The prospects for ultrahigh-speed VLSI GaAs digital logic

Recent advances of GaAs integrated circuit fabrication technology have made possible the demonstration of ultrahigh performance GaAs digital ICs with up to 64 gate MSI circuit complexities and with gate areas and power dissipations sufficiently low to make VLSI circuits achievable. The authors evaluate, based on the current state of GaAs IC technology and the fundamental device physics involved, the prospects of achieving an ultrahigh-speed VLSI GaAs IC technology. GaAs IC fabrication and logic circuit approaches is reviewed. The experimental performance results are compared for the leading GaAs logic circuit approaches, both for simple ring oscillators and for more complex sequential logic circuits.     

Design of a Standard Floating-Point Chip

Some aspects of design of a VLSI floating-point chip, which provides the WE/spl registered/32100 microprocessor with math acceleration capabilities, are described. The chip is implemented in 1.75-/spl mu/m twin-tub CMOS II technology [2] and contains 140 000 transistors.     

A 10 K-gate 950 MHz CML demonstrator circuit made with a 1-μmtrench-isolated bipolar silicon technology

A quad 512-b static shift register consuming 1.8 mW/stage designed to demonstrate the capabilities of an advanced bipolar silicon technology is discussed. The process uses 1-μm lithography, trench isolation, polyemitter transistors, polysilicon resistors, and polycide layer for local interconnections. This VLSI circuit (over 35 K transistors, 86-mm² chip) has been implemented on a sea-of-cells structure. An appropriate scheme has been used for the clock distribution. The experimental results show operation at a clock frequency up to 950 MHz     

Noise-margin limitations on gallium-arsenide VLSI

Two factors which limit the complexity of GaAs MESFET VLSI circuits are considered. Power dissipation sets an upper complexity limit for a given logic circuit implementation and thermal design. Uniformity of device characteristics and the circuit configuration determines the electrical functional yield. Projection of VLSI complexity based on these factors indicates that logic chips of 15000 gates are feasible with the most promising static circuits if a maximum power dissipation of 5 W per chip is assumed. While lower power per gate and therefore more gates per chip can be obtained by using a popular E/D FET circuit, yields are shown to be small when practical device parameter tolerances are applied. Further improvements in materials, devices, and circuits will be needed to extend circuit complexity to the range currently dominated by silicon     

The prospects for ultrahigh-speed VLSI GaAs digital logic

Recent advances in the state of GaAs integrated circuit fabrication technology have made possible the demonstration of ultrahigh performance (tau_{d} sim 100ps) GaAs digital IC's with up to 64 gate MSI circuit complexities and with gate areas and power dissipations sufficiently low to make VLSI circuits achievable. It is the purpose of this paper to evaluate, based on the current state of GaAs IC technology and the fundamental device physics involved, the prospects of achieving an ultrahigh-speed VLSI GaAs IC technology. The paper includes a performance comparison analysis of Si and GaAs FET's and switching circuits which indicates that, for equivalent speed-power product operation, GaAs IC's should be about six times faster than Si IC's. The state of the art in GaAs IC fabrication and logic circuit approaches is reviewed, with particular emphasis on those approaches which are LSI/VLSI compatible in power and density. The experimental performance results are compared for the leading GaAs logic circuit approaches, both for simple ring oscillators and for more complex sequential logic circuits (which have demonstrated equivalent gate delays as low astau_{d} = 110ps).     

A VLSI memory management chip: design considerations and experience

The design of a VLSI memory measurement chip which provides the WE 32001 microprocessor with an extensive set of memory management capabilities is described. The chip is implemented in 1.75 /spl mu/m twin-tub CMOS II technology and contains 92000 transistors. Highlights of the technology are the use of twin tubs for independently optimized n- and p-channel transistors, local oxidation and self-aligned channel-stops for parasitic field protection, and the use of an n/SUP +/ substrate for latchup protection. In addition, a composite layer of TaSi over n/SUP +/ polysilicon is used to achieve a fivefold reduction in sheet resistance over the conventional n/SUP +/ polysilicon. Electrical channel lengths for n-channel and p-channel transistors are nominally 1.5 /spl mu/m.     

MARVLE: a VLSI chip for data compression using tree-based codes

Describes the architecture and design of a CMOS VLSI chip for data compression and decompression using tree-based codes. The chip, called MARVLE, implements a memory-based architecture for variable length encoding and decoding based on tree-based codes. The architecture is based on an efficient scheme of mapping the tree representing any binary code onto a memory device. A prototype 2-mm CMOS VLSI chip has been designed, verified, and fabricated by the MOSIS facility. The chip has a 512×12 static RAM with an access time of 4 ns and logic circuitry for compression as well as decompression. The chip occupies a silicon area of 6.8 mm×6.9 mm and consists of 49695 transistors. The prototype chip yields a compression rate of 95.2 Mb/s and a decompression rate of 60.6 Mb/s with a clock rate of 83.3 MHz. The VLSI hardware can be used to implement the JPEG baseline compression scheme     

Design methodology for full custom CMOS microcomputers

A highly structured design methodology is necessary to be successful in the design of VLSI integrated circuits with more than 100000 transistors on a chip. Such a methodology is described: it is based on the regularity of the circuit architecture with an associated chip floor plan and on a new layout technique named metal oriented layout.This methodology has been tested with the design of a 13500 MOS microcomputer. From the instruction set and through different levels of instruction interpretation, the architecture and associated chip floor plan are generated. The detailed logic design is made directly in symbolic layout with the chip floor plan in mind.The proposed design methodology can be best appreciated by the short development time and small chip area required for the designed 13500 MOS microcomputer.     

