Finally, many experimental results demonstrate that our cache configuration simulation algorithms obtain accurate cache hit/miss counts and run up to 249 times faster than a conventional cache simulator.

Since target applications in embedded systems are limited, we can optimize its cache configuration. A very fast and exact cache simulation algorithm, CRCB, has been recently proposed. CRCB assumes LRU as a cache replacement policy but FIFO- or PLRU-based cache is often used due to its low hardware cost. This paper proposes exact and fast L1 cache simulation algorithms for PLRU- or FIFO-based caches. First, we prove that CRCB can be applied to FIFO and PLRU. Next, we show several properties for FIFO- and PLRU-based caches and propose their associated cache-simulation speed-up algorithms. Experiments demonstrate that our algorithms run up to 300 times faster than a well-known cache simulator.

Since target applications running on an embedded processor are much limited in embedded systems, we can optimize its cache configuration based on the number of sets, block size, and associativities. An extremely fast cache configuration simulation method, CRCB (Configuration Reduction approach by the Cache Behavior), has been recently proposed which can calculate cache hit/miss counts accurately for possible cache configurations when the three parameters above are changed. The CRCB method assumes LRU-based (Least Recently Used-based) cache but many recent processors use FIFO-based (First In First Out-based) cache or PLRU-based (Pseudo LRU-based) cache due to its hardware cost. In this paper, we propose exact and fast L1 cache configuration simulation algorithms for embedded applications that use PLRU or FIFO as a cache replacement policy. Firstly, we prove that the CRCB method can be applied not only to LRU but also to other cache replacement policies including FIFO and PLRU. Secondly, we prove several properties for FIFO- and PLRU-based caches and we propose associated cache simulation algorithms which can simulate simultaneously more than one cache configurations with different cache associativities accurately for FIFO or PLRU. Finally, many experimental results demonstrate that our cache configuration simulation algorithms obtain accurate cache hit/miss counts and run up to 249 times faster than a conventional cache simulator. © 2011 Information Processing Society of Japan.

As device feature size decreases, the reliability improvement against soft errors becomes quite necessary. A fault-secure system, in which concurrent error detection is realized, is one of the solutions to this problem. On the other hand, average interconnection delays exceed gate delays which leads to a serious timing closure problem. By using regular-distributed-register architecture (RDR architecture), we can estimate interconnection delays very accurately and their influence can be much reduced even in behavioral-level design. In this paper, we propose a fault-secure high-level synthesis algorithm for an RDR architecture. In fault-secure high-level synthesis, a recomputation CDFG as well as a normal-computation CDFG must be scheduled to control steps and bound to functional units. Firstly, our algorithm re-uses vacant areas on RDR islands to allocate new function units additionally for the recomputation CDFG. Secondly, we propose an efficient edge-break algorithm which considers comparison nodes' scheduling/binding. We can have small-latency scheduling/binding for both the normal CDFG and recomputation CDFG. Our algorithm reduces the required control steps by up to 53% compared with the conventional approach. © 2011 Information Processing Society of Japan.

Large-scale network and multimedia application LSIs include application specific arithmetic units. A multiply-accumulator unit or a MAC unit which is one of these optimized units arranges partial products and decreases carry propagations. However, there is no method similar to MAC to execute "subtractmultiplication". In this paper, we propose a high-speed subtract-multiplication unit that decreases latency of a subtract operation by bit-level transformation using selector logics. By using bit-level transformation, its partial products are calculated directly. The proposed subtract-multiplication units can be applied to any types of systems using subtract-multiplications and a butterfly operation in FFT is one of their suitable applications. We apply them effectively to Radix- 2 butterfly units and Radix-4 butterfly units. Experimental results show that our proposed operation units using selector logics improves the performance by up to 13.92%, compared to a conventional approach. © 2011 Information Processing Society of Japan.

A scan-path test is one of the most important testing techniques, but it can be used as a side-channel attack against a cryptography circuit. Scan-based attacks are techniques to decipher a secret key using scanned data obtained from a cryptography circuit. Public-key cryptography, such as RSA and elliptic curve cryptosystem (ECC), is extensively used but conventional scan-based attacks cannot be applied to it, because it has a complicated algorithm as well as a complicated architecture. This paper proposes a scan-based attack which enables us to decipher a secret key in ECC. The proposed method is based on detecting intermediate values calculated in ECC. We focus on a 1-bit sequence which is specific to some intermediate values. By monitoring the 1-bit sequence in the scan path, we can find out the register position specific to the intermediate value in it and we can know whether this intermediate value is calculated or not in the target ECC circuit. By using several intermediate values, we can decipher a secret key. The experimental results demonstrate that a secret key in a practical ECC circuit can be deciphered using 29 points over the elliptic curve E within 40 seconds. © 2011 Information Processing Society of Japan.

Scan based side channel attacks retrieve a secret key in a cryptography circuit by analyzing scanned data Since they must be considerable threats to a cryptosystem LSI we have to protect cryptography circuits from them RSA is one of the most important cryptography algorithms because it effectively realizes a public key cryptography system RSA is extensively used but conventional scan based side channel attacks cannot be applied to it because It has a complicated algorithm This paper proposes a scan based side channel attack which enables us to retrieve a secret key in an RSA circuit The proposed method is based on detecting intermediate values calculated in an RSA circuit We focus on a I bit time sequence which is specific to some intermediate values By monitoring the I bit time sequence in the scan path we can find out the register position specific to the intermediate value and we can know whether this intermediate value is calculated or not in the target RSA circuit We can retrieve a secret key one bit by one bit from MSB to LSB The experimental results demonstrate that a 1 024 bit secret key used in the target RSA circuit can be retrieved using 30 2 input messages within 98 3 seconds and its 2 048 bit secret key can be retrieved using, 34 4 input within 634 0 seconds

Due to the limitations of scan structure, the second vector in transition delay test is usually applied either by shift operation or by functional launch, which possibly results in unsatisfying transition delay fault (TDF) coverage. To overcome such a limitation for higher TDF coverage, a novel improved launch delay test technique that combines the pros of launch-on-shift and launch-on-capture tests is introduced in this paper. The proposed method can achieve near perfect TDF coverage with fewer test patterns without the need for a global fast scan enable signal. Experimental results on ISCAS89 and ITC99 benchmark circuits are included to show the effectiveness of the proposed method.

Scan test is a powerful and popular test technique because it can control and observe the internal states of the circuit under test. However, scan path would be used to discover the internals of crypto hardware, which presents a significant security risk of information leakage. An interesting design-for-test technique by inserting inverters into the internal scan path to complicate the scan structure has been recently presented. Unfortunately, it still carries the potential of being attacked through statistical analysis of the information scanned out from chips. Therefore, in this paper we propose secure scan architecture, called dynamic variable secure scan, against scan-based side channel attack. The modified scan flip-flops are state-dependent, which could cause the output of each State-dependent Scan FF to be inverted or not so as to make it more difficult to discover the internal scan architecture.

In this paper, we propose a high-level synthesis method targeting generalized distributed-register architecture in which we introduce shared/local registers and global/local controllers. Functional units on a critical path use local registers and local controllers and functional units on non-critical path use shared register and global controller in our architecture. Our method is based on iterative improvement of scheduling/binding and floorplanning. Using iterative flow, we obtains a generalized distributed-register architecture where its scheduling/binding as well as floorplanning are simultaneously optimized. Experimental results show that 8.6% performance improvement can be achieved compared to the conventional high-performance method.

Scan-based attacks are techniques to decipher a secret key using scanned data obtained from a cryptography circuit. Public-key cryptography, such as RSA and elliptic curve cryptosystem (ECC), is extensively used but conventional scan-based attacks cannot be applied to it, because it has a complicated algorithm as well as a complicated architecture. This paper proposes a scan-based attack which enables us to decipher a secret key in ECC. The proposed method is based on detecting intermediate values calculated in ECC. By monitoring the 1-bit sequence in the scan path, we can find out the register position specific to the intermediate value in it and we can know whether this intermediate value is calculated or not in the target ECC circuit. By using several intermediate values, we can decipher a secret key. The experimental results demonstrate that a secret key in a practical ECC circuit can be deciphered using 29 points over the elliptic curve E within 40 seconds.

Intra-frame coding is one of the most important technologies in H.264/AVC, which made significant contributions to the enhancement of coding efficiency of H.264/AVC at the cost of computation complexity. To address this problem, in this paper we present an efficient VLSI implementation of a computation efficient intra prediction algorithm for H.264/AVC encoding. Unlike most of existing fast intra-mode selection techniques, in the proposed method the directional differences are computed using a few selected original pixels to obtain the candidate modes with the minimal direction cost. The proposed method is hardware-friendly and provides more processing parallelism for H.264 intra-frame encoding with less overhead and less power consumption, which is expected to be utilized as a favourable accelerator hardware module in a real-time HDTV (1920x1080p) H.264 encoder.

Large-scale network and multimedia application LSIs include application specific arithmetic units. A multiplyaccumulator unit (MAC unit) which is one of these optimized units arranges partial products and decreases carry propagations. However, there is no method similar to MAC to execute "subtract-multiplication". In this paper, we propose a high-speed subtract-multiplication unit that decreases latency of a subtract operation by bit-level transformation using selector logics. By using bit-level transformation, its partial products are calculated directly. The proposed subtract-multiplication units can be applied to even any types of systems using subtract-multiplications and a butterfly operation in FFT is one of their suitable applications. Experimental results show that our proposed arithmetic units using selector logics improves the performance by 13.92%, compared to a conventional approach.

Network-on-chip (NoC) architectures are emerged as a promising solution to the lack of scalability in multi-processor systems-on-chips (MPSoCs). In this paper, A busmesh network-on-chip (BMNoC) architecture is proposed, together with simulation results. It is comprised of bus-based connection and global mesh routers to enhance the performance of on-chip communication. Furthermore, MPEG-4, H.264 and a hybrid application mixed MPEG-4 and H.264 on our architecture illustrates the better performance than earlier studies and feasibility of BMNoC.

Recently, two-level cache, L1 cache and L2 cache, is commonly used in a processor. Particularly in an embedded system whereby a single application or a class of applications is repeatedly executed on a processor, its cache configuration can be customized such that an optimal one is achieved. An optimal two-level cache configuration can be obtained which minimizes overall memory access time or memory energy consumption by varying the three cache parameters: the number of sets, a line size, and an associativity, for L1 cache and L2 cache. In this paper, we first extend the L1 cache simulation algorithm so that we can explore two-level cache configuration. Second, we propose two-level cache design space exploration algorithms: CRCB-T1 and CRCB-T2, each of which is based on applying Cache Inclusion Proper v to two-level cache configuration. Each of the proposed algorithms realizes exact cache simulation but decreases the number of cache hit/miss judgments by a factor of several thousands. Experimental results show that. by using our approach. the number of cache hit/miss judgments required to optimize a cache configurations is reduced to 1/50-1/5500 compared to the exhaustive approach. As a result, our proposed approach totally runs an average of 1398.25 times faster compared to the exhaustive approach. Our proposed cache simulation approach achieves the world fastest two-level cache design space exploration.

A scan chain is one of the most important testing techniques, but it can be used as side-channel attacks against a cryptography LSI. We focus on scan-based attacks, in which scan chains are targeted for side-channel attacks. The conventional scan-based attacks only consider the scan chain composed of only the registers in a cryptography circuit. However, a cryptography LSI usually uses many circuits such as memories, micro processors and other circuits. This means that the conventional attacks cannot be applied to the practical scan chain composed of various types of registers. In this paper, a scan-based attack which enables to decipher the secret key in an AES cryptography LSI composed of an AES circuit and other circuits is proposed. By focusing on bit pattern of the specific register and monitoring its change, Our scan-based attack eliminates the influence of registers included in other circuits than AES. Our attack does not depend on scan chain architecture, and it can decipher practical AES cryptography LSIs.

As device feature size decreases, interconnection delay becomes the dominating factor of circuit total delay. Distributed-register architectures call reduce the influence of interconnection delay. They may, however, increase circuit area because they require many local registers. Moreover original distributed-register architectures do not consider control signal delay, which may be the bottleneck in a circuit. In this paper. we propose it high-level synthesis method targeting generalized distributed-register architecture in which we introduce shared/local registers aid global/local controllers. Our method is based on iterative improvement of scheduling/binding and floorplanning. First, we prepare shared-register groups with global controllers, each of which corresponds to it single functional unit. As iterations proceed, we use local registers and local controllers for functional units on it critical path. Shared-register groups physically located close to each other are merged into a single group. Accordingly, global controllers are merged. Finally, our method obtains it generalized distributed-register architecture where its scheduling/binding as well as floorplanning are simultaneously optimized. Experimental results show that the area is decreased by 4.7% while maintaining the performance of the circuit equal with that using original distributed-register architectures.

This paper presents a novel X-handling technique, which removes the effect of unknowns on compacted test response with maximal compaction ratio. The proposed method combines with the current X-tolerant compactors and inserts masking cells on scan paths to selectively mask X's. By doing this, the number of unknown responses in each scan-out cycle could be reduced to a reasonable level such that the target X-tolerant compactor would tolerate with guaranteed possible error detection, It guarantees no test loss due to the effect of X's, and achieves the maximal compaction that the target response compactor could provide as well. Moreover, because the masking cells are only inserted on the scan paths, it has no performance degradation of the designs. Experimental results demonstrate the effectiveness of the proposed method.

Modular multiplication is the most dominant arithmetic operation in elliptic curve cryptography (ECC), that is a type of public-key cryptography. Montgomery multiplier is commonly used to compute the modular multiplications and requires scalability because the bit length of operands varies depending on its security level. In addition, ECC is performed in GF(P) or GF(2(n)), and unified architecture for multipliers in GF(P) and GF(2(n)) is required. However, in previous works, changing frequency is necessary to deal with delay-time difference between GF(P) and GF(2(n)) multipliers because the critical path of the GF(P) multiplier is longer. This paper proposes unified dual-radix architecture for scalable Montgomery multiplications in GF(P) and GF(2(n)). This proposed architecture unifies four parallel radix-2(16) multipliers in GF(P) and a radix-2(64) multiplier in GF(2(n)) into a single unit. Applying lower radix to GF(P) multiplier shortens its critical path and makes it possible to compute the operands in the two fields using the same multiplier at the same frequency so that clock dividers to deal with the delay-time difference are not required. Moreover, parallel architecture in GF(P) reduces the clock cycles increased by dual-radix approach. Consequently, the proposed architecture achieves to compute a GF(P) 256-bit Montgomery multiplication in 0.28 mu s. The implementation result shows that the area of the proposal is almost the same as that of previous works: 39 kgates.

In an embedded system where a single application or a class of applications is repeatedly executed on a processor, its cache configuration can be customized such that an optimal one is achieved. We can have an optimal cache configuration which minimizes overall memory access time by varying the three cache parameters: the number of sets, a line size, and an associativity. In this paper, we first propose two cache simulation algorithms: CRCB1 and CRCB2, based on Cache Inclusion Property. They realize exact cache simulation but decrease the number of cache hit/miss judgments dramatically. We further propose three more cache design space exploration algorithms: CRMF1, CRMF2, and CRMF3, based on our experimental observations. They can find an almost optimal cache configuration from the viewpoint of access time. By using our approach, the number of cache hit/miss judgments required for optimizing cache configurations is reduced to 1/10-1/50 compared to conventional approaches. As a result, our proposed approach totally runs an aver-age of 3.2 times faster and a maximum of 5.3 times faster compared to the fastest approach proposed so far. Our proposed cache simulation approach achieves the world fastest cache design space exploration when optimizing total memory access time.

