This paper presents a design of a new multiple-valued matching VLSI processor for high-speed reasoning. It is useful in the application for real-time rule-based systems with large knowledge bases which are programmable. In order to realize high-speed reasoning, the matching VLSI processor can perform the fully parallel pattern matching between an input data and rules. On the based of direct multiple-valued encoding of each attribute in an input data and rules, pattern matching can be described by using only a programmable delta literal. Moreover, the programmable delta literal circuit can be easily implemented using two kinds of floating-gate MOS devices whose threshold voltages are controllable. In fact, it is demonstrated that four kinds of threshold voltages in a practical floating-gate MOS device can be easily programmable by appropriately controlling the gate, the drain and the source voltage. Finally, the inference time of the quaternary matching VLSI processor with 256 rules and conflict resolution circuits is estimated at about 360 (ns), and the chip area is reduced to about 30 percent, in comparison with the equivalent binary implementation.

A new multiple-valued current-mode (MVCM) logic circuit using substrate bias control is proposed for low-power VLSI systems at higher clock frequency. Since a multi-level threshold value is represented as a threshold voltage of an MOS transistor, a voltage comparator is realized by a single MOS transistor. As a result, two basic components, a comparator and an output generator in the MVCM logic circuit can be merged into a single MOS differential-pair circuit where the threshold voltages of MOS transistors are controlled by substrate biasing. Moreover, the leakage current is also reduced using substrate bias control. As a typical example of an arithmetic circuit, a radix-2 signed-digit full adder using the proposed circuit is implemented in a 0.18- µm CMOS technology. Its dynamic and static power dissipations are reduced to about 79 percent and 14 percent, respectively, in comparison with those of the corresponding binary CMOS implementation at the supply voltage of 1.8 V and the clock frequency of 500 MHz.

The development of energy-efficient neural network hardware using magnetic tunnel junction (MTJ) devices has been widely investigated. One of the issues in the use of MTJ devices is large write energy. Since MTJ devices show stochastic behaviors, a large write current with enough time length is required to guarantee the certainty of the information held in MTJ devices. This paper demonstrates that quantized neural networks (QNNs) exhibit high tolerance to bit errors in weights and an output feature map. Since probabilistic switching errors in MTJ devices do not have always a serious effect on the performance of QNNs, large write energy is not required for reliable switching operations of MTJ devices. Based on the evaluation results, we achieve about 80% write-energy reduction on buffer memory compared to the conventional method. In addition, it is demonstrated that binary representation exhibits higher bit-error tolerance than the other data representations in the range of large error rates.

A new logic-in-memory circuit is proposed for a fine-grain pipelined VLSI system. Dynamic-storage elements are distributed over a logic-circuit plane. A functional pass gate is a key component, where a linear summation and threshold function are merged compactly using charge-storage and charge-coupling effect with a DRAM-cell-based circuit structure. The use of dynamic logic based on pass-transistor network using functional pass gates makes it possible to realize any logic circuits compactly with small power dissipation. As a typical example, a 54-bit pipelined multiplier is implemented by using the proposed circuit technology. Its power dissipation and chip area are reduced to about 63 percent and 72 percent, respectively, in comparison with those of a corresponding binary CMOS implementation under 0.35-µm CMOS technology.

This paper presents a multiple-valued logic-in-memory circuit with real-time programmability. The basic component, in which a dynamic storage function and a multiple-valued threshold function are merged, is implemented compactly by using charge storage and capacitive coupling with a DRAM-cell-based circuit structure under a 0.8-µm CMOS technology. The pass-transistor network using these basic components makes it possible to realize any multiple-valued-inputs binary-outputs logic circuits compactly. As a typical example, a fully parallel multiple-valued magnitude comparator is also implemented by using the proposed DRAM-cell-based pass-transistor network. Its execution time and power dissipation are reduced to about 11 percent and 29 percent, respectively, in comparison with those of a corresponding binary implementation. A prototype chip is also fabricated to confirm the basic operation of the proposed DRAM-cell-based logic-in-memory circuit.