Design Considerations for Single-Chip Computers of the Future

In the mid 1980's it will be possible to put a million devices (transistors or active MOS gate electrodes) onto a single silicon chip. General trends in the evolution of silicon integrated circuits are reviewed and design constraints for emeging VLSI circuits are analyzed. Desirable architectural features in modem computers are then discussed and consequences for an implementation with large-scale integrated circuits are investigated. The resulting recommended processor design includes features such as an on-chip memory hierarchy, multiple homogeneous caches for enhanced execution parallelism, support for complex data structures and high-level languages, a flexible instruction set, and communication hardware. It is concluded that a viable modular building block for the next generation of computing systems will be a self-contained computer on a single chip. A tentative allocation of the one million transistors to the various functional blocks is given, and the result is a memory intensive design.     

A 64-point Fourier transform chip for video motion compensationusing phase correlation

Details of a new low power fast Fourier transform (FFT) processor for use in digital television applications are presented. This has been fabricated using a 0.6-μm CMOS technology and can perform a 64 point complex forward or inverse FFT on real-time video at up to 18 Megasamples per second. It comprises 0.5 million transistors in a die area of 7.8×8 mm² and dissipates 1 W. The chip design is based on a novel VLSI architecture which has been derived from a first principles factorization of the discrete Fourier transform (DFT) matrix and tailored to a direct silicon implementation     

A VLSI chip set for a large-scale parallel inference machine: PIM/m

The authors present three VLSI chips-a processor (PU) chip, a cache memory (CU) chip, and a network control (NU) chip-for a large-scale parallel inference machine. The PU chip has been designed to be adapted to logic programming languages such as PROLOG. The CU chip implements a hardware support called `trial buffer' which is suitable for the execution of the PROLOG-like languages. The NU chip makes it possible to connect 256 processing elements in a mesh network. The parallel inference machine (PIM/m) runs a PROLOG-like network-based operating system called PIMOS as well as many applications and has a peak performance of 128 mega logical inferences per second (MLIPS). The PU chip containing 384000 transistors is fabricated in a 0.8 μm double-metal CMOS technology. The CU chip and the NU chip contain 610000 and 329000 transistors, respectively. They are fabricated in a 1.0 μm double-metal CMOS technology. A cell-based design method is used to reduce the layout design time     

CMOS layout design of the hysteresis McCulloch-Pitts neuron

A McCulloch-Pitts neuron is the simplified neuron model which has been successfully used for many optimisation problems. The neural network with the hysteresis property can suppress the oscillatory behaviours of neural dynamics so that the convergence time is shortened. In this paper, digital CMOS layout design of the hysteresis McCulloch-Pitts neuron is presented. Based on simulation results using the hysteresis McCulloch-Pitts binary neuron model, a 6-bit fixed point 2's complement arithmetic was adopted for the calculation of the input U of each neuron. Each neuron needs 204 transistors and requires a 399 lambda *368 lambda layout area using the MOSIS scalable CMOS/bulk (SCMOS) VLSI technology with 2 mu m rule of P well, double level metal. Layout design of the hysteresis McCulloch-Pitts neuron chip was completed, and fabrication of the chip and the design for the test circuit for the fabricated CMOS VLSI chip are underway at present.<>     

Design of a 32 bit microprocessor, TX1

Implementation of the TX1 VLSI microprocessor is described. Particular emphasis is placed on the design method, which meets the requirements of short design time with reasonable chip size. A one-phase clock system, which is a better solution for high-speed operation but requires careful design for evading the skew problem, is discussed. Design for testability is embedded in the chip. The TX1 is fabricated with a 1.0 μm two-layer metal CMOS process. The chip contains 450 K transistors in a 10.89×10.27 mm² die     

Modeling technology impact on cluster microprocessor performance

The growing speed gap between transistors and wire interconnects is forcing the development of distributed, or clustered, architectures. These designs partition the chip into small regions with fast intracluster communication. Longer latency is required to communicate between clusters. The hardware and/or software are responsible for scheduling instructions to clusters such that critical path communication occurs within a cluster. This paper presents GENEric SYstems Simulator (GENESYS), a technology modeling tool that captures a broad range of materials, device, circuit, and interconnect parameters across current and future semiconductor technology. This tool is used to explore the relationship between key technology parameters (intercluster wire delay and transistor switching delay) and key architecture parameters (superscalar versus multithreaded instruction dispatch, and value prediction support). GENESYS is used to predict intercluster latencies as VLSI technology advances. The study provides quantitative data showing how conventional superscalar performance is degraded with increasing wire latency. Threaded designs are more tolerant to wire delay. Optimal thread size changes with advancing VLSI technology, suggesting a highly adaptive architecture. Value prediction is shown to be useful in all cases, but provides more benefit to the multithreaded design.     