Scan technology carries the potential of being misused as a "side channel" to leak out the secret information of crypto cores. To address such a design challenge, this paper proposes a design-for-secure-test (DFST) solution for crypto cores by adding a stimuli-launched flip-flop into the traditional scan flip-flop to maintain the high test quality without compromising the security.

In recent years, the gap between the cycle time of processors and memory access time has been increasing. One of the solutions to solve this problem is to use a cache. But just using a large cache may not reduce the total memory access time. We can have an optimal cache configuration which minimizes overall memory access time by varying the three cache parameters: a cache set size, a line size, and an associativity. In this paper, we propose two exact cache simulation algorithms: CRCB1 and CRCB2, based on Cache Inclusion Property. They realize exact cache simulation but increase simulation speed dramatically. By using our approach, the number of cache hit/miss judgments required for simulating all the cache configurations is reduced to 31.4%-93.6% compared to conventional approaches. As a result, our proposed approach totally runs an average of 1.8 times faster and a maximum of 3.3 times faster compared to the fastest approach proposed so far. Our proposed exact cache simulation approach achieves the world fastest L1 cache simulation.

This paper presents a unified test compression technique for scan stimulus and unknown masking data with seamless integration of test generation, test compression and all unknown response masking for high quality manufacturing test cost reduction. Unlike prior test compression methods. the proposed approach considers the unknown responses during test pattern generation procedure, and then selectively encodes, the less specified bits (either Is or Os) in each scan slice for compression while at the same time masks the unknown responses before sending them to the response compactor. The proposed test scheme could dramatically reduce test data volume as well as the number of required test channels by using only c tester channels to drive N internal scan chains, where c = inverted right perpendicular log(2) N inverted left perpendicular + 2. In addition, because all the unknown responses could be exactly masked before entering into the response compactor, test loss due to unknown responses would be eliminated. Experimental results oil both benchmark circuits and larger designs indicated the effectiveness of the proposed technique.

In this paper, we propose a high-level synthesis method targeting distributed/shared-register architectures. Our method repeats (1) scheduling/ FU binding, (2) register allocation, (3) register binding, and (4) module placement. By feeding back floorplan information from (4) to (1), our method obtains a distributed/shared-register architecture where its scheduling/binding as well as floorplaning are simultaneously optimized. Experimental results show that the area is decreased by 13.2% while maintaining the performance of the circuit equal with that using distributed-register architectures. © 2008 Information Processing Society of Japan.

Reducing the power dissipation for LDPC code decoder is a major challenging task to apply it to the practical digital communication systems. In this paper, we propose a low power LDPC code decoder architecture based on an intermediate message-compression technique which features as follows: (i) An intermediate message compression technique enables the decoder to reduce the required memory capacity and write power dissipation. (H) A clock gated shift register based intermediate message memory architecture enables the decoder to decompress the compressed messages in a single clock cycle while reducing the read power dissipation. The combination of the above two techniques enables the decoder to reduce the power dissipation while keeping the decoding throughput. The simulation results show that the proposed architecture improves the power efficiency up to 52% and 18% compared to that of the decoder based on the overlapped schedule and the rapid convergence schedule without the proposed techniques respectively.

In this paper, we presented a Design-for-Secure-Test (DFST) technique for pipelined AES to guarantee both the security and the test quality during testing. Unlike previous works, the proposed method can keep all the secrets inside and provide high test quality and fault diagnosis ability as well. Furthermore, the proposed DFST technique can significantly reduce test application time, test data volume, and test generation effort as additional benefits.

In this paper, we propose a high-level synthesis method targeting distributed/shared-register architectures. Our method repeats (1) scheduling/FU binding, (2) register allocation, (3) register binding, and (4) module placement. By feeding back floorplan information from (4) to (1), our method obtains a distributed/shared-register architecture where its scheduling/binding as well as floorplaning are simultaneously optimized. Experimental results show that the area is decreased by 13.2% while maintaining the performance of the circuit equal with that using distributed-register architectures.

In this paper, we propose a high-level synthesis method targeting distributed/shared-register architectures. Our method repeats (1) scheduling/FU binding, (2) register allocation, (3) register binding, and (4) module placement. By feeding back floorplan information from (4) to (1), our method obtains a distributed/shared-register architecture where its scheduling/binding as well as floorplaning are simultaneously optimized. Experimental results show that the area is decreased by 13.6% while maintaining the performance of the circuit equal with that using distributed-register architectures.

Modular multiplication is the most dominant arithmetic operation in elliptic curve cryptography (ECC), which is a type of public-key cryptography. Montgomery multiplication is commonly used as a technique for the modular multiplication and required scalability since the bit length of operands varies depending on the security levels. Also, ECC is performed in GF(P) or GF(2), and unified architectures for GF(P) and GF(2(n)) Multiplier are needed. However, in previous works, changing frequency or dual-radix architecture is necessary to deal with delay-time difference between GF(P) and GF(2(n)) circuits of the multiplier because the critical path of GF(P) circuit is longer. This paper proposes a scalable unified dual-radix architecture for Montgomery multiplication in GF(P) and GF(2(n)). The proposed architecture unifies 4 parallel radix-2(16) multipliers in GF(P) and a radix-2(64) multiplier in GF(2(n)) into a single unit Applying lower radix to GF(P) multiplier shortens its critical path and makes it possible to compute the operands in the two fields using the same multiplier at the same frequency so that clock dividers to deal with the delay-time difference are not required. Moreover, parallel architecture in GF(P) reduces the clock cycles increased by dual-radix approach. Consequently, the proposed architecture achieves to compute GF(P) 256-bit Montgomery multiplication in 0.23 mu s.

This paper introduces GECOM technology, a novel test compression method with seamless integration of test GEneration, test COmpression (i.e. integrated compression on scan stimulus and masking bits) and all unknown scan responses Masking for manufacturing test cost reduction. Unlike most of prior methods, the proposed method considers the unknown responses during ATPG procedure and selectively encodes the specified 1 or 0 bits (either Is or Os) in scan slices for compression while at the same time masks the unknown responses before sending them to the response compactor. The proposed GECOM technology consists of GECOM architecture and GECOM ATPG technique. In the GECOM architecture, for a circuit with N internal scan chains, only c tester channels, where c = [log(2) N] +2, are required. GECOM ATPG generates test patterns for the GECOM architecture thus not only the scan inputs could be efficiently compressed but also all the unknown responses would be masked. Experimental results on both benchmark circuits and real industrial designs indicated the effectiveness of the proposed GECOM technique.

This paper presents a new unknown response masking technique to minimize the effect on test loss due to over-masking. Unlike previous works where the scan responses are masked before entering the response compactor, the proposed method could mask the Xs when they are transformed on the scan path. Meanwhile, the masking cells are inserted along the scan paths, thus they would have no degradation on the performance of the designs. In addition, the test data required to mask unknown responses is only one bit for each test pattern. Experimental results show the effectiveness of the proposed method.

Recently a demand for high-speed wireless network service on mobile devices is rapidly increasing. Error correcting codes are used to enhance network communication quality. Particularly, LDPC (Low Density Parity Check) codes show high throughput and achieve information rates very close to the Shannon limit. In this paper, we propose a dynamically reconfigurable architecture for mufti-rate compatible regular LDPC decoding. Our proposed decoder deals with mufti-rate codes by introducing a mufti-rate compatible 1st-2nd minimum searching unit. The proposed decoder shows the better throughput over the wide range of S/N ratio compared to conventional rate-fixed LDPC decoders.

Reconfigurable processors are those whose contexts are dynamically reconfigured while they are working. We focus on a reconfigurable processor called FE-GA (Flexible Engine/Generic ALU array) for digital media processing. Currently, FE-GA does not have its dedicated behavior synthesis tool. In this paper, we design FIR filters and propose an algorithm to map them onto it automatically. For given an order and coefficients of an FIR filter, the algorithm generates a dedicated assembly code which represents a given FIR filter for FE-GA. Then an editor called FEEditor reads the generated assembly code and implements its corresponding FIR filter on FE-GA. The proposed algorithm achieves automatic mapping of FIR filters of all orders within the range of the specification of FE-GA architecture. Furthermore, it is proved that a minimum cycle is achieved to execute FIR filtering if there is no thread switching.

近年撮像デバイスが普及し,ユーザの保有するマルチメディアコンテンツが増大している.そのため,画像や動画を圧縮する技術から効率的に扱う(検索・グルーピング等)技術が注目されている.本研究ではデジタル写真を検索・グルーピングするデジタル写真アルバムに着目し,その機能を携帯機器へ載せるための提案を行う.現在,コンテンツの特徴量(メタデータ)を抽出し検索等を行うメタデータ生成処理がボトルネックとなり携帯機器へ実装されていない.また,携帯機器供給者がメタデータを用いて様々なサービスを提供することを見据え,本研究ではメタデータ生成処理部分を対象とする.本研究で設計するメタデータ生成システムはMPEG-A Photo Playerが規格化しているメタデータを生成する.本システムは携帯機器での実行を前提とし,1秒以内でのメタデータ生成を目指し,メタデータ生成手法の処理量の削減と専用ハードウェアを用いた処理の高速化を提案する.具体的にはメタデータ生成において最大のボトルネックとなるクラスタリングの代替,Radon変換の高速化,画像の縮小を用いて処理量を削減する.この処理量の削減とともに専用ハードウェアを用いることで約200万画素の画像から0.87秒でメタデータを生成することができ,制約とした1秒以内での実行を実現した.

PackedSIMD型命令セットを持つプロセッサを対象としたHW/SW協調合成システムSPADESでは,PackedSIMD型命令を持つプロセッサを対象とした並列化コンパイラを必要とする.この並列化コンパイラは,まずSPADESが対象とするプロセッサコアにおいて,付加可能なすべてのハードウェアユニットを持つ仮想的なプロセッサを仮定する.仮想プロセッサ上で,入力アプリケーションの命令レベル並列度をPackedSIMD型命令を用いて最大限に抽出し,実行時間最小となるようにスケジューリングされたアセンブリコードを出力する.並列化コンパイラの出力により,合成するプロセッサの初期構成を得られる.本稿では,並列化コンパイラの中核をなす多重ループ構造の並列化アルゴリズムを提案する.提案手法では入力アプリケーション内の多重ループ構造に対して命令レベル並列性を抽出し,PackedSIMD型命令を生成可能とする.提案手法により,アプリケーションの命令レベル並列性をPackedSIMD型命令を利用して抽出し,高速化されたアセンブリコードを出力する.計算機実験により提案手法の有効性を評価する.

本稿では,SIMD型プロセッサコアを対象としたハードウェア/ソフトウェア(HW/SW)協調合成システム「SPADES」においてHW/SW分割フレームワークを提案する.SPADESはアプリケーションに特化した面積・性能に過不足のないプロセッサコアの自動合成を目的とする.SPADESの核となるHW/SW分割では,アプリケーションを最速処理可能となるようなプロセッサコア構成からプロセッサコアに付加される演算器やレジスタ等のハードウェアユニットを削減していき,最適なプロセッサコア構成を探索する.提案するHW/SW分割フレームワークにより,プロセッサコア構成探索を可能とし,アプリケーションの性能要求に応じた小面積のSIMD型プロセッサコア構成を得ることが出来る.また,提案フレームワークはモジュール単位で構成されているため,変更および拡張を容易に実現できる.計算機実験により提案フレームワークを評価し,結果を報告する.

画像処理や音声処理などのアプリケーションを実行する処理系プロセッサは,対象とするアプリケーションに特化した演算ユニットをもつことで,汎用のプロセッサよりも高速,小型かつ低費消費電力を両立することができる.このような,特殊なデータパスを持つ演算ユニットを我々はプロセッシングユニットと呼んでいる.アプリケーションに応じてプロセッシングユニットを設計を行うには時間的なコストがかかるため,プロセッシングユニットを自動合成できるシステムが必要とされている.本稿では,アプリケーションに特化したSIMD型プロセッサコアの自動合成において,合成されるプロセッサコアに付加するプロセッシングユニットを最適化する手法を提案する.提案手法では,アプリケーションのControl Data Flow Graph(CDFG)構造から複数の算術演算および論理演算ノードに対してクラスタリングを行ない,プロセッシングユニットの自動生成を行なう.生成されたプロセッシングユニットをプロセッサコアに組み込み,プロセッサコアの最適化と同時にプロセッシングユニットの再構成を行ないながら最適な構成の探索を行なう.計算機実験により提案手法を評価し結果を報告する.

This paper proposes the power-efficient LDPC decoder architecture which features (1) a FIFO buffering based rapid convergence schedule which enables the decoder to accelerate the decoding throughput without increasing the required number of memory bits, (2) an intermediate message compression technique based on a clock gated shift register which reduces the read and write, power dissipation for the intermediate messages. Simulation results show that the proposed decoder achieves 1.66 times faster decoding throughput, and improves the power efficiency (which is defined by the power dissipation per Mbps) up to 52% compared to the decoder based on the conventional overlapped schedule.

Cryptography plays an important role in the security of data transmission. To ensure the correctness of crypto hardware, we should conduct testing at fabrication and infield. However, the state-of-the-art scan-based test techniques, to achieve high test qualities, need to increase the testability of the circuit under test, which carries a potential of being misused to reveal the secret information of the crypto hardware. Thus, to develop efficient test strategies for crypto hardware to achieve high test quality without compromising security becomes an important task. In this paper we discuss the development of a Design-for-Secure-Test (DFST) technique for pipelined AES to overcome the above contradiction between security and test quality in testing crypto hardware. Unlike previous works, the proposed method can keep all the secrets inside and provide high test quality and fault diagnosis ability as well. Furthermore, the proposed DFST technique can significantly reduce test application time, test data volume, and test generation effort as additional benefits.

In this paper, we propose a power-efficient LDPC decoder architecture based on an accelerated message-passing schedule. The proposed decoder architecture is characterized as follows: (i) Partitioning a pipelined operation not to read and write intermediate messages simultaneously enables the accelerated message-passing schedule to be implemented with single-port SRAMs. (H) FIFO-based buffering reduces the number of SRAM banks and words of the LDPC. decoder based on the accelerated message-passing schedule.. The proposed LDPC decoder keeps a single message for each non-zero bit in a parity check matrix as well as a classical schedule while achieving the accelerated message-passing schedule. Implementation results in 0.18 [mu m] CMOS technology show that the proposed decoder architecture reduces an area of the LDPC decoder by 43% and a power dissipation by 29% compared to the conventional architecture based on the accelerated message-passing schedule.

近年,携帯電話などのモバイル端末での大容量コンテンツ配信など,無線環境での高速通信の需要が高まりつつある.無線通信の品質を向上する必要がある中で1つの手段として誤り訂正符号が用いられる.本稿では,線形時間のオーダで復号が可能で従来の誤り訂正符号よりも高い復号特性を持つ新しい誤り訂正符号であるLDPC(Low Density Parity Check)符号を対象アプリケーションとした復号器を提案する.提案する復号器はmin-sum法を使用し,通信路の雑音特性に応じたマルチレート符号語の復号に対応する.マルチレート対応型最小・第2最小値探索回路を提案し設計した.このことにより,従来までの符号化率固定型の復号器と比較して幅広いS/N比において,より優れた復号スループットを出すことが可能である

携帯電話による位置情報サービスとインターネットサービスの普及により,歩行者を対象とした地図サービスの利用が拡大している.これに伴い,表示面積の狭いモバイル端末に有効な略地図を自動生成するための各種技術の研究が盛んに行われている.道路形状の水平・垂直化,交差点角度の量子化を基本とする従来手法では碁盤の目のようなデザイン性の高い略地図の生成が可能であるが,それらがユーザにとって迷いにくい地図であるとは限らない.本稿では,人間の方向判断基準や交差点形状が歩行者に与える影響を反映させた略地図を生成するための簡略化処理アルゴリズムを提案し,携帯電話のような微小画面においても視認性が良く,かつ迷いにくい略地図の生成を目指す.ノード数400程度の道路ネットワークデータに対して本手法を適用し略地図が生成されることを確認した.