Toward the age of ultra-high-density digital ULSI systems, the development of new integrated circuits suitable for an ultimately fine geometry feature size will be an important issue. Resonant-tunneling (RT) diodes and transistors based on quantum effects in deep submicron geometry are such kinds of key devices in the next-generation ULSI systems. From this point of view, there has been considerable interests in RT diodes and transistors as functional devices for circuit applications. Especially, it has been recognized that RT functional devices with multiple peaks in the current-voltage (I-V) characteristic are inherently suitable for implementing multiple-valued circuits such as a multiple-state memory cell. However, very few types of the other multiple-valued logic circuits have been reported so far using RT devices. In this paper, a new multiple-valued programmable logic array (MVPLA) based on RT devices is proposed for the next-generation ULSI-oriented hardware implementation. The proposed MVPLA consists of 3 basic building blocks: a universal literal circuit, an AND circuit and a linear summation circuit. The universal literal circuit can be directly designed by the combination of the RT diodes with one peak in the I-V characteristic, which is programmable by adjusting the width of quantum well in each RT device. The other basic building blocks can be also designed easily using the wired logic or current-mode wired summation. As a result, a highdensity RT-diode-based MVPLA superior to the corresponding binary implementation can be realized. The device-model-based design method proposed in this paper is discussed using static characteristics of typical RT diode models.

Rapid advances in integrated circuit technology based on binary logic have made possible the fabrication of digital circuits or digital VLSI systems with not only a very large number of devices on a single chip or wafer, but also high-speed processing capability. However, the advance of processing speeds and improvement in cost/performance ratio based on conventional binary logic will not always continue unabated in submicron geometry. Submicron integrated circuits can handle multiple-valued signals at high speed rather than binary signals, especially at data communication level because of the reduced interconnections. The use of nonbinary logic or discrete-analog signal processing will not be out of the question if the multiple-valued hardware algorithms are developed for fast parallel operations. Moreover, in VLSI or ULSI processors the delay time due to global communications between functional modules or chips instead of each functional module itself is the most important factors to determine the total performance. Locally computable hardware implementation and new parallel hardware algorithms natural to multiple-valued data representation and circuit technologies are the key properties to develop VLSI processors in submicron geometry. As a result, multiple-valued VLSI processors make it possible to improve the effective chip density together with the processing speed significantly. In this paper, we summarize several potential advantages of multiple-valued VLSI processors in submicron geometry due to great reduction of interconnection and due to the suitability to locally computable hardware implementation, and demonstrate that some examples of special-purpose multiple-valued VLSI processors, which are a signed-digit arithmetic VLSI processor, a residue arithmetic VLSI processor and a matching VLSI processor can achieve higher performance for real-world computing system.

This paper presents highly reliable multiple-valued one-phase signalling for an asynchronous on-chip communication link under process, supply-voltage and temperature variations. New multiple-valued dual-rail encoding, where each code is represented by the minimum set of three values, makes it possible to perform asynchronous communication between modules with just two wires. Since an appropriate current level is individually assigned to the logic value, a sufficient dynamic range between adjacent current signals can be maintained in the proposed multiple-valued current-mode (MVCM) circuit, which improves the robustness against the process variation. Moreover, as the supply-voltage and the temperature variations in smaller dimensions of circuit elements are dominated as the common-mode variation, a local reference voltage signal according to the variations can be adaptively generated to compensate characteristic change of the MVCM-circuit component. As a result, the proposed asynchronous on-chip communication link is correctly operated in the operation range from 1.1 V to 1.4 V of the supply voltage and that from -50 to 75 under the process variation of 3σ. In fact, it is demonstrated by HSPICE simulation in a 0.13-µm CMOS process that the throughput of the proposed circuit is enhanced to 435% in comparison with that of the conventional 4-phase asynchronous communication circuit under a comparable energy dissipation.