A dual voltage-frequency VLSI chip for image watermarking in DCT domain

In this brief, we present a new VLSI architecture that can insert invisible or visible watermarks in images in the discrete cosine transform domain. The proposed architecture incorporates low-power techniques such as dual voltage, dual frequency, and clock gating to reduce the power consumption and exploits pipelining and parallelism extensively in order to achieve high performance. The supply voltage level and the operating frequency are chosen for each module so as to maintain the required bandwidth and throughput match among the different modules. A prototype VLSI chip was designed and verified using various Cadence and Synopsys tools based on TSMC 0.25-/spl mu/m technology with 1.4 M transistors and 0.3 mW of estimated dynamic power.     

An advanced bipolar-MOS-I2L technology with a thin epitaxial layer for analog-digital VLSI

A novel Bi-MOS technology, Advanced Bipolar CMOS (ABC), is proposed. Bipolar transistors (n-p-n, p-n-p, I²L) and MOS transistors (both n- and p-channel) have been successfully fabricated on the same chip with no decrease in performance by using a 3-µm design rule. Thin epitaxial layer (leq 2 microm) is used in order to obtain small-size high-performance (3-GHz) bipolar devices. Device size is reduced by using a shallow junction and self-aligning technique. n-channel MOS transistors are formed in p-well regions designed to reach p-type substrate, and p-channel MOS transistors are formed in epitaxial layer with an n⁺buried layer. This technology has the potential for monolithic multifunctional analog-digital VLSI.     

Automatic hardware synthesis

The complexity of the circuit that can fit on an integrated circuit (IC) chip has reached the level of a million transistors with the advent of Very-large-Scale Integration (VLSI). Several automatic synthesis systems have evolved that "aid" the human designer in managing this complexity. This paper surveys such efforts. The synthesis is viewed as the process of transforming a high-level design specification into a lower level design specification that includes more structural details, leading to the physical design of the IC. The characteristics of ten automatic synthesis systems are summarized.     

An ultra-high-speed direct digital frequency synthesizer implemented in GaAs HBT technology

This paper presents a 10-GHz 8-bit direct digital synthesizer(DDS)microwave monolithic integrated circuit implemented in 1 μm GaAs HBT technology.The DDS takes a double-edge-trigger(DET)8-stage pipeline accumulator with sine-weighted DAC-based ROM-less architecture,which can maximize the utilization ratio of the GaAs HBT's high-speed potential.With an output frequency up to 5 GHz,the DDS gives an average spurious free dynamic range of23.24 dBc through the first Nyquist band,and consumes 2.4 W of DC power from a single-4.6 V DC supply.Using1651 GaAs HBT transistors,the total area of the DDS chip is 2.4 × 2.0 mm2.     

A 64-b microprocessor with multimedia support

A 167 MHz 64 b VLSI CPU chip is described. The chip executes a 333-MFLOPS (peak) with an estimated system performance of 270SPECint92/380SPECfp92 (@167 MHz, 2 MB E-cache). The 17.7×17.8 mm die is fabricated with a 0.5 micron CMOS technology with four metal layers and contains 5.2 M transistors. The superscalar processor is capable of sustaining an execution rate of four instructions per cycle even in the presence of conditional branches and cache misses. Four fully pipelined 8×16 b multipliers and four single-cycle latency 16 b adders combine to speed up image processing, 2-D, 3-D graphics, video compression/decompression by up to an order of magnitude. High clock speed was obtained by the use of delayed reset logic, a new register file design; and novel comparators. Strict design methodology allowed fully functional first silicon which met all speed targets. The power dissipation of the chip is 28 W     

Resistive interpolation biasing: a technique for compensatinglinear variation in an array of MOS current sources

A new technique called resistive interpolation biasing for accurately biasing a large number of analog cells on a VLSI chip is presented. Variations in oxide thickness, mobility, doping concentration, etc., cause inaccuracies in current ratios of two identically biased transistors if they are placed sufficiently far apart on a chip. The proposed technique compensates for these inaccuracies without using any sampling or switching. The technique has been verified using a 2 μm n-well CMOS process. Measurements show a factor of 3 improvement in terms of current ratio accuracy when the resistive interpolation technique is used. The circuit can be implemented with a small chip area and low power dissipation. This technique finds applications where extensive current duplication over a large area is required (e.g., analog memories, D/A converters, continuous-time filters, imaging arrays, neural networks, and fuzzy logic systems)