現在,携帯電話を用いて目的地まで利用者を経路案内するサービスが展開している.しかしながら,展開されている経路案内サービスの対象は屋外空間のみである.本稿では,経路案内サービスの対象を屋内空間とし,地下街や百貨店等の屋内空間に特化したネットワークデータを構築する.さらに,携帯端末を用いて屋内用歩行者ナビゲーションサービスを利用することを想定し,歩行者の嗜好性を反映した最適経路を提供することを目的として,屋内に特化した経路探索手法を提案する.提案手法の有効性を示すために2種類のシミュレーション実験を実施し,ユーザにとって最適な経路を出力することができることを示す.

近年,移動体通信網の発達などを背景に,携帯情報端末を利用した歩行者用ナビゲーションシステムに関する研究が盛んに実施されている.この種のシステムでは,ユーザの嗜好を反映させたナビゲーションを可能にすることでナビゲーションシステムの使いやすさの向上が期待できる.そこで本稿では,デパートなどの屋内環境を対象とした飲食店を推測するナビゲーションシステムにおいて,ユーザの詳細な要求に対応した目的地推測手法を提案する.本システムでは,ユーザの食事に対する嗜好性や履歴情報,および店の混雑時間をもとに,ユーザの嗜好を満足する目的地の候補を求め,ユーザ・システム間のインタラクションによって目的地を決定する.本手法による実験システムを構築した結果,その有効性を確認した.

近年のITS技術の進化に伴ってカーナビにおける測位精度や経路案内の技術が高まりつつあるが,出発地から目的地までの所要時間の測定の精度は未だに十分とは言えず,いかにして正確な渋滞情報を取得するかが課題となっている.特に車線ごとに混雑状況が異なることが大きな影響を及ぼすため,既存の渋滞情報の検出方法の問題を起こすことなく,交差点などにおいて車線ごとに混雑状況が異なる場合があったとしてもそれを個別に検出することが必要である.そこで本稿では,車車間通信および路車間通信を用いることにより,一般道路の各交差点において車線別に渋滞情報をリアルタイムに検出するための手法を提案する.ビーコンから通信を開始して車車間通信を繰り返すことで車両の情報を収集し,その情報から渋滞を通過するまでに要した時間を算出する.また,シミュレーションによりこの手法の有効性を示す.

This paper presents a test input data compression technique, Selective Low-Care Coding (SLC), which can be used to significantly reduce input test data volume as well as the external test channel requirement for multiscan-based designs. In the proposed SLC scheme, we explored the linear dependencies of the internal scan chains, and instead of encoding all the specified bits in test cubes, only a smaller amount of specified bits are selected for encoding, thus greater compression can be expected. Experiments on the larger benchmark circuits show drastic reduction in test data volume with corresponding savings on test application time can be indeed achieved even for the well-compacted test set.

In this paper, we propose a partially-parallel LDPC decoder which achieves a high-efficiency message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the column operation module to compute every message in each bit node which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay when the row and column operations are performed concurrently. Therefore, the proposed decoder performs the column operations more frequently in a single iterative decoding, and achieves a high-efficiency message-passing schedule within the limited decoding delay time. Hardware implementation on an FPGA and simulation results show that the proposed partially-parallel LDPC decoder improves the decoding throughput and bit error performance with a small hardware overhead.

In this paper, we propose a partially-parallel LDPC decoder which achieves a high-efficiency message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the column operation module to compute every message in each bit node which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay when the row and column operations are performed concurrently. Therefore, the proposed decoder performs the column operations more frequently in a single iterative decoding, and achieves a high-efficiency message-passing schedule within the limited decoding delay time. Hardware implementation on an FPGA and simulation results show that the proposed partially-parallel LDPC decoder improves the decoding throughput and bit error performance with a small hardware overhead.

This paper presents a test input data compression technique, Selective Low-Care Coding (SLC), which can be used to significantly reduce input test data volume as well as the external test channel requirement for multiscan-based designs. In the proposed SLC scheme, we explored the linear dependencies of the internal scan chains, and instead of encoding all the specified bits in test cubes, only a smaller amount of specified bits are selected for encoding, thus greater compression can be expected. Experiments on the larger benchmark circuits show drastic reduction in test data volume with corresponding savings on test application time can be indeed achieved even for the well-compacted test set.

In this paper, we propose a partially-parallel LDPC decoder which achieves a high-efficiency message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the column operation module to compute every message in each bit node which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay when the row and column operations are performed concurrently. Therefore, the proposed decoder performs the column operations more frequently in a single iterative decoding, and achieves a high-efficiency message-passing schedule within the limited decoding delay time. Hardware implementation on an FPGA and simulation results show that the proposed partially-parallel LDPC decoder improves the decoding throughput and bit error performance with a small hardware overhead.

Elliptic curve cryptosystems are expected to be a next standard of public-key cryptosystems. A security level of elliptic curve cryptosystems depends on a difficulty of a discrete logarithm problem on elliptic curves. The security level of a elliptic curve cryptosystem which has a public-key of 160-bit is equivalent to that of a RSA system which has a public-key of 1024-bit. We propose an elliptic curve cryptosystem LSI architecture embedding word-based Montgomery multipliers. A Montgomery multiplication is an efficient method for a finite field multiplication. We can design a scalable architecture for an elliptic curve cryptosystem by selecting structure of word-based Montgomery multipliers. Experimental results demonstrate effectiveness and efficiency of the proposed architecture. In the hardware evaluation using 0.18 mu m CMOS library, the highspeed design using 126 Kgates with 20 x 8-bit multipliers achieved operation times of 3.6 ms for a 160-bit point multiplication.

Elliptic curve cryptosystems are expected to be a next standard of public-key cryptosystems. A security level of elliptic curve cryptosystems depends on a difficulty of a discrete logarithm problem on elliptic curves. The security level of a elliptic curve cryptosystem which has a public-key of 160-bit is equivalent to that of a RSA system which has a public-key of 1024-bit. We propose an elliptic curve cryptosystem LSI architecture embedding word-based Montgomery multipliers. A Montgomery multiplication is an efficient method for a finite field multiplication. We can design a scalable architecture for an elliptic curve cryptosystem by selecting structure of word-based Montgomery multipliers. Experimental results demonstrate effectiveness and efficiency of the proposed architecture. In the hardware evaluation using 0.18 mu m CMOS library, the highspeed design using 126 Kgates with 20 x 8-bit multipliers achieved operation times of 3.6 ms for a 160-bit point multiplication.

本稿では,アプリケーションに特化したSIMD型プロセッサコアの自動合成において,合成されるプロセッサコアのパイプライン構成を最適化する手法を提案する.画像処理アプリケーションなどで頻繁に処理されるSIMD乗算や乗加算は演算実行遅延が大きくプロセッサの動作周波数を決定することが多い.SIMD演算を実行可能なSIMD型プロセッサコアの自動合成では,SIMD演算をマルチサイクル化することにより,少ない面積オーバーヘッドでより高速なプロセッサコア構成を得ることが期待できる.但し,ハザードの影響により必ずしもパイプライン段数を増やすことがアプリケーション実行時間の短縮につながるとは限らず,プロセッサコアの自動合成では与えられた実行時間制約の中で最適なパイプライン構成を探索するのが望ましい.提案手法では,まず段数の異なる複数のパイプライン構成を定義する.次に定義された全てのパイプライン構成について,与えられた実行時間制約を満たすように,付加するハードウェアユニットの種類と数を最適化する.最後に得られた全てのパイプライン構成の中で最も小面積の構成を選択し,最適なパイプライン構成として出力する.計算機実験により本手法の有効性を評価した.

本稿ではシステムLSI設計のためのHW/SW分割システムを提案する.本システムは,制約に応じて列挙するIPの数と種類を変えるデータベースと,制約を満たす解がない場合に改良に取り掛かる設計者を対話により補助する設計ナビゲーションを持つ.データベースのIPを再利用することにより新規設計部分を減らし開発期間の短縮が可能となり,列挙するIPを削減することで探索時間も短縮できる.また,設計ナビゲーションによりボトルネックや改良が容易なアーキテクチャを短時間で検討できる.これらの有効性を計算機実験により確認した.

Elliptic curve cryptosystems are expected to be a next standard of public-key cryptosystems. A security level of elliptic curve cryptosystems depends on a difficulty of a discrete logarithm problem on elliptic curves. The security level of a elliptic curve cryptosystem which has a public-key of 160-bit is equivalent to that of a RSA system which has a public-key of 1024-bit. We propose an elliptic curve cryptosystem LSI architecture embedding word-based Montgomery multipliers. A Montgomery multiplication is an efficient method for a finite field multiplication. We can design a scalable architecture for an elliptic curve cryptosystem by selecting structure of word-based Montgomery multipliers. Experimental results demonstrate effectiveness and efficiency of the proposed architecture. In the hardware evaluation using 0.18 mu m CMOS library, the highspeed design using 126 Kgates with 20 x 8-bit multipliers achieved operation times of 3.6 ms for a 160-bit point multiplication.

This paper proposes a parallel LSI architecture for LDPC decoder which improves a message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the decoder to perform every column operation using the messages which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay. Hardware imp mentation and simulation results show that the proposed decoder improves the decoding throughput and bit error performance with a small hardware overhead.

This paper proposes a new multiscan-based test input data compression technique by employing a Fan-out Compression Scan Architecture (FCSCAN) for test cost reduction. The basic idea of FCSCAN is to target the minority specified 1 or 0 bits (either 1 or 0) in scan slices for compression. Due to the low specified bit density in test cube set, FCSCAN can significantly reduce input test data volume and the number of required test channels so as to reduce test cost. The FCSCAN technique is easy to be implemented with small hardware overhead and does not need any special ATPG for test generation. In addition, based on the theoretical compression efficiency analysis, improved procedures are also proposed for the FCSCAN to achieve further compression. Experimental results on both benchmark circuits and one real industrial design indicate that drastic reduction in test cost can be indeed achieved.

In SoC designs, efficient communication between the hardware IPs and the on-chip processor becomes very important, however the interface is usually affacted by the processor core specification. Thus in this paper, we focus on developing an efficient interface circuit architecture for the communications between the on-chip processor and embedded hardware IP cores. we also propose a method to synthesize it. Experimental results show that our method could obtain optimal interface circuits and works well through designing a MPEG-4 encode application.

This paper proposes a memory-efficient accelerating schedule for LDPC decoder. Important properties of the proposed techniques are as follows: (i) Partitioning a pipelined operation not to read and write intermediate messages simultaneously enables the accelerated message-passing schedule to be implemented with single-port memories. (ii) FIFO-based buffering reduces the number of memory banks and words for the decoder based on the accelerated message-passing schedule. The proposed decoder reduces the memories for intermediate messages by half compared to the conventional one based on the accelerated message-passing schedule.

This paper proposes a parallel LSI architecture for LDPC decoder which improves a message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the decoder to perform every column operation using the messages which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay. Hardware imp mentation and simulation results show that the proposed decoder improves the decoding throughput and bit error performance with a small hardware overhead.

This paper presents a test input data compression technique, Selective Low-Care Coding (SLC), which can he used to significantly reduce input test data volume as well as the external test channel requirement for multiscan-based designs. In the proposed SLC scheme, we explored the linear dependencies of the internal scan chains, and instead of encoding all the specified bits in test cubes, only a smaller amount of specified bits are selected for encoding, thus greater compression can be expected. Experiments on the larger benchmark circuits show drastic reduction in test data volume with corresponding savings on test application time can be indeed achieved even for the well-compacted test set. Copyright © 2006 The Institute of Electronics, Information and Communication Engineers.

This paper proposes a new multiscan-based test input data compression technique by employing a Fan-out Compression Scan Architecture (FCSCAN) for test cost reduction. The basic idea of FCSCAN is to target the minority specified 1 or 0 bits (either 1 or 0) in scan slices for compression. Due to the low specified bit density in test cube set, FCSCAN can significantly reduce input test data volume and the number of required test channels so as to reduce test cost. The FCSCAN technique is easy to be implemented with small hardware overhead and does not need any special ATPG for test generation. In addition, based on the theoretical compression efficiency analysis, improved procedures are also proposed for the FCSCAN to achieve further compression. Experimental results on both benchmark circuits and one real industrial design indicate that drastic reduction in test cost can be indeed achieved.

In SoC designs, efficient communication between the hardware IPs and the on-chip processor becomes very important, however the interface is usually affacted by the processor core specification. Thus in this paper, we focus on developing an efficient interface circuit architecture for the communications between the on-chip processor and embedded hardware IP cores. we also propose a method to synthesize it. Experimental results show that our method could obtain optimal interface circuits and works well through designing a MPEG-4 encode application.

H.264/AVCは高い符号化効率を実現する反面,符号化に必要な処理量が多い問題点があり,その90%以上は動き予測処理が占めている.符号化効率を向上させるために導入された複数参照フレーム,可変ブロックサイズ,1/4画素精度動き補償がその主因である.これを高速化させるため,複数参照フレーム,可変ブロックサイズに対応した整数精度動き予測処理アーキテクチャが提案されている.しかし,これらのアーキテクチャは探索場所の移動において変則的なメモリアクセスを要し,メモリバンド幅が制限されるDSP組み込み等の用途では性能向上が難しい.本稿では,この問題に対応するため,画素サブサンプリング手法を用いたDSP組み込み整数精度動き予測処理アーキテクチャを提案する.画素サブサンプリングは演算に用いる画素を間引くことにより,一般的にハードウェア量削減に用いられる.提案アーキテクチャではサブサンプリングパターンを一般的なチェスボード状から縦縞状に変更することにより,演算器のデータ読み込みサイクルを削減し動き予測処理の高速化を可能とする.提案するアーキテクチャは200MHzで動作させた場合,CIF画像の予測処理を86.5fpsで実行可能である.

This paper proposes a parallel LSI architecture for LDPC decoder which improves a message-passing schedule. The proposed LDPC decoder is characterized as follows: (i) The column operations follow the row operations in a pipelined architecture to ensure that the row and column operations are performed concurrently. (ii) The proposed parallel pipelined bit functional unit enables the decoder to perform every column operation using the messages which is updated by the row operations. These column operations can be performed without extending the single iterative decoding delay. Hardware imp mentation and simulation results show that the proposed decoder improves the decoding throughput and bit error performance with a small hardware overhead.

DSM(Deep SubMicron technology)時代では高位設計の際, フロアプランや配線抵抗などを考慮する必要が出でくる.また, 高位設計で繰り返し行われるグローバル配線遅延の見積もりの際, インダクタンスの影響が無視できない.本稿ではインダクタンスを考慮してグローバル配線遅延を見積もる方法について述べる.本稿ではドライバ-RLC配線-負荷モデルのステップ応答のが50%に達するまでの時間(50%遅延)を見積もる.提案する見積もり式は, あらかじめ素子値を説明変数として重回帰分析により得られた1次式を用いる.本手法は遅延の内, time of flightが支配的な場合に適用可能で, SPICEで計算した値との誤差を最大約15%, 平均約2.5%で見積もることができる.