This paper introduces open-wire fault-resilient multiple-valued codes for reliable asynchronous point-to-point global communication links. In the proposed encoding, two communication modules assign complementary codewords that change between two valid states without an open-wire fault. Under an open-wire fault, at each module, the codewords don't reach to one of the two valid states and remains as “invalid” states. The detection of the invalid states makes it possible to stop sending wrong codewords caused by an open-wire fault. The detectability of the open-wire fault based on the proposed encoding is proven for m-of-n codes. The proposed code used in the multiple-valued asynchronous global communication link is capable of detecting a single open-wire fault with 3.08-times higher coding efficiency compared with a conventional multiple-valued code used in a triple-modular redundancy (TMR) link that detects an open-wire fault under the same dynamic range of logical values.

This paper presents a 4-valued content-addressable memory (CAM) for fully parallel template-matching operations in real-time cellular logic image processing with fixed templates. A universal literal is essential to perform a multiple-valued template-matching operation. It is decomposed of a pair of a threshold operation in a CAM cell and a logic-value conversion shared by CAM cells in the same column of a CAM cellular array, which makes a CAM cell function simple. Since a threshold operation together with a 4-valued storage element can be designed by using a single floating-gate MOS transistor, a high-density 4-valued universal-literal CAM with a single-transistor cell can be implemented by using a multi-layer interconnection technology. It is demonstrated that the performance of the proposed CAM is much superior to that of conventional CAMs under the same function.

We have developed a long-range asynchronous on-chip data-transmission link based on multiple-valued single-track signaling for a highly reliable asynchronous Network-on-Chip. In the proposed signaling, 1-bit data with control information is represented by using a one-digit multi-level signal, so serial data can be transmitted asynchronously using only a single wire. The small number of wires alleviates the routing complexity of wiring long-range interconnects. The use of current-mode signaling makes it possible to transmit data at high speed without buffers or repeaters over a long interconnect wire because of the low-voltage swing of signaling, and it leads to low-latency data transmission. We achieve a latency of 0.45 ns, a throughput of 1.25 Gbps, and energy dissipation of 0.58 pJ/bit with a 10-mm interconnect wire under a 0.13 µm CMOS technology. This represents an 85% decrease in latency, a 150% increase in throughput, and a 90% decrease in energy dissipation compared to a conventional serial asynchronous data-transmission link.

An energy-efficient nonvolatile FPGA with assuring highly-reliable backup operation using a self-terminated power-gating scheme is proposed. Since the write current is automatically cut off just after the temporal data in the flip-flop is successfully backed up in the nonvolatile device, the amount of write energy can be minimized with no write failure. Moreover, when the backup operation in a particular cluster is completed, power supply of the cluster is immediately turned off, which minimizes standby energy due to leakage current. In fact, the total amount of energy consumption during the backup operation is reduced by 66% in comparison with that of a conventional worst-case-based approach where the long time write current pulse is used for the reliable write.

A new static storage component, a quaternary flip-flop which consists of two-bit storage elements and three four-level voltage comparators, is proposed for a high-performance multiple-valued VLSI-processor datapath. A key circuit, a differential-pair circuit (DPC), is used to realize a high-speed multi-level voltage comparator. Since PMOS cross-coupled transistors are utilized as not only active loads of the DPC-based comparator but also parts of each storage element, the critical delay path of the proposed flip-flop can be shortened. Moreover, a dynamic logic style is also used to cut steady current paths through current sources in DPCs, which results in great reduction of its power dissipation. It is evaluated with HSPICE simulation in 0.18 µm CMOS that the power dissipations of the proposed quaternary flip-flop is reduced to 50 percent in comparison with that of a corresponding binary CMOS one.

A new common-bus architecture with temporal and spatial parallel access capabilities under wire-resource constraint is proposed to transfer vast quantities of data between modules inside a VLSI chip. Since bus controllers are distributed into modules, the proposed bus architecture can directly transfer data from one module to another without any central bus control unit like a Direct Memory Access (DMA) controller, which enables to reduce communication steps for data transfer between modules. Moreover, when a start address and the number of block data in both source/destination modules are determined at the first step of a data-transfer scheme, no additional address setting for the data transfer is required in the rest of the scheme, which allows us to use all the wire resources as only the "data bus." Therefore, the bus function is dynamically programmed, which results in achieving high throughput of bus communication. For example, in case of a 64-line common bus, it is evaluated that the maximum data throughput in the proposed architecture with dynamic bus-function programming is four times higher than that in the conventional DMA bus architecture with fixed 32-bit-address/32-bit-data buses.

A nonvolatile field-programmable gate array (NV-FPGA), where the circuit-configuration information still remains without power supply, offers a powerful solution against the standby power issue. In this paper, an NV-FPGA is proposed where the programmable logic and interconnect function blocks are described in a hardware description language and are pushed through a standard-cell-based design flow with nonvolatile flip-flops. The use of the standard-cell-based design flow makes it possible to migrate any arbitrary process technology and to perform architecture-level simulation with physical information. As a typical example, the proposed NV-FPGA is designed under 55nm CMOS/100nm magnetic tunnel junction (MTJ) technologies, and the performance of the proposed NV-FPGA is evaluated in comparison with that of a CMOS-only volatile FPGA.

This paper presents FIR digital filters based on stochastic/binary hybrid computation with reduced hardware complexity and high computational accuracy. Recently, some attempts have been made to apply stochastic computation to realization of digital filters. Such realization methods lead to significant reduction of hardware complexity over the conventional filter realizations based on binary computation. However, the stochastic digital filters suffer from lower computational accuracy than the digital filters based on binary computation because of the random error fluctuations that are generated in stochastic bit streams, stochastic multipliers, and stochastic adders. This becomes a serious problem in the case of FIR filter realizations compared with the IIR counterparts because FIR filters usually require larger number of multiplications and additions than IIR filters. To improve the computational accuracy, this paper presents a stochastic/binary hybrid realization, where multipliers are realized using stochastic computation but adders are realized using binary computation. In addition, a coefficient-scaling technique is proposed to further improve the computational accuracy of stochastic FIR filters. Furthermore, the transposed structure is applied to the FIR filter realization, leading to reduction of hardware complexity. Evaluation results demonstrate that our method achieves at most 40dB improvement in minimum stopband attenuation compared with the conventional pure stochastic design.

A multiple-valued current-mode (MVCM) circuit using current-flow control is proposed for a power-greedy sequential linear-array system. Whenever operation is completed in processing element (PE) at the present stage, every possible current source in the PE at the previous stage is cut off, which greatly reduces the wasted power dissipation due to steady current flows during standby states. The completion of the operation can be easily detected using "operation monitor" that observes input and output signals at latches, and that generates control signal immediately at the time completed. Since the wires of data and control signals are shared in the proposed MVCM circuit, no additional wires are required for current-flow control. In fact, it is demonstrated that the power consumption of the MVCM circuit using the proposed method is reduced to 53 percent in comparison with that without current-source control.

A new circuit technique based on pass-gate logic with dynamic supply-voltage and clock-frequency control is proposed for a low-power motion-vector detection VLSI processor. Since the pass-gate logic style has potential advantages that have small equivalent stray capacitance and small number of short-circuit paths, its circuit implementation makes it possible to reduce the power dissipation with maintaining high-speed switching capability. In case the calculation result is obtained on the way of calculation steps, additional power saving is also achieved by combining the pass-gate logic circuitry with a mechanism that dynamically scales down the supply voltage and the clock frequency while maintaining the calculation throughput. As a typical example, a sum of absolute differences (SAD) unit in a motion-vector detection VLSI processor is implemented and its efficiency in power saving is demonstrated.

This paper presents an asynchronous data transfer scheme using 2-color 2-phase dual-rail encoding based on a differential operation and its circuit realization. The proposed encoding enables seamless asynchronous data transfer without inserting a spacer, because each logic value is represented by two kinds of codewords with dual-rail, called "color" data. Since the difference x-x between components of a codeword (x,x) becomes constant in every valid state, the data-arrival state can be detected by calculating the difference x-x. From the viewpoint of circuit implementation, during the state transition, since the dual-rail x and x are defined so as to transit differentially, the compatibility with a comparator using a differential amplifier becomes high, which results in reduction of the cycle time. It is evaluated using HSPICE simulation with a 0.18 µm CMOS technology that communication speed using the proposed dual-rail encoding becomes 1.4 times faster than that using conventional dual-rail encoding.