近年のLSI設計プロセスの微細化に伴い, 配線による遅延の割合がゲート遅延に対し相対的に増加してきおり, 高位合成の段階においてもフロアプランを考慮する必要がある.レジスタ分散型アーキテクチャを用いると, レジスタ間データ転送を利用することにより配線遅延が回路の性能に与える影響を削除することが可能であるが, レジスタ数が増大し面積増加を招いてしまうという問題点が生じる.本稿では, レジスタ分散型とレジスタ共有型を併用するレジスタ分散・共有型を対象とし, (1)スケジューリング, (2)レジスタアロケーション, (3)レジスタバインディング, (4)モジュール配置の工程を繰り返し(4)から得られたフロアプラン情報をフィードバックすることにより, 解を収束させる高位合成手法を提案する.この手法はレジスタ分散型アーキテクチャと同等の回路の性能を維持しながら面積を削減することが可能となる.また, 計算機実験によって, 提案手法の有効性を示す.

This paper proposes a reconfigurable adaptive FEC system based on Reed-Solomon (RS) code with interleaving. In adaptive FEC schemes, error correction capability t is changed dynamically according to the communication channel condition. For given error correction capability t, we can implement an optimal RS decoder composed of minimum hardware units for each t. If the hardware units of the RS decoder can be reduced for any given error correction capability t, we can embed as large deinterleaver as possible into the RS decoder for each.t. Reconfiguring the RS decoder embedded with the expanded deinterleaver dynamically for each error correction capability t allows us to decode larger interleaved codes which are more robust error correction codes to burst errors. In a reliable transport protocol, experimental results show that our system achieves up to 65% lower packet error rate and 5.9% higher data transmission throughput compared to the adaptive FEC scheme on a conventional fixed hardware system. In an unreliable transport protocol, our system achieves up to 76% better bit error performance with higher code rate compared to the adaptive FEC scheme on a conventional fixed hardware system.

This paper focuses on SIMD processor synthesis and proposes a SIMD instruction set/functional unit synthesis algorithm. Given an initial assembly code and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with optimal SIMD functional units. It also synthesizes a SIMD instruction set. The input initial assembly code is assumed to run on a full-resource SIMD processor (virtual processor) which has all the possible SIMD functional units. In our algorithm, we introduce the SIMD operation decomposition and apply it to the initial assembly code and the full-resource SIMD processor. By gradually reducing SIMD operations or decomposing SIMD operations, we can finally find a processor core with small area under the given timing constraint. The promising experimental results are also shown.

This paper proposes a reconfigurable adaptive FEC system based on Reed-Solomon (RS) code with interleaving. In adaptive FEC schemes, error correction capability t is changed dynamically according to the communication channel condition. For given error correction capability t, we can implement an optimal RS decoder composed of minimum hardware units for each t. If the hardware units of the RS decoder can be reduced for any given error correction capability t, we can embed as large deinterleaver as possible into the RS decoder for each.t. Reconfiguring the RS decoder embedded with the expanded deinterleaver dynamically for each error correction capability t allows us to decode larger interleaved codes which are more robust error correction codes to burst errors. In a reliable transport protocol, experimental results show that our system achieves up to 65% lower packet error rate and 5.9% higher data transmission throughput compared to the adaptive FEC scheme on a conventional fixed hardware system. In an unreliable transport protocol, our system achieves up to 76% better bit error performance with higher code rate compared to the adaptive FEC scheme on a conventional fixed hardware system.

This paper focuses on SIMD processor synthesis and proposes a SIMD instruction set/functional unit synthesis algorithm. Given an initial assembly code and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with optimal SIMD functional units. It also synthesizes a SIMD instruction set. The input initial assembly code is assumed to run on a full-resource SIMD processor (virtual processor) which has all the possible SIMD functional units. In our algorithm, we introduce the SIMD operation decomposition and apply it to the initial assembly code and the full-resource SIMD processor. By gradually reducing SIMD operations or decomposing SIMD operations, we can finally find a processor core with small area under the given timing constraint. The promising experimental results are also shown.

A b-bit SIMD functional unit has n k-bit sub-functional units in itself, where b = k x n. It can execute n-parallel k-bit operations. However, all the b-bit functional units in a processor core do not necessarily execute n-parallel operations. Depending on an application program, some of them just execute n/2-parallel operations or even n/4-parallel operations. This means that we can modify a b-bit SIMD functional unit so that it has n/2 k-bit sub-functional units or n/4 k-bit sub-functional units. The number of k-bit sub-functional units in a SIMD functional unit is called sub-operation parallelism. We incorporate a sub-operation parallelism optimization algorithm into SIMD functional unit optimization. Our proposed algorithm gradually reduces sub-operation parallelism of a SIMD functional unit while the timing constraint of execution time satisfied. Thereby, we can finally find a processor core with small area under the given timing constraint. We expect that we can obtain processor core configurations of smaller area in the same timing constraint rather than a conventional system. The promising experimental results are also shown.

本稿では, 動的フローに適応したネットワークプロセッサを提案する.提案ネットワークプロセッサは, ネットワークプロセッサ内の各処理ユニットの中から, ボトルネック処理をコマンド待ちキュー情報・各処理ユニットに必要とされるサイクルカウントを元に判断し, 全体スループットを改善する為に次に行うべき処理を採択し, それを実行するハードウェア機構を備えている.この処理状態管理ハードウェアとネットワークプロセッサ内の各処理ユニットに必要な命令・データパスを備えたDynamic-Micro Packet Processorが, 随時ボトルネック処理の並列化を行うことで, ネットワークプロセッサ全体のスループット向上を実現する.提案アーキテクチャをHW記述言語であるVHDLを用いて実装し, 計算機実験を行うことで, その有効性を評価する.

次世代の公開鍵暗号技術として楕円曲線暗号が注目されている.楕円曲線暗号の安全性は楕円曲線上の離散対数問題に依っている.このため, 1024ビットのRSA暗号と同程度の安全性を160ビットの暗号鍵で実現できる.本稿では, ワードベースモンゴメリ乗算器を搭載した楕円曲線暗号LSIを提案する.モンゴメリ乗算は有限体上の乗算を実現するアルゴリズムであり, ワードベースモンゴメリ乗算器の構成を変えることで, 小面積で高速な楕円曲線暗号LSIを実現することができる.計算機実験により提案する楕円曲線暗号LSIを評価する.評価結果において, 0.18μmプロセスを用い8ビット乗算器を搭載した楕円曲線暗号LSIは126Kgatesで160ビットのスカラー倍算を実行した場合の処理時間は3.6msであった.

In this paper, we propose a sub-operation parallelism optimization algorithm in SIMD) processor synthesis. Given an initial assembly code and timing constraints, our algorithm synthesizes a processor core with sub-operation parallelism optimization for SIMD) functional units. First we consider an initial processor which has sufficient hardware units for executing an initial assembly code. An initial processor core includes the maximum sub-operation parallelism for each SIMD) functional unit. By gradually reducing sub-operation parallelism, we can finally have a processor core with small area meeting a given timing constraints. We show the effectiveness of our proposed algorithm through experimental results.

This paper proposes a partially-parallel LDPC decoder based on a high-efficiency message-passing algorithm. Our proposed partially-parallel LDPC decoder performs the column operations for bit nodes in conjunction with the row operations for check nodes. Bit functional unit with pipeline architecture in our LDPC decoder allows us to perform column operations for every bit node connected to each of check nodes which are updated by the row operations in parallel. Our proposed LDPC decoder improves the timing when the column operations are performed, accordingly it improves the message-passing efficiency within the limited number of iterations for decoding. We implemented the proposed partially-parallel LDPC decoder on an FPGA, and simulated its decoding performance. Practical simulation shows that our proposed LDPC decoder reduces the number of iterations for decoding, and it improves the bit error performance with a small hardware overhead.

This paper presents a new DFT technique that can significantly reduce test data volume as well as scan-in power consumption for multiscan-based designs. It can also help to reduce test time and tester channel requirements with small hardware overhead In the proposed approach, we start with a pre-computed test cube set and fill the don't-cares with proper values for joint reduction of test data volume and scan power consumption. In addition we explore the linear dependencies of the scan chains to construct a fanout structure only with inverters to achieve further compression. Experimental results for the larger ISCAS'89 benchmarks show the efficiency of the proposed technique.

This paper proposes a reconfigurable adaptive FEC system with interleaving. For adaptive FEC schemes, we can implement an optimal RS decoder composed of minimum hardware units for any given error correction capability t. If the hardware units of the RS decoder can be reduced for any given t, we can embed as large deinterleaver as possible into the RS decoder for each t. Reconfiguring the RS decoder embedded with the expanded deinterleaver dynamically for each t allows us to decode larger interleaved codes which are more robust FEC codes to burst errors. Our reconfigurable adaptive FEC system with interleaving achieves better packet error rate and higher throughput than fixed hardware systems.

This paper proposes a new design methodology for SoCs reusing hardware IPs. In our approach, after system-level HW/SW partitioning, we use IPs for hardware parts, but synthesize a new processor core instead of reusing a processor core IP. System performs efficient parallel execution of hardware and software by taking account of a response time of hardware IP obtained by the proposed calculation algorithm. We can use optimal hardware IPs selected by the proposed hardware IPs selection algorithm. The experimental results show effectiveness of our new design methodology.

In this paper, we propose a sub-operation parallelism optimization algorithm in SIMD) processor synthesis. Given an initial assembly code and timing constraints, our algorithm synthesizes a processor core with sub-operation parallelism optimization for SIMD) functional units. First we consider an initial processor which has sufficient hardware units for executing an initial assembly code. An initial processor core includes the maximum sub-operation parallelism for each SIMD) functional unit. By gradually reducing sub-operation parallelism, we can finally have a processor core with small area meeting a given timing constraints. We show the effectiveness of our proposed algorithm through experimental results.

This paper proposes a partially-parallel LDPC decoder based on a high-efficiency message-passing algorithm. Our proposed partially-parallel LDPC decoder performs the column operations for bit nodes in conjunction with the row operations for check nodes. Bit functional unit with pipeline architecture in our LDPC decoder allows us to perform column operations for every bit node connected to each of check nodes which are updated by the row operations in parallel. Our proposed LDPC decoder improves the timing when the column operations are performed, accordingly it improves the message-passing efficiency within the limited number of iterations for decoding. We implemented the proposed partially-parallel LDPC decoder on an FPGA, and simulated its decoding performance. Practical simulation shows that our proposed LDPC decoder reduces the number of iterations for decoding, and it improves the bit error performance with a small hardware overhead.

A b-bit SIMD functional unit has n k-bit sub-functional units in itself, where b = k × n. It can execute n-parallel k-bit operations. However, all the b-bit functional units in a processor core do not necessarily execute n-parallel operations. Depending on an application program, some of them just execute n/2-parallel operations or even n/4-parallel operations. This means that we can modify a b-bit SIMD functional unit so that it has n/2 k-bit sub-functional units or n/4 k-bit sub-functional units. The number of k-bit sub-functional units in a SIMD functional unit is called sub-operation parallelism. We incorporate a sub-operation parallelism optimization algorithm into SIMD functional unit optimization. Our proposed algorithm gradually reduces sub-operation parallelism of a SIMD functional unit while the timing constraint of execution time satisfied. Thereby, we can finally find a processor core with small area under the given timing constraint. We expect that we can obtain processor core configurations of smaller area in the same timing constraint rather than a conventional system. The promising experimental results are also shown. Copyright © 2005 The Institute of Electronics, Information and Communication Engineers.

This paper proposes a new design methodology for SoCs reusing hardware IPs. In our approach, after system-level HW/SW partitioning, we use IPs for hardware parts, but synthesize a new processor core instead of reusing a processor core IP. System performs efficient parallel execution of hardware and software by taking account of a response time of hardware IP obtained by the proposed calculation algorithm. We can use optimal hardware IPs selected by the proposed hardware IPs selection algorithm. The experimental results show effectiveness of our new design methodology.

In this paper, we propose a sub-operation parallelism optimization algorithm in SIMD processor synthesis. Given an initial assembly code and timing constraints, our algorithm synthesizes a processor core with sub-operation parallelism optimization for SIMD functional units. First we consider an initial processor which has sufficient hardware units for executing an initial assembly code. An initial processor core includes the maximum sub-operation parallelism for each SIMD functional unit. By gradually reducing sub-operation parallelism, we can finally have a processor core with small area meeting a given timing constraints. We show the effectiveness of our proposed algorithm through experimental results. © 2005 IEEE.

This paper proposes a partially-parallel LDPC decoder based on a high-efficiency message-passing algorithm. Our proposed partially-parallel LDPC decoder performs the column operations for bit nodes in conjunction with the row operations for check nodes. Bit functional unit with pipeline architecture in our LDPC decoder allows us to perform column operations for every bit node connected to each of check nodes which are updated by the row operations in parallel. Our proposed LDPC decoder improves the timing when the column operations are performed, accordingly it improves the message-passing efficiency within the limited number of iterations for decoding. We implemented the proposed partially-parallel LDPC decoder on an FPGA, and simulated its decoding performance. Practical simulation shows that our proposed LDPC decoder reduces the number of iterations for decoding, and it improves the bit error performance with a small hardware overhead.

A b-bit SIMD functional unit has n k-bit sub-functional units in itself, where b = k × n. It can execute n-parallel k-bit operations. However, all the b-bit functional units in a processor core do not necessarily execute n-parallel operations. Depending on an application program, some of them just execute n/2-parallel operations or even n/4-parallel operations. This means that we can modify a b-bit SIMD functional unit so that it has n/2 k-bit sub-functional units or n/4 k-bit sub-functional units. The number of k-bit sub-functional units in a SIMD functional unit is called sub-operation parallelism. We incorporate a sub-operation parallelism optimization algorithm into SIMD functional unit optimization. Our proposed algorithm gradually reduces sub-operation parallelism of a SIMD functional unit while the timing constraint of execution time satisfied. Thereby, we can finally find a processor core with small area under the given timing constraint. We expect that we can obtain processor core configurations of smaller area in the same timing constraint rather than a conventional system. The promising experimental results are also shown. Copyright © 2005 The Institute of Electronics, Information and Communication Engineers.

In this paper, we propose a reconfigurable adaptive FEC system. In adaptive FEC schemes, the error correction capability t is changed dynamically according to the communication channel condition. If a particular error correction capability t is given, we can implement an FEC decoder which is optimal for t by taking the number of operations into consideration. Thus, reconfiguring the optimal FEC decoder dynamically for each error correction capability allows us to maximize the throughput of each decoder within a limited hardware resource. Based on this concept, our reconfigurable adaptive FEC system can reduce the packet dropping rate more efficiently than conventional fixed hardware systems. We can improve data transmission throughput for a reliable transport protocol. Practical simulation results are also shown.

This paper proposes a thread partitioning algorithm in low power high-level synthesis. The algorithm is applied to high-level synthesis systems. In the systems, we can describe parallel behaving circuit blocks (threads) explicitly. First it focuses on a local register file RF in a thread. It partitions a thread into two sub-threads, one of which has RF and the other does not have RE The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. Then we can synthesize a low power circuit with a low area overhead, compared to the original circuit. Experimental results demonstrate effectiveness and efficiency of the algorithm.

In this paper, we propose a reconfigurable adaptive FEC system. In adaptive FEC schemes, the error correction capability t is changed dynamically according to the communication channel condition. If a particular error correction capability t is given, we can implement an FEC decoder which is optimal for t by taking the number of operations into consideration. Thus, reconfiguring the optimal FEC decoder dynamically for each error correction capability allows us to maximize the throughput of each decoder within a limited hardware resource. Based on this concept, our reconfigurable adaptive FEC system can reduce the packet dropping rate more efficiently than conventional fixed hardware systems. We can improve data transmission throughput for a reliable transport protocol. Practical simulation results are also shown.

This paper proposes a thread partitioning algorithm in low power high-level synthesis. The algorithm is applied to high-level synthesis systems. In the systems, we can describe parallel behaving circuit blocks (threads) explicitly. First it focuses on a local register file RF in a thread. It partitions a thread into two sub-threads, one of which has RF and the other does not have RE The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. Then we can synthesize a low power circuit with a low area overhead, compared to the original circuit. Experimental results demonstrate effectiveness and efficiency of the algorithm.

This paper proposes a hardware/software cosynthesis algorithm for processors with heterogeneous registers. Given a CDFG corresponding to an application program and a timing constraint, the algorithm generates a processor configuration minimizing area of the processor and an assembly code on the processor. First, the algorithm configures a datapath which can execute several DFG nodes with data dependency at one cycle. The datapath can execute the application program at the least number of cycles. The branch and bound algorithm is applied and all the number of functional units and memory banks are tried. For an assumed number of functional units and memory banks, an appropriate number of heterogeneous registers and connections to functional units and registers are explored. The experimental results show effectiveness and efficiency of the algorithm.

This paper proposes a hardware/software cosynthesis algorithm for processors with heterogeneous registers. Given a CDFG corresponding to an application program and a timing constraint, the algorithm generates a processor configuration minimizing area of the processor and an assembly code on the processor. First, the algorithm configures a datapath which can execute several DFG nodes with data dependency at one cycle. The datapath can execute the application program at the least number of cycles. The branch and bound algorithm is applied and all the number of functional units and memory banks are tried. For an assumed number of functional units and memory banks, an appropriate number of heterogeneous registers and connections to functional units and registers are explored. The experimental results show effectiveness and efficiency of the algorithm.

本稿では,既存のIPを元に「処理の並列度」や「処理の反復性」に着目し,新たなIPを類推する手法を提案する.提案手法は,MPEG-4,JPEG,JPEG2000などの画像処理アプリケーションを対象とする.上記のアプリケーションを構成する処理(DCT,量子化など)を解析し,それぞれの処理に対して「処理の並列度に着目した類推手法」または「処理の反復性に着目した類推手法」を適用する.提案手法を適用することにより,解の探索空間が広がる.その結果,既存のIPのみから構成される解と比較して,同一の処理時間かつ小面積の解を得ることが可能になる.提案手法を計算機上に実装し,その有効性を報告する.

本稿ではインターリーブを考慮したReconfigurable Adaptive FEC を提案する.誤り訂正処理回路のハードウェア構成を訂正能力について動的に最適化することにより,ハードウェアで考慮できるインターリーブの深さを訂正能力ごとに与えることができる.これにより従来のASIC等で実現する固定的なハードウェアに比べて,よりバースト誤りに強い符号を生成することが可能となる.誤り訂正処理における復号器の最適化とインターリーブを考慮したハードウェア構成を提案し,本ハードウェアによる有効性を計算機シミュレーションにより評価する.

本稿では，携帯機器を対象としたJava 動的コンパイラにおけるプロファイリングシステムを提案する．本システムはJava 仮想マシンの実行中に，アプリケーションにおいて頻繁に実行されるメソッド(ホットメソッド) を検出する．ホットメソッドはコンパイラによってネイティブコードにコンパイルされ，ヒープ領域に格納される．ネイティブコードに使用するヒープ領域をプロファイラが決定することにより，ガーベッジコレクションの起動を抑制することが可能である．提案プロファイラはJava仮想マシンの処理時間の3%程度のオーバーヘッドで動作可能である．ガーベッジコレクションを元の仮想マシンの2 倍程度に抑え，プロファイリングによって得られたメソッドをコンパイルすることにより，平均で約7 倍程度の高速化を実現した．This paper proposes a lightweight profiling system of Java dynamic compiler for handheld devices. The system detects the methods frequently invoked in application (hot method) during execution of a Java virtual machine. A hot method is compiled into a native code by the compiler, and is stored in a heap area. The profiler determines the heap area used for a native code, and it is possible to reduce a garbage collection. Our technique can profiles method informations with 3% overhead of the processing time of a Java virtual machine. By compiling the hot method, as a result, we achieve approximately 7 times speedup in average, by suppressing a garbage collections to approximately 2 times of the original virtual machine.

Test data volume and scan power are two Major concerns in SoC test. In this paper we present an alternative run-length coding method through scan chain reconfiguration to reduce both test data volume and scan-in power consumption. The proposed method analyzes the compatibility of the internal scan cells for a given test set and then divides the scan cells into compatible classes. To extract the compatible scan cells we apply a heuristic algorithm by solving the graph coloring problem; and then a simple greedy algorithm is used to configure the scan chain for the minimization of scan power Experimental results for the larger ISCAS'89 benchmarks show that the proposed approach leads to highly reduced test data volume with significant power savings during scan test.

This paper focuses on SIMD processor synthesis and proposes a SIMD instruction set/functional unit synthesis algorithm. Given an initial assembly code and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with optimal SIMD functional units. It also synthesizes a SIMD instruction set. The input initial assembly code is assumed to run on a full-resource SIMD processor (virtual processor) which has all the possible SIMD functional units. In our algorithm, we introduce the SIMD operation decomposition and apply it to the initial assembly code and the full-resource SIMD processor. By gradually reducing SIMD operations or decomposing SIMD operations, we can finally find a processor core with small area under the given timing constraint. The promising experimental results are also shown.

This paper proposes a hardware/software cosynthesis algorithm for processors with heterogeneous registers. Given a CDFG corresponding to an application program and a timing constraint, the algorithm generates a processor configuration minimizing area of the processor and an assembly code on the processor. First, the algorithm configures a datapath which can execute several DFG nodes with data dependency at one cycle. The datapath can execute the application program at the least number of cycles. The branch and bound algorithm is applied and all the number of functional units and memory banks are tried. For an assumed number of functional units and memory banks, an appropriate number of heterogeneous registers and connections to functional units and registers are explored. The experimental results show effectiveness and efficiency of the algorithm.

This paper proposes a thread partitioning algorithm in low power high-level synthesis. The algorithm is applied to high-level synthesis systems. In the systems, we can describe parallel behaving circuit blocks(threads) explicitly. First it focuses on a local register file RF in a thread. It partitions a thread into two sub-threads, one of which has RF and the other does not have, RF. The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. Then we can synthesize a low power circuit with a low area overhead, compared to the original circuit. Experimental results demonstrate effectiveness and efficiency of the algorithm.

This paper proposes a reconfigurable adaptive FEC system. For adaptive FEC schemes, we can implement an FEC decoder which is optimal for error correction capability t by taking the number of operations into consideration. Reconfiguring the optimal FEC decoder dynamically for each t allows us to maximize the throughput of each decoder within a limited hardware resource. Our system can reduce packet dropping rate more efficiently than conventional fixed hardware systems for a reliable transport protocol.

This paper presents a thread partitioning algorithm for high-level synthesis systems which generate low energy circuits. In the algorithm, we partitions a thread into two sub-threads, one of which has RF and the other does not have RE The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. We achieve 33% energy reduction when we apply our proposed algorithm to a JPEG encoder.

Test data volume and scan power are two Major concerns in SoC test. In this paper we present an alternative run-length coding method through scan chain reconfiguration to reduce both test data volume and scan-in power consumption. The proposed method analyzes the compatibility of the internal scan cells for a given test set and then divides the scan cells into compatible classes. To extract the compatible scan cells we apply a heuristic algorithm by solving the graph coloring problem; and then a simple greedy algorithm is used to configure the scan chain for the minimization of scan power Experimental results for the larger ISCAS'89 benchmarks show that the proposed approach leads to highly reduced test data volume with significant power savings during scan test.

We describe the optimization of a complex video encoder systems based on target architecture. We implemented the MPEG-4 encoder using hardware/software codesign approach, mapped together based on a target architecture. We proposed a target architecture template and an optimization methodology. In our design flow, we searched for a bottleneck module constraining the system. After investigating the computational complexity, quality, and the simplicity of algorithms, we chose the best algorithm for hardware implementation, and then mapped the selected algorithm onto the hardware with different architecture, what does the best architecture for the algorithm and which is the best architecture of components. We chose one of the architectures meet the constraints and also made tradeoffs among speed, chip area, and memory bandwidth for different architecture. The proposed system architecture was used to reduce the design decisions and iterations, provided flexible and scalable systems. The evaluations resulted in effective optimization of the motion estimation module and better tradeoffs that optimized the overall system.

This paper presents a new test data compression technique for multiscan-based designs through dictionary-based encoding on the single or sequences scan-inputs. In spite of its simplicity, it achieves significant reduction in test data volume. Unlike some previous approaches on test data compression, our approach eliminates the need for additional synchronization and handshaking between the CUT and the ATE, so it is especially suitable to be integrated in a low cost test scheme for SoC test In addition in contrast to previous dictionary-based coding techniques, even for the CUT with a small number of scan chains, the proposed approach can achieve satisfied reduction in test data volume. Experimental results showed the proposed test scheme works particularly well for the large ISCAS'89 benchmarks.

This paper focuses on SIMD processor synthesis and proposes a SIMD instruction set/functional unit synthesis algorithm. Given an initial assembly code and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with optimal SIMD functional units. It also synthesizes a SIMD instruction set. The input initial assembly code is assumed to run on a full-resource SIMD processor (virtual processor) which has all the possible SIMD functional units. In our algorithm, we introduce the SIMD operation decomposition and apply it to the initial assembly code and the full-resource SIMD processor. By gradually reducing SIMD operations or decomposing SIMD operations, we can finally find a processor core with small area under the given timing constraint. The promising experimental results are also shown.

This paper proposes a hardware/software cosynthesis algorithm for processors with heterogeneous registers. Given a CDFG corresponding to an application program and a timing constraint, the algorithm generates a processor configuration minimizing area of the processor and an assembly code on the processor. First, the algorithm configures a datapath which can execute several DFG nodes with data dependency at one cycle. The datapath can execute the application program at the least number of cycles. The branch and bound algorithm is applied and all the number of functional units and memory banks are tried. For an assumed number of functional units and memory banks, an appropriate number of heterogeneous registers and connections to functional units and registers are explored. The experimental results show effectiveness and efficiency of the algorithm.

This paper proposes a thread partitioning algorithm in low power high-level synthesis. The algorithm is applied to high-level synthesis systems. In the systems, we can describe parallel behaving circuit blocks(threads) explicitly. First it focuses on a local register file RF in a thread. It partitions a thread into two sub-threads, one of which has RF and the other does not have, RF. The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. Then we can synthesize a low power circuit with a low area overhead, compared to the original circuit. Experimental results demonstrate effectiveness and efficiency of the algorithm.

This paper proposes a reconfigurable adaptive FEC system. For adaptive FEC schemes, we can implement an FEC decoder which is optimal for error correction capability t by taking the number of operations into consideration. Reconfiguring the optimal FEC decoder dynamically for each t allows us to maximize the throughput of each decoder within a limited hardware resource. Our system can reduce packet dropping rate more efficiently than conventional fixed hardware systems for a reliable transport protocol.

This paper presents a thread partitioning algorithm for high-level synthesis systems which generate low energy circuits. In the algorithm, we partitions a thread into two sub-threads, one of which has RF and the other does not have RE The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. We achieve 33% energy reduction when we apply our proposed algorithm to a JPEG encoder.

Test data volume and scan power are two Major concerns in SoC test. In this paper we present an alternative run-length coding method through scan chain reconfiguration to reduce both test data volume and scan-in power consumption. The proposed method analyzes the compatibility of the internal scan cells for a given test set and then divides the scan cells into compatible classes. To extract the compatible scan cells we apply a heuristic algorithm by solving the graph coloring problem; and then a simple greedy algorithm is used to configure the scan chain for the minimization of scan power Experimental results for the larger ISCAS'89 benchmarks show that the proposed approach leads to highly reduced test data volume with significant power savings during scan test.

We describe the optimization of a complex video encoder systems based on target architecture. We implemented the MPEG-4 encoder using hardware/software codesign approach, mapped together based on a target architecture. We proposed a target architecture template and an optimization methodology. In our design flow, we searched for a bottleneck module constraining the system. After investigating the computational complexity, quality, and the simplicity of algorithms, we chose the best algorithm for hardware implementation, and then mapped the selected algorithm onto the hardware with different architecture, what does the best architecture for the algorithm and which is the best architecture of components. We chose one of the architectures meet the constraints and also made tradeoffs among speed, chip area, and memory bandwidth for different architecture. The proposed system architecture was used to reduce the design decisions and iterations, provided flexible and scalable systems. The evaluations resulted in effective optimization of the motion estimation module and better tradeoffs that optimized the overall system.

This paper presents a new test data compression technique for multiscan-based designs through dictionary-based encoding on the single or sequences scan-inputs. In spite of its simplicity, it achieves significant reduction in test data volume. Unlike some previous approaches on test data compression, our approach eliminates the need for additional synchronization and handshaking between the CUT and the ATE, so it is especially suitable to be integrated in a low cost test scheme for SoC test In addition in contrast to previous dictionary-based coding techniques, even for the CUT with a small number of scan chains, the proposed approach can achieve satisfied reduction in test data volume. Experimental results showed the proposed test scheme works particularly well for the large ISCAS'89 benchmarks.

This paper focuses on SIMD processor synthesis and proposes a SIMD instruction set/functional unit synthesis algorithm. Given an initial assembly code and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with optimal SIMD functional units. It also synthesizes a SIMD instruction set. The input initial assembly code is assumed to run on a full-resource SIMD processor (virtual processor) which has all the possible SIMD functional units. In our algorithm, we introduce the SIMD operation decomposition and apply it to the initial assembly code and the full-resource SIMD processor. By gradually reducing SIMD operations or decomposing SIMD operations, we can finally find a processor core with small area under the given timing constraint. The promising experimental results are also shown.

This paper proposes a hardware/software cosynthesis algorithm for processors with heterogeneous registers. Given a CDFG corresponding to an application program and a timing constraint, the algorithm generates a processor configuration minimizing area of the processor and an assembly code on the processor. First, the algorithm configures a datapath which can execute several DFG nodes with data dependency at one cycle. The datapath can execute the application program at the least number of cycles. The branch and bound algorithm is applied and all the number of functional units and memory banks are tried. For an assumed number of functional units and memory banks, an appropriate number of heterogeneous registers and connections to functional units and registers are explored. The experimental results show effectiveness and efficiency of the algorithm.

This paper proposes a thread partitioning algorithm in low power high-level synthesis. The algorithm is applied to high-level synthesis systems. In the systems, we can describe parallel behaving circuit blocks(threads) explicitly. First it focuses on a local register file RF in a thread. It partitions a thread into two sub-threads, one of which has RF and the other does not have, RF. The partitioned sub-threads need to be synchronized with each other to keep the data dependency of the original thread. Since the partitioned sub-threads have waiting time for synchronization, gated clocks can be applied to each sub-thread. Then we can synthesize a low power circuit with a low area overhead, compared to the original circuit. Experimental results demonstrate effectiveness and efficiency of the algorithm.

This letter proposes a new hardware/software partitioning algorithm for processor cores with SIMD instructions. Given a compiled assembly code including SIMD instructions and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with a new assembly code. Firstly, we assume for each operation type a super SIMD functional unit which can execute all the SIMD instructions. Secondly we reduce a SIMD instruction or "sub-function" of each super functional unit, one by one, while the timing constraint is satisfied. At the same time, we update the assembly code so that it can run on the new processor configuration. By repeating this process, we finally find SIMD functional unit configuration as well as a processor core architecture. The promising experimental results are also shown.

A packed SIMD type operation or a SIMD operation is n-parallel b/n-bit sub-operations executed by the modified n-bit functional unit. Such a functional unit is called a SIMD functional unit and a processor core which can execute SIMD operations is called a SIMD processor core. SIMD operations can be effectively applied to image processing applications. This paper focuses on hardware/software cosynthesis of SIMD processor cores and particularly proposes a new simulator generator which simulates pipelined instructions for a SIMD processor. Generally, a SIMD functional unit has many options and then we can have so many different SIMD functional unit instances. However, since our hardware/software cosynthesis system synthesizes a special-purpose processor core for an input application program, it uses very limited SIMD functional unit instances. In the proposed approach, we consider a SIMD operation to be a set of SIMD sub-operations. By adding up the appropriate SIMD sub-operations, we construct a single SlMD operation. Then a SIMD functional unit behavior can be characterized by a collection of SIMD operations. This approach has the advantage that: if we have a small number of behavior libraries for SIMD suboperations, we can instantiate a particular SIMD functional unit behavior. Experimental results demonstrate the effectiveness of the proposed approach.

This letter proposes a new hardware/software partitioning algorithm for processor cores with SIMD instructions. Given a compiled assembly code including SIMD instructions and a timing constraint, the proposed algorithm synthesizes an area-optimized processor core with a new assembly code. Firstly, we assume for each operation type a super SIMD functional unit which can execute all the SIMD instructions. Secondly we reduce a SIMD instruction or "sub-function" of each super functional unit, one by one, while the timing constraint is satisfied. At the same time, we update the assembly code so that it can run on the new processor configuration. By repeating this process, we finally find SIMD functional unit configuration as well as a processor core architecture. The promising experimental results are also shown.

A packed SIMD type operation or a SIMD operation is n-parallel b/n-bit sub-operations executed by the modified n-bit functional unit. Such a functional unit is called a SIMD functional unit and a processor core which can execute SIMD operations is called a SIMD processor core. SIMD operations can be effectively applied to image processing applications. This paper focuses on hardware/software cosynthesis of SIMD processor cores and particularly proposes a new simulator generator which simulates pipelined instructions for a SIMD processor. Generally, a SIMD functional unit has many options and then we can have so many different SIMD functional unit instances. However, since our hardware/software cosynthesis system synthesizes a special-purpose processor core for an input application program, it uses very limited SIMD functional unit instances. In the proposed approach, we consider a SIMD operation to be a set of SIMD sub-operations. By adding up the appropriate SIMD sub-operations, we construct a single SlMD operation. Then a SIMD functional unit behavior can be characterized by a collection of SIMD operations. This approach has the advantage that: if we have a small number of behavior libraries for SIMD suboperations, we can instantiate a particular SIMD functional unit behavior. Experimental results demonstrate the effectiveness of the proposed approach.

本研究室ではCAM(連想メモリ)を使用するプロセッサを合成するシステムを構築中である.現在のシステムはC言語で記述されたアプリケーション記述を入力としてそのアプリケーションを実行するプロセッサの最適なハードウェア構成を出力する.本稿では現在のシステムに面積制約機能を付加する.提案手法ではCAMの一部をRAMに置換してプロセッサの面積を削減する.本システムは面積制約を満足した上で実行時間を最小化する,CAMからRAMへ置換する量を導出できる.計算機実験により,面積制約を満たした上で,システムに入力されたアプリケーションを最遠に実行するプロセッサの構成を出力できる事を確認した.

我々はCAM(連想メモリ)を使用するプロセッサを対象としたハードウェア/ソフトウェア協調合成システムを構築中である.現在のシステムはC言語で記述されたアプリケーション記述を入力としてそのアプリケーションを実行するプロセッサの最適なハードウェア構成を出力する.本稿では,現在のシステムを拡張し,面積制約機能を付加したCAMプロセッサ向けハードウェア/ソフトウェア協調合成システムを提案する.提案手法では面積制約を満足した上で実行時間を最小化するCAMワード数を導出し,CAMの一部をRAMに置換してプロセッサの面積を削減する.計算機実験により,面積制約を満たした上で,システムに入力されたアプリケーションを最速に実行するプロセッサの構成を出力できる事を確認した.

IPマルチキャスト環境において,転送信頼性を確保する高信頼性マルチキャストが盛んに研究されている.現在までに様々な再送手法が提案されているが,それぞれ適したネットワークの状況やマルチキャストグループの規模は異なる.本研究では,時間的に変化するネットワーク状況や受信ホスト数に対応し,その状況に適した再送手法を選択する高信頼性マルチキャストプロトコルを提案する.提案プロトコルでは,隣接する他の受信ホストまでのホップ数である分岐距離という指標を用いて再送手法の決定を行う.幾つかの再送手法を単独で用いた場合と提案プロトコルをシュミレーションして比較した結果,提案プロトコルの配信完了所要時間が減少したことを確認した.

Content addressable memory (CAM) is one of the functional memories which realize word-parallel equivalence search. Since a CAM unit is generally used in a particular application program, we consider that appropriate design for CAM units is required depending on the requirements for the application program. This paper proposes a hardware/software cosynthesis system for CAM processors. The input of the system is an application program written in C including CAM functions and a constraint for execution time (or CAM processor area). Its output is hardware descriptions of a synthesized processor and a binary code executed on it. Based on the branch-and-bound method, the system determines which CAM function is realized by a hardware and which CAM function is realized by a software with meeting the given timing constraint (or area constraint) and minimizing the CAM processor area (or execution time of the application program). We expect that we can realize optimal CAM processor design for an application program. Experimental results for several application programs show that we can obtain a CAM processor whose area is minimum with meeting the given timing constraint.

Content addressable memory (CAM) is one of the functional memories which realize word-parallel equivalence search. Since a CAM unit is generally used in a particular application program, we consider that appropriate design for CAM units is required depending on the requirements for the application program. This paper proposes a hardware/software cosynthesis system for CAM processors. The input of the system is an application program written in C including CAM functions and a constraint for execution time (or CAM processor area). Its output is hardware descriptions of a synthesized processor and a binary code executed on it. Based on the branch-and-bound method, the system determines which CAM function is realized by a hardware and which CAM function is realized by a software with meeting the given timing constraint (or area constraint) and minimizing the CAM processor area (or execution time of the application program). We expect that we can realize optimal CAM processor design for an application program. Experimental results for several application programs show that we can obtain a CAM processor whose area is minimum with meeting the given timing constraint.

This paper proposes a new hardware/software partitioning algorithm for processor cores with SIMD instructions. Given a compiled assembly code including SIMD instructions, a timing constraint of execution time, and available hardware units, the proposed algorithm synthesizes an area-optimized processor core with a new assembly code. Firstly, we assume an initial processor core on which an input assembly code can run with the shortest execution time. Secondly we reduce a hardware unit added to a processor core one by one while the timing constraint is satisfied. At the same time, we update the assembly code so that it can run on the new processor configuration. By repeating this process, we finally obtain a processor core architecture with small area under the given timing constraint. We expect that we can obtain a processor core which has appropriate SIMD functional units for running the input application program. The promising experimental results are also shown.

This paper proposes a new hardware/software partitioning algorithm for processor cores with SIMD instructions. Given a compiled assembly code including SIMD instructions, a timing constraint of execution time, and available hardware units, the proposed algorithm synthesizes an area-optimized processor core with a new assembly code. Firstly, we assume an initial processor core on which an input assembly code can run with the shortest execution time. Secondly we reduce a hardware unit added to a processor core one by one while the timing constraint is satisfied. At the same time, we update the assembly code so that it can run on the new processor configuration. By repeating this process, we finally obtain a processor core architecture with small area under the given timing constraint. We expect that we can obtain a processor core which has appropriate SIMD functional units for running the input application program. The promising experimental results are also shown.

本稿では,ハードウェアIPの応答時間を考慮したプロセッサコアのハードウェア/ソフトウェア分割手法を提案する.我々は対象とするアプリケーションに応じて利用するハードウェアIPを始めに決定した上で,機能・性能に過不足のないプロセッサコアを合成するシステムLSI設計アプローチを提案している.そこで適切な構成をもったプロセッサコアを合成するためには,ハードウェアIPの応答時間を考慮したハードウェア/ソフトウェア分割が有効である.提案手法はハードウェアIPの応答時間を命令レベルで考慮することで既存手法を拡張しており,これによりプロセッサコアとハードウェアIPが独立したタスクを効率良く並列実行することが可能となる.計算機実験により提案手法を評価し,本設計アプローチの有効性を示す.

本稿では,ハードウェアIPの応答時間を考慮したプロセッサコアの合成システムと,これを利用したシステムLSI設計のフレームワークを提案する.ハードウェアIPを利用したシステムLSI設計では,システムLSIに要求される性能に対して必要にして十分な性能のIPが必ずしも用意されているとは限らない.そこでシステムのハードウェア/ソフトウェア分割後,ハードウェアで実現する機能にはIPを用いるが,ソフトウェアを動作させるプロセッサコアにはIPを用いず,アプリケーションに応じて性能に過不足のないプロセッサコアを自動合成する.提案するフレームワークに沿ってJPEGエンコーダを設計し,計算機実験により提案する合成システムとフレームワークの有効性を示す.

This paper proposes a new hardware/software partitioning algorithm for processor cores with SIMD instructions. Given a compiled assembly code including SIMD instructions, a timing constraint of execution time, and available hardware units, the proposed algorithm synthesizes an area-optimized processor core with a new assembly code. Firstly, we assume an initial processor core on which an input assembly code can run with the shortest execution time. Secondly we reduce a hardware unit added to a processor core one by one while the timing constraint is satisfied. At the same time, we update the assembly code so that it can run on the new processor configuration. By repeating this process, we finally obtain a processor core architecture with small area under the given timing constraint. We expect that we can obtain a processor core which has appropriate SIMD functional units for running the input application program. The promising experimental results are also shown.

This paper proposes a high-level energy-optimizing algorithm which can synthesize low energy system VLSIs. Given an initial system hardware obtained from an abstract behavioral description, the proposed algorithm applies to it the three energy reduction techniques, 1) reducing supply voltage, 2) selecting lower energy modules, and 3) applying gated clocks. By incorporating our area/delay/power estimation, the proposed algorithm can obtain low energy system VLSIs meeting the constraints of area, delay, and execution time. The proposed algorithm has been incorporated into a high-level synthesis system and experimental results demonstrate effectiveness and efficiency of the algorithm.

The motion estimation can choose the most suitable algorithm for different kinds of motion types, formats, and characteristics. The video encoding system can be optimized for quality, speed, and power consumption. In this paper, we propose a reconfigurable approach to a motion estimation algorithm and hardware architecture. The proposed algorithm determines motion type and then selects adapted block-matching algorithm for different kinds of motion sequences. The quality of our algorithm is better than that of the TSS and the BBGDS algorithm, or comparable to the performance of the better of the two, and the computational complexity of our algorithm is significantly less than that of the TSS. We also propose hardware architecture for realizing two kinds of motion estimations in the same hardware. We implemented the flexible and reconfigurable hardware architecture by using address generator unit, delay unit, and parameters and by using the hardware description language (VHDL) and the SYNOPSYS synthesis design tools. We analyze the performance of the algorithm and present adapted algorithm for a low cost real time application.

This paper proposes a high-level energy-optimizing algorithm which can synthesize low energy system VLSIs. Given an initial system hardware obtained from an abstract behavioral description, the proposed algorithm applies to it the three energy reduction techniques, 1) reducing supply voltage, 2) selecting lower energy modules, and 3) applying gated clocks. By incorporating our area/delay/power estimation, the proposed algorithm can obtain low energy system VLSIs meeting the constraints of area, delay, and execution time. The proposed algorithm has been incorporated into a high-level synthesis system and experimental results demonstrate effectiveness and efficiency of the algorithm.

The motion estimation can choose the most suitable algorithm for different kinds of motion types, formats, and characteristics. The video encoding system can be optimized for quality, speed, and power consumption. In this paper, we propose a reconfigurable approach to a motion estimation algorithm and hardware architecture. The proposed algorithm determines motion type and then selects adapted block-matching algorithm for different kinds of motion sequences. The quality of our algorithm is better than that of the TSS and the BBGDS algorithm, or comparable to the performance of the better of the two, and the computational complexity of our algorithm is significantly less than that of the TSS. We also propose hardware architecture for realizing two kinds of motion estimations in the same hardware. We implemented the flexible and reconfigurable hardware architecture by using address generator unit, delay unit, and parameters and by using the hardware description language (VHDL) and the SYNOPSYS synthesis design tools. We analyze the performance of the algorithm and present adapted algorithm for a low cost real time application.

Content Addressable Memory(CAM)を使用したプロセッサを対象とした自動合成システムを提案する.本システムではC言語によるCAMユニットの機能を使用したアプリケーションプログラムを入力とし,CAMプロセッサのハードウェア記述とCAMプロセッサ上で動作するバイナリコードを出力する.一致検索や閾値検索機能を持つ機能の異なる10種類のCAMセルアレーから合成することで,アプリケーションに適したCAMプロセッサを設計できる.計算機上に本システムを構築し,アプリケーションプログラムの入力によってCAMプロセッサのハードウェア記述およびバイナリコードが得られることを確認した.

本稿では,Bluetoothを用いたストリーミングを主目的としたアクセスネットワーク環境を考える.一般に無線回線は有線回線と比べてパケットの欠落が多いためストリーミングデータの劣化が起こる.人間の耳は100[ms]程度の遅延時間は感知しないという特徴がある。そこでデータに許容可能な遅延時間を考え,許容遅延時間以内でエラー訂正可能ならばエラー訂正することで,ストリーミングデータの劣化が抑えられうと考えられる.Bluetoothを用いたアクセスネットワークの環境では,一般的にサーバの存在する回線側は端末の存在するBluetooth回線側と比べて通信速度が速いという特徴がある.そのためエラー訂正等で増えるサーバ側回線の通信量がストリーミングデータへ与える影響は少ないと考えられ,この手法はBluetoothを用いたアクセスネットワークの環境での利用に適していると考える.本稿では上記の特徴を使って最大許容遅延という指標を導入し,再送処理・帯域制御することで品質の良いストリーミングデータ伝送することを提案する.さらに,計算機実験によってその有効性を示す。

Packed SIMD型命令を持つプロセッサのハードウェア/ソフトウェア協調合成システムSPADE-simdでは,Packed SIMD型命令を持つプロセッサを対象とした並列化コンパイラが必要となる.この並列化コンパイラは,まず,SPADE-simdが対象とするプロセッサコアにおいて,付加可能な全てのハードウェアを持つ仮想的なプロセッサを仮定する.仮想プロセッサ上で,入力アプリケーションの命令レベル並列性をPacked SIMD型命令を用いて最大限に抽出し,アセンブリコードを出力する.並列化コンパイラの出力により,合成するプロセッサの初期構成が得られる.本稿では,並列化コンパイラの中核をなすレジスタ内SIMD並列化アルゴリズム,命令併合アルゴリズムを提案する.レジスタ内SIMD並列化アルゴリズムは,入力アプリケーションの命令レベル並列性を抽出し,レジスタ内に低精度データを梱包・整列することでPacked SIMD型命令を利用する.命令併合アルゴリズムは,複数のPacked SIMD型命令を1命令に併合し,シフト・飽和演算を含むPacked SIMD型命令やPacked SIMD型乗加算命令を利用可能とする.提案手法により,アプリケーションの命令レベル並列性をPacked SIMD型命令を用いて抽出し,高速化したアセンブリコードを出力する.提案手法を用いて計算機上に並列化コンパイラを実装し,有効性を評価する.

本稿では,Packed SIMD型命令セットを持った画像処理プロセッサを対象としたハードウェア/ソフトウェア協調合成システムSPADE-simdにおけるハードウェア/ソフトウェア分割手法を提案する.画像処理では,プロセッサの基本ピット幅に比較して小さいビット幅を持ったデータを演算対象とする.Packed SIMD型命令は1つの演算器を用いて複数の小ピットデータに対する演算を実行するため,画像処理に適用することでデータの並列演算による処理の高速化が可能となる.アプリケーションプログラムに記述される全てのPacked SIMD型命令について,命令実行機能をプロセッサコアに含まれる対応Packed SIMD型演算器のいずれかに割り当てる.割り当てられた命令の組合せに応じて,各々のPacked SIMD型演算器のハードウェア構成は変更が可能である.プロセッサに付加されるPacked SIMD型演算器の命令実行機能構成を適切に決定することで,演算器およびプロセッサコアのハードウェアコストを削減することができる.計算機実験により提案手法を評価し,結果を報告する.

本稿では，Packed SIMD型命令を持つプロセッサを対象としたハードウェア/ソフトウェア協調合成システムのためのハードウェアユニット生成手法を提案する．ハードウェアユニットを複数の部分機能を実現するハードウェアの組合せで構成することにより，与えられた命令集合を実行できるハードウェアユニットの構成を，高速に複数列挙し，ハードウェアユニットの面積と遅延を見積もることができ，Packed SIMD型命令を持つプロセッサコアを協調合成システムにより得ることができるようになる．計算機実験により本手法の有効性を評価した．This paper proposes a hardware unit generation algorithm for ahardware/software cosynthesis system of digital signal processors withpacked SIMD type instructions. Given a set of instructions, the proposedalgorithm extracts a set of subfunctions to be required by the hardwareunit and generates more than one architecture candidates for hardwareunits. The algorithm also outputs the estimated area and delay of eachof the generated hardware units. The execution time of the proposedalgorithm is very short and thus it can be easily incorporated into theprocessor core synthesis system. Experimental results demonstrateeffectiveness and efficiency of the alogorithm.

At high-level synthesis for system VLSIs, their power consumption is efficiently reduced by applying gated clocks to them. Since using gated clocks causes the reduction of power consumption and the increase of area/delay, estimating tradeoff between power and area/delay by applying gated clocks is very important. In this paper. we discuss the amount of variance of area, delay and power by applying gated clocks. We propose a simple gate-level circuit model and estimation equations. We vary parameters in our proposed circuit model, and evaluate power consumption by back-annotating gate-level simulation results to the original circuit. This paper also proposes a conditional expression for applying gated clocks The expression shows whether or not we can reduce power consumption by applying gated clocks. We confirm the accuracy of proposed estimation equations by experiments.

At high-level synthesis for system VLSIs, their power consumption is efficiently reduced by applying gated clocks to them. Since using gated clocks causes the reduction of power consumption and the increase of area/delay, estimating tradeoff between power and area/delay by applying gated clocks is very important. In this paper. we discuss the amount of variance of area, delay and power by applying gated clocks. We propose a simple gate-level circuit model and estimation equations. We vary parameters in our proposed circuit model, and evaluate power consumption by back-annotating gate-level simulation results to the original circuit. This paper also proposes a conditional expression for applying gated clocks The expression shows whether or not we can reduce power consumption by applying gated clocks. We confirm the accuracy of proposed estimation equations by experiments.

制御処理ハードウェアの高位合成システムにおける,ハードウェアの実行時間の削減を実現するコントロールデータフローグラフ変形手法を提案する.本高位合成システムは,アプリケーションプログラム記述から,状態数が最小で且つ高速な演算器が割り当てられ高速に動作する初期状態、遷移グラフを初期ハードウェア候補として出力する.提案するコントロールデータフローグラフ変形手法は2つのアプローチにより,(1)初期状態遷移グラフの実行クロックサイクル数の削減,(2)初期状態遷移グラフのクロック周期の削減,を実現する.(1)を部分コントロールデータフローグラフの複製処理によるパス長の削減により実現する.(2)を一算術演算処理内のメモリ読み込み処理と演算処理を異なるイタレーションに分散させるパイプライン実行により実現する.計算機実験により提案手法の有効性を示す.

本稿では,1)類推列挙機構を有し,2)定性指標を考慮した,動画像処理向けVLSIを対象とするハードウェア/ソフトウェア協調設計システムを提案する.本システムでは,過去に設計したIPをデータベースに蓄積し,類推列挙に利用する.類推列挙とは,データベースに存在するIPから,その並列処理の度合いを変化させたときの面積/遅延の変化を類推する機能のことである.定性指標とは,他のハードウェア部品との接続しやすさといった,面積/時間/電力以外の定性的な指標のことである.1)を導入することで仮想的に解候補数を増やすことができ,面積/時間制約を満たす解が得られやすくなる.2)を導入することで定量指標が同等な解候補の中から,設計コストの低い候補を選ぶことができる.これらの機能を含むシステムを構築し,計算機実験によってその有効性を示す.

If motion estimation can choose the most suitable algorithm according to the changing characteristics of input image signals, we can get benefits, which improve quality and performance, reduce power consumption, and an optimize system. In this paper we propose a reconfigurable approach to motion estimation algorithm and architecture. The propose algorithm determines motion type and then selects adapted algorithm in order to improve quality and performance of images. We implemented the flexible and reconfigurable architecture by hardware with address generator unit, delay unit, and parameters. Our architecture supports more than one block-matching algorithm and parameters providing to optimize system. We are implementing our architecture by using hardware description language (VHDL) and synthesis design tools. We analyze the performance of architecture and present adaption to algorithm for a low cost real time application.

Let us consider to synthesize a processor core with SIMD instructions by a hardware/software cosynthesis system. The system is required to configure functional units executing SIMD instructions and obtain the area and delay of the functional units to evaluate the synthesized processor core. This paper proposes a hardware unit generation algorithm for a hardwaxe/software cosynthesis system of processors with SIMD instructions. Given a set of instructions to be executed by a hardware unit and constraints for area and delay of the hardware unit, the proposed algorithm extracts a set of subfunctions to be required by the hardware unit and generates more than one architecture candidates for the hardware unit. The algorithm also outputs the estimated area and delay of each of the generated hardware units. The execution time of the proposed algorithm is very short and thus it can be easily incorporated into the processor core synthesis system. Experimental results demonstrate effectiveness and efficiency of the algorithm.

The authors consider the synthesis of a processor core with SIMD instructions by a hardware/software cosynthesis system. The system is required to configure functional units executing SIMD instructions and obtain the area and delay of the functional units to evaluate the synthesized processor core. This paper proposes a hardware unit generation algorithm for a hardware/software cosynthesis system of processors with SIMD instructions. Given a set of instructions to be executed by a hardware unit and constraints for area and delay of the hardware unit, the proposed algorithm extracts a set of subfunctions to be required by the hardware unit and generates more than one architecture candidates for the hardware unit. The algorithm also outputs the estimated area and delay of each of the generated hardware units. The execution time of the proposed algorithm is very short and thus it can be easily incorporated into the processor core synthesis system. Experimental results demonstrate effectiveness and efficiency of the algorithm.

Let us consider to synthesize a processor core with SIMD instructions by a hardware/software cosynthesis system. The system is required to configure functional units executing SIMD instructions and obtain the area and delay of the functional units to evaluate the synthesized processor core. This paper proposes a hardware unit generation algorithm for a hardwaxe/software cosynthesis system of processors with SIMD instructions. Given a set of instructions to be executed by a hardware unit and constraints for area and delay of the hardware unit, the proposed algorithm extracts a set of subfunctions to be required by the hardware unit and generates more than one architecture candidates for the hardware unit. The algorithm also outputs the estimated area and delay of each of the generated hardware units. The execution time of the proposed algorithm is very short and thus it can be easily incorporated into the processor core synthesis system. Experimental results demonstrate effectiveness and efficiency of the algorithm.

本稿では, 面積/遅延/実行時間の制約を満たしながら低消費電力なシステムVLSIを合成可能な高位合成システムを提案する.低消費電力化手法として, 1)電源電圧の低減, 2)低消費電力なモジュールの選択, 3)Gated Clockの3つの手法を採用した.一般に, これら3つの手法の適用により消費電力は低減可能であるが, 面積/遅延/実行時間は増加してしまう.提案する手法では, 面積/遅延/実行時間の変化量を予測することによって, これらの各制約を満たしながら初期ハードウェアよりも消費電力を低減したハードウェアを合成することができる.さらに, 計算機実験により消費電力が低減されていることを確認した.

This letter proposes a hardware/software partitioning algorithm for digital signal processor cores with two register files. Given a compiled assembly code and a timing constraint of execution time, the proposed algorithm generates a processor core configuration with a new assembly code running on the generated processor core. The proposed algorithm considers two register files and determines the number of registers in each of register files. Moreover the algorithm considers two or more types of functional units for each arithmetic or logical operation and assigns functional units with small area to a processor core without causing performance penalty. A generated processor core will have small area compared with processor cores which have a single register file or those which consider only one type of functional units for each operation. The experimental results demonstrate the effectiveness and efficiency of the proposed algorithm.

Hardware/software partitioning is one of the key processes in a hardware/software cosynthesis system for digital signal processor cores. In hardware/software partitioning, area and delay estimation of a processor core plays an unportant role since the hardware/software partitioning process must determine which part of a processor core should be realized by hardware units and which part should be realized by a sequence of instructions based on execution time of an input application program and area of a synthesized processor core. This paper proposes area and delay estimation equations for digital signal processor cores. For area estimation, we show that total area for a processor core can be derived from the sum of area for a processor kernel and area for additional hardware units. Area for a processor kernel can be mainly obtained by minimum area for a processor kernel and overheads for adding hardware units and registers. Area for a hardware unit can be mainly obtained by its type and operation bit width. For delay estimation, we show that critical path delay for a processor core can be derived from the delay of a hardware unit which is on the critical path in the processor core. Experimental results demonstrate that errors of area estimation are less than 2% and errors of delay estimation are less than 2 ns when comparing estimated area and delay with logic-synthesized area and delay.

演算語長をnビットからn/2ビットに縮小することでプロセッサの面積削減を図る.一般的に, このときnビットの結果を得るには少なくとも2回の演算命令を実行しなければならない.しかし, ここでアプリケーションプログラムの内部変数のデータ長がn/2ビットのみであると仮定する.このとき, 演算命令を一度実行するだけで結果を得られる.我々の提案しているハードウェア/ソフトウェア協調合成システムでは, これまで, アプリケーションプログラムのデータ長はプロセッサコアの演算語長と等しいことを前提としていた.本稿では, 演算語長を縮小する手法を提案する.この手法は, 内部変数の演算精度に応じて各nビットの演算命令を1つ, もしくは2つ以上のn/2ビットの演算命令に繰り返し置き換える.

This letter proposes a hardware/software partitioning algorithm for digital signal processor cores with two register files. Given a compiled assembly code and a timing constraint of execution time, the proposed algorithm generates a processor core configuration with a new assembly code running on the generated processor core. The proposed algorithm considers two register files and determines the number of registers in each of register files. Moreover the algorithm considers two or more types of functional units for each arithmetic or logical operation and assigns functional units with small area to a processor core without causing performance penalty. A generated processor core will have small area compared with processor cores which have a single register file or those which consider only one type of functional units for each operation. The experimental results demonstrate the effectiveness and efficiency of the proposed algorithm.

Hardware/software partitioning is one of the key processes in a hardware/software cosynthesis system for digital signal processor cores. In hardware/software partitioning, area and delay estimation of a processor core plays an unportant role since the hardware/software partitioning process must determine which part of a processor core should be realized by hardware units and which part should be realized by a sequence of instructions based on execution time of an input application program and area of a synthesized processor core. This paper proposes area and delay estimation equations for digital signal processor cores. For area estimation, we show that total area for a processor core can be derived from the sum of area for a processor kernel and area for additional hardware units. Area for a processor kernel can be mainly obtained by minimum area for a processor kernel and overheads for adding hardware units and registers. Area for a hardware unit can be mainly obtained by its type and operation bit width. For delay estimation, we show that critical path delay for a processor core can be derived from the delay of a hardware unit which is on the critical path in the processor core. Experimental results demonstrate that errors of area estimation are less than 2% and errors of delay estimation are less than 2 ns when comparing estimated area and delay with logic-synthesized area and delay.

This paper proposes an area/time optimizing algorithm in a high-level synthesis system for control-based hardwares. Given a call graph whose node corresponds to a control flow of an application program. the algorithm generates a set of state-transition graphs which represents the input call graph under area and timing constraint. In the algorithm. first state-transition graphs which satisfy only timing constraint are generated and second they are transformed so that they can satisfy area constraint. Since the algorithm is directly applied to control-flow graphs, it can deal with control flows such as bitwise processed and conditional branches. Further, the algorithm synthesizes more than one hardware architecture candidates from a single call graph for an application program. Designers of an application program can select several good hardware architectures among candidates depending on multiple design criteria. Experimental results for several control-based hardwares demonstrate effectiveness and efficiency of the algorithm.

This paper proposes an area/time optimizing algorithm in a high-level synthesis system for control-based hardwares. Given a call graph whose node corresponds to a control flow of an application program. the algorithm generates a set of state-transition graphs which represents the input call graph under area and timing constraint. In the algorithm. first state-transition graphs which satisfy only timing constraint are generated and second they are transformed so that they can satisfy area constraint. Since the algorithm is directly applied to control-flow graphs, it can deal with control flows such as bitwise processed and conditional branches. Further, the algorithm synthesizes more than one hardware architecture candidates from a single call graph for an application program. Designers of an application program can select several good hardware architectures among candidates depending on multiple design criteria. Experimental results for several control-based hardwares demonstrate effectiveness and efficiency of the algorithm.

1個のbビット演算ユニットを用いてn個のb/nビット演算を実現するPacked SIMD型命令は画像処理などに有効である. Packed SIMD型命令を持つプロセッサコアをハードウェア/ソフトウェア協調合成システムによって合成するとき, 必要な命令が実行できるPacked SIMD型演算ユニットを構成し面積と遅延を高速に見積もることが要求される. そこで, 本稿では複数のハードウェアユニットを高速に構成するPakced SIMD型ハードウェアユニット生成手法を提案する. 本手法は, 1つのハードウェアユニットで実行される命令の集合と, 生成されるハードウェアユニットの面積と遅延の制約を入力とし, ハードウェアユニットに必要となる部分機能を抽出して, その部分機能を実現するハードウェアを組み合わせることでハードウェアユニット構成を複数列挙し面積と遅延の見積もり値を出力する. 提案手法を計算機上に実装しPacked SIMD型演算ユニットに適用した結果を示しその有効性を評価する.

本稿では設計資産の再利用が実現するシステムLSIを対象としたハードウェア/ソフトウェア分割システムを提案する。システムはアーキテクチャデータベースと実現可能割付け系で構成される。アーキテクチャデータベースはアルゴリズム名と既設計事例を対応付けるデータベースである。アプリケーションは機能ブロック図で表す。機能ブロック図は機能モジュールと機能モジュールの接続関係からなり,各機能モジュールにはアルゴリズム名を与える。実現可能割付け系にアプリケーションの実行時間制約と機能ブロック図を入力として与える。実現可能割付け系は時間制約を満たすモジュール-既設計事例割付けを全て列挙する。MPEG-4エンコーダを入力とした適用事例を示す。

画像処理アプリケーションの性能向上を目的として多くの汎用およびディジタル信号処理向けプロセッサで拡張命令セットが提供されている。主にこの機能拡張はPacked SIMD型の命令の追加によるものである。Packed SIMD型命令を用いることで,アプリケーションのデータ並列性を活用することができる。Packed SIMD型命令は多くの種類が存在するが,アプリケーションは限られた命令しか使用しない。そこでプロセッサをアプリケーションスペシフィック合成をすることで,実行時間制約を満足するプロセッサの面積を大きく削減することができる。本稿では,Packed SIMD型命令セットをもつプロセッサのハードウェア/ソフトウェア協調合成システムとそのコンパイラのレジスタ内SIMD並列化アルゴリズムを提案する。C言語によるアプリケーション記述とアプリケーションデータを入力とし,プロセッサのハードウェア記述,プロセッサ上で動作するオブジェクトコードおよびソフトウェア環境を出力とする。並列化コンパイラはプロセッサコアに付加可能な全てのハードウェアを付加したプロセッサ上で動作するオブジェクトコードを生成する。コンパイラは命令レベル並列性,レジスタ内SIMD並列性を抽出し,実行時間の最小化を目指す。計算機実験によって本コンパイラを評価した結果を報告する。

本稿では，制御処理ハードウェアの高位合成システムのための面積／遅延見積もり手法を提案する．面積／遅延見積もりでは本システム構成の1つである面積／時間最適化において構築された状態遷移グラフを入力としてその状態遷移グラフに対する面積見積もり値および遅延見積もり値を出力する．提案見積もり手法では状態数および演算器の種類に依存した見積もり式を定式化することでハードウェアの制御部分を含めた面積，遅延の見積もり値を得ている．提案手法をハフマン符号化を始めとするいくつかの制御処理アプリケーションプログラムに適用し，その有効性を評価する．This paper proposes an area/delay estimation technique in high-level synthesis for control flow based hardwares. At area/delay estimation, the input is the state-transition graph, which is generated by the area/time optimizing. The output is estimated area and delay value for the state-transition graph. Our estimation technique gives area and delay including control part of hardware, using an estimation equation. The equation has been decided by number of operations, number of states and type of operations. Experimental results for several control-based hardware demonstrate effectiveness and efficiency of the technique.

Hardware/software partitioning is one of the key processes in a hardware/software cosynthesis system for digital signal processor cores. In hardware/software partitioning, area and delay estimation of a processor core plays an important role since the hardware/software partitioning process must determine which part of a processor core should be realized by hardware units and which part should be realized by a sequence of instructions based on execution time of an input application program and area of a synthesized processor core. This paper proposes area and delay estimation equations for digital signal processor cores. For area estimation, we show that total area for a processor core can be derived from the sum of area for a processor kernel and area for additional hardware units. Area for a processor kernel can be mainly obtained by minimum area for a processor kernel and overheads for adding hardware units and registers. Area for a hardware unit can be mainly obtained by its type and operation bit width. For delay estimation, we show that critical path delay for a processor core can be derived from the delay of a hardware unit which is on the critical path in the processor core. Experimental results demonstrate that errors of area estimation are less than 2% and errors of delay estimation are less than 2ns when comparing estimated area and delay with logic-synthesized area and delay.

本稿では, ディジタル信号処理ハードウェアのデータパス設計を対象とした計算時間予測に基づき解を導出するリソースバインディング手法を提案する.提案手法は, 発見的算法に基づくリソースバインダと分枝限定法に基づくリソースバインダを組み合わせたものである.まず, 発見的算法に基づくリソースバインダが割り当てるリソース数を変化させ, 残りのリソースを分枝限定法に基づくリソースバインダで割り当てた場合, 計算時間がどのように増減するかを予測する.その予測に基づき, 設計者の与える計算時間制約を満足するように発見的算法に基づくリソースバインダで割り当てるリソース数を決定し, 実際に割当を実行する.残りのリソースの割当を分枝限定法に基づくリソースバインダで決定することにより, 最終的なリソースバインディングの解を得る.計算機実験により, 本手法の有効性を確認する.

近年, システムの動作中にシステムの一部の論理を書き換える動的再構成可能システムが研究されている.従来, 動的再構成可能システムはそのハードウェア構成が固定されており, そのため, 特定のアプリケーションに適用するにはそのシステム上の構成では機能が冗長, あるいは, 不足するという問題があった.本研究では, これらの問題を解決するためにパラメータ付けされた動的再構成可能システムを提案する.提案する動的再構成可能システムはPCIインターフェース部, 演算を実行する機能部, それらを制御する制御部から成る.機能部の処理速度や回路規模, 再構成時間, ピン数などのデバイス構成やデバイス間の接続方式はアプリケーションに応じて可変である.デバイスのパラメータにはその性能に応じてコストが割り当てられておりユーザの指定した制約コスト内での最速処理をするデバイスのパラメータとそのデバイスから成るシステム構成の決定が可能である.本研究ではアプリケーションの適用例として画像符号化処理とパケット処理を挙げ, システムの有効性を評価する.

Hardware/software partitioning is one of the key processes in a hardware/software cosynthesis system for digital signal processor cores. In hardware/software partitioning, area and delay estimation of a processor core plays an important role since the hardware/software partitioning process must determine which part of a processor core should be realized by hardware units and which part should be realized by a sequence of instructions based on execution time of an input application program and area of a synthesized processor core. This paper proposes area and delay estimation equations for digital signal processor cores. For area estimation, we show that total area for a processor core can be derived from the sum of area for a processor kernel and area for additional hardware units. Area for a processor kernel can be mainly obtained by minimum area for a processor kernel and overheads for adding hardware units and registers. Area for a hardware unit can be mainly obtained by its type and operation bit width. For delay estimation, we show that critical path delay for a processor core can be derived from the delay of a hardware unit which is on the critical path in the processor core. Experimental results demonstrate that errors of area estimation are less than 2% and errors of delay estimation are less than 2ns when comparing estimated area and delay with logic-synthesized area and delay.

CAM (Content Addressable Memory) units are generally designed so that thc) carl be applied to variety of application programs. However, if a particular application runs on CAM units, some functions in CAM units may be often used and other functions may never be used. We consider that appropriate design for CAM units is required depending on the requirements for a given application program. This paper proposes a CAM processor synthesis system based on behavioral descriptions. The input of the system is an application programs written in C including CAM functions, and its output is hardware descriptions of a synthesized processor and a binary code executed on it. Since the system determines functions in CAM units and synthesizes a CAM processor depending on the requirements of an application program, we expect that a synthesized CAM processor can execute the application program with small processor area and delay. Experimental results demonstrate its efficiency and effectiveness.

CAM (Content Addressable Memory) units are generally designed so that thc) carl be applied to variety of application programs. However, if a particular application runs on CAM units, some functions in CAM units may be often used and other functions may never be used. We consider that appropriate design for CAM units is required depending on the requirements for a given application program. This paper proposes a CAM processor synthesis system based on behavioral descriptions. The input of the system is an application programs written in C including CAM functions, and its output is hardware descriptions of a synthesized processor and a binary code executed on it. Since the system determines functions in CAM units and synthesizes a CAM processor depending on the requirements of an application program, we expect that a synthesized CAM processor can execute the application program with small processor area and delay. Experimental results demonstrate its efficiency and effectiveness.

In digital signal processing, bit width of intermediate variables should be longer than that of input and output variables in order to execute intermediate operations with high precision. Then a processor cole for digital signal processing is required to have two types of register files, one of which is used by input and output variables and the other one is used by intermediate variables. This paper proposes a hardware/software cosynthesis system for digital signal processor cores with two types of register files. Given an application program and its data, the system synthesizes a hardware description of a processor cure, an object code running on the processor core, and software environments. A synthesized processor core can be composed of a processor kernel, multiple data memory buses, hardware loop units, addressing units, and multiple functional units. Furthermore it can have two types of register files RF1 and RF2. The bit width and number of registers in RF1 or RF2 will be determined based on a given application program. Thus a synthesized processor core will have small area with keeping high precision of intermediate operations compared with a processor core with only one register file. The experimental results demonstrate the effectiveness of the proposed system.

歩行者を対象とする地図データ配信システムとはベクトル型の地図データを格納する地図サーバ, アンテナを設置している基地局, 及び歩行者が所持している複数個の携帯端末から構成されるシステムである.歩行者はGPSにより位置情報を取得し, 位置情報を基地局を通し地図サーバに送信する.地図サーバは位置を中心とする地図データを基地局を通し, 携帯端末に送信する.本稿では, 携帯端末の位置情報が地図サーバに与えられたとき, 位置情報を中心とする質問長方形の内部に含まれる, または交差する地図データを探索する地図データ探索問題を取り上げる.これを高速に処理する地図サーバ専用プロセッサとして連想メモリを適用した地図サーバプロセッサおよび専用ハードウェアを持った地図サーバプロセッサを提案する.それぞれのプロセッサハードウェアをハードウェア記述言語で記述し, これらを論理合成することで, ハードウェアコスト, クリティカルパス遅延を見積もり, プロセッサの性能を評価した結果を報告する.

In digital signal processing, bit width of intermediate variables should be longer than that of input and output variables in order to execute intermediate operations with high precision. Then a processor cole for digital signal processing is required to have two types of register files, one of which is used by input and output variables and the other one is used by intermediate variables. This paper proposes a hardware/software cosynthesis system for digital signal processor cores with two types of register files. Given an application program and its data, the system synthesizes a hardware description of a processor cure, an object code running on the processor core, and software environments. A synthesized processor core can be composed of a processor kernel, multiple data memory buses, hardware loop units, addressing units, and multiple functional units. Furthermore it can have two types of register files RF1 and RF2. The bit width and number of registers in RF1 or RF2 will be determined based on a given application program. Thus a synthesized processor core will have small area with keeping high precision of intermediate operations compared with a processor core with only one register file. The experimental results demonstrate the effectiveness of the proposed system.

This paper proposes an area/time optimizing algorithm in high-level synthesis for control-based hardwares. Given a call graph whose node corresponds to a control flow of an application program, the algorithm generates a set of state-transition graphs which represents the input call graph under area and timing constraint. In the algorithm, first state-transition graphs which satisfy only timing constraint are generated and second they are transformed so that they can satisfy area constraint. Since the algorithm is directly applied to control-flow graphs, it can deal with control flows such as bit-wise processes and conditional branches. Further, the algorithm synthesizes more than one hardware architecture candidates from a single call graph for an application program. Designers of an application program can select several good hardware architectures among candidates depending on multiple design criteria. Experimental results for several control-based hardwares demonstrate effectiveness and efficiency of the algorithm. © 2000 IEEE.

This paper proposes a hardware/software partitioning algorithm for digital signal processor cores with two register files. Given a compiled assembly code and a timing constraint of execution time, the proposed algorithm generates a processor core configuration with a new assembly code running on the generated processor core. The proposed algorithm considers two register files and determines the number of registers in each of register files. Moreover the algorithm considers two or more functional units for each arithmetic or logical operation and assigns functional units with small area to a processor core without causing performance penalty. A generated processor core will have small area compared with processor cores which have a single register file or those which have only one functional unit for each operation. The experimental results demonstrate the effectiveness and efficiency of the proposed algorithm.

This paper proposes a hardware/software partitioning algorithm for digital signal processor cores with two register files. Given a compiled assembly code and a timing constraint of execution time, the proposed algorithm generates a processor core configuration with a new assembly code running on the generated processor core. The proposed algorithm considers two register files and determines the number of registers in each of register files. Moreover the algorithm considers two or more functional units for each arithmetic or logical operation and assigns functional units with small area to a processor core without causing performance penalty. A generated processor core will have small area compared with processor cores which have a single register file or those which have only one functional unit for each operation. The experimental results demonstrate the effectiveness and efficiency of the proposed algorithm.

This paper proposes a hardware/software cosynthesis system for digital signal processor cores and a hardware/software partitioning algorithm which is one of the key issues for the system. The target processor has a VLIW-type core which can be composed of a processor kernel, multiple data memory buses (X-bus and Y-bus), hardware loop units, addressing units, and multiple functional units. The processor kernel includes five pipeline stages (RISC-type kernel) or three pipeline stages (DSP-type kernel). Given an application program written in the C language and a set of application data, the system synthesizes a processor core by selecting an appropriate kernel (RISC-type or DSP-type kernel) and required hardware units according to the application program/data and the hardware costs. The system also generates the object code for the application program and a software environment (compiler and simulator) for the processor core. The experimental results demonstrate that the system synthesizes processor cores effectively according to the features of an application program and the synthesized processor cores execute most application programs with the minimum number of clock cycles compared with several existing processors.

2種類のレジスタファイルを持つディジタル信号処理向けプロセッサのハードウェア/ソフトウェア協調合成システムでは,ハードウェア/ソフトウェア分割の評価値として,アプリケーションプログラムの実行時間の見積り値と生成されるプロセッサコアの面積の見積り値が必要となる.これら見積り値を得るためには,実際にシステムを用いてハードウェアユニットを変化させ得られたプロセッサコア記述を論理合成ツールで論理合成した結果を解析し,見積り式を導出する必要がある.本稿では,プロセッサコアの面積見積り式および遅延見積り式の導出方法とその検証結果について報告する.面積見積り式の導出では,まず,プロセッサコアの面積がプロセッサカーネルとカーネルに付加されるハードウェアユニットの面積の和として表されることを示す.しかも,プロセッサカーネルの面積が付加するハードウェアユニットに依存する部分と汎用レジスタ数に依存する部分に分離して考えられる点に注目する.導出した面積見積り式によるプロセッサコアの面積見積り値は,論理合成結果後の面積値と比較して,誤差を2%程度に抑えられることが分かった.遅延見積り式の導出では,クリティカルパスを構成する演算器ごとに見積り式を導出することにより誤差を小さくできることを示す.導出した遅延見積り式によるプロセッサコアの1クロック周期は,論理合成結果後の1クロック周期と比較して,誤差を2ns以下に抑えられることが分かった.

本稿では,制御処理ハードウェアの高位合成システムのための面積/時間最適化アルゴリズムを提案する.面積/時間最適化アルゴリズムは,入力としてコールグラフおよびコールグラフを構成するコントロールフローグラフ集合を取り,面積制約および時間制約のもとに,コールグラフ全体を表す状態遷移グラフ集合を合成する.まず,時間制約のみを満足する状態遷移グラフを構築し,その後,面積制約を満足するよう状態遷移グラフを変換する.提案アルゴリズムは,コントロールフローグラフを直接的に操作するため,ビット処理および条件分岐処理といった制御処理を扱うことができ,しかも,アプリケーションプログラム全体を表す1個のコールグラフから,面積制約および時間制約を満足する複数個のハードウェア候補を列挙することができる.提案アルゴリズムをハフマン符号化を始めとする,いくつかの制御処理アプリケーションプログラムに適用し,その有効性を評価する.