Abstract: According to some embodiments, an interconnect structure includes write and read structures.

Abstract: A global clock self-timed circuit initiates a precharge pulse in response to which a domino node is precharged. A self-terminating precharge circuit coupled to the global clock self-timed circuit and the domino node terminates the precharge pulse after the domino node has been precharged.

Abstract: A variable width pulse generator. The pulse generator includes a pulse circuit responsive to a reset signal to provide a pulse circuit signal. A variable delay reset loop path, coupled to the pulse circuit, is responsive to the pulse circuit signal to provide the reset signal. A control signal may vary the width of a pulse generated by the circuit by varying the length of a delay associated with the reset loop path. Both a coarse control signal, such as a signal that selectively removes a logic element in the reset loop path, and a fine control signal, such as a signal that controls a tunable delay element in the reset loop path, may be used to adjust the pulse width.

Abstract: A logic structure adapted to receive pulsed active input signals produces a logical output with a very small inherent switching delay. Pull-down transistors and complementary pull-up transistors are ratioed such that the default logical output level remains close to nominal even when the logic structure sinks or sources a DC current. When the pulsed input signals are inactive, no DC current path is enabled.

Abstract: A parallel self-resetting parallel binary adder provides high speed addition and subtraction. The adder combines the advantages of a fully custom design methodology with the higher performance potential of self-resetting complementary metal oxide semiconductor (CMOS) circuits. The adder logic architecture is carry look-ahead with two bit groups and requires six rows of merge logic to calculate the carry out of the Most Significant Bit (MSB). Loading on the critical path of the adder is reduced by moving as many merge blocks as possible to later rows. This allows the fan-out per stage in the critical path to be reduced from around three to two or less. The adder utilizes a bubble pipelined circuit architecture. For the adder, a bubble pipe segment consists of a row of self-resetting circuit blocks.

Abstract: A virtual multi-port RAM (VMPRAM) structure has automatic port sequencing and single-port array density and speed. VMPRAM employs input-triggered, self-resetting macros in a pipelined architecture to provide multiple self-timed on-chip cycles during one machine cycle. The VMPRAM incorporates an SRAM segmented into many input triggered, self-resetting, fast cycling blocks. A timing signal is derived from a selected SRAM block for releasing the next select signals and data to the SRAM blocks. The SRAM block inputs are only the data input bus and the decoded signals needed to select a wordline and a bitline pair, and the SRAM block cycle is only the time needed to provide adequate pulse width for word lines and bitlines. Each SRAM block, and all the circuit blocks in the path to access the SRAM blocks, are input-triggered and self-resetting. The multiple address and data input latches are multiplexed at the driver to the true and complement buses to the SRAM segments, and those buses are self-resetting.

Abstract: A high-performance register file is implemented in a multi-block structure consisting of sub-arrays and associated multiplexing circuits. For a given port, the outputs of all multiplexer circuits are dotted together to form a single global output. The multiplexer circuits may be completely external to the cells ("standard" approach), or distributed and integrated into the cells ("alternate" approach). The register cells arranged as such, may or may not contain extra latches, pass gates, and controls arranged so that the file may be fully tested via LSSD.

Abstract: A fast comparator circuit, including a plurality of first switches operating in parallel. A first data bit from a first data word is input into a first input of each first switch, and a corresponding second data bit from a second data word is respectively input into a second input of each first switch. Each first switch provides a first logic state output when the first data bit matches the corresponding second data bit or a second logic state output when the first data bit does not match the second data bit. A plurality of second switches receive the respective logic state outputs and produce a combined output, indicating an all match or a mismatch, to a third switch combination connected to a first branch node and a second branch node to create a first voltage difference between the first and second branch nodes when an all match output results and a second voltage difference between the first and second branch node when a mismatch output results.

Abstract: A virtual multi-port RAM (VMPRAM) structure has automatic port sequencing and single-port array density and speed. VMPRAM employs input-triggered, self-resetting macros in a pipelined architecture to provide multiple self-timed on-chip cycles during one machine cycle. The VMPRAM incorporates an SRAM segmented into many input triggered, self-resetting, fast cycling blocks. A timing signal is derived from a selected SRAM block for releasing the next select signals and data to the SRAM blocks. The SRAM block inputs are only the data input bus and the decoded signals needed to select a wordline and a bitline pair, and the SRAM block cycle is only the time needed to provide adequate pulse width for word lines and bitlines. Each SRAM block, and all the circuit blocks in the path to access the SRAM blocks, are input-triggered and self-resetting. The multiple address and data input latches are multiplexed at the driver to the true and complement buses to the SRAM segments, and those buses are self-resetting.

Abstract: A circuit incorporating a clocked first stage and unlocked second and third stages with amplification for driving capacitance loads with the output nodes each coupling its changed state back to the earlier stages to reset them independent of the clock. The circuit may use complementary metal-oxide semiconductor devices of various gate widths.

Abstract: An AND gate includes first and second opposite-type field effect transistors, each including first and second conduction path terminals and a control electrode. The gate's output terminal is connected, in common, to the second conduction path terminals of the transistors. A first logic input is connected to the first conduction path terminal of the first transistor and a second logic input is connected in common to the control electrode of the first transistor and to the first conduction path terminal of the second transistor. A third logic input is applied to the control electrode of the second transistor. In a standby state prior to the application of logic signals all three logic inputs are in the same state. This assures no conduction of logic signals, while conditioning the gate for rapid selection when logic signals are applied. Subsequently, logic signals are applied to the inputs with the third input being the complement level of the logic signal on the first input.

Abstract: CMOS ECL drive circuit for providing regulated ECL logic levels. A CMOS logic circuit is connected by parallel N channel and P channel devices to serially connected N and P channel devices. The serially connected N and P channel devices are connected across a CMOS power supply with gate connections connected to the logic circuit. The parallel devices provide a regulating feedback current to one of the serially connected P channel and N channel devices during each of first and second ECL logic states. The feedback current effectively controls the bias on the gate connections of the serially connected P and N channel devices. The voltage at the junction of the serially connected P and N channel devices is regulated by each of the parallel connected devices.

Abstract: A semiconductor random access memory chip wherein the cycle time is less than the access time for any combination of read or write sequence. The semiconductor random access memory chip is partitioned into relatively small sub-arrays with local decoding and precharging. The memory chip operates in a pipelined manner with more than one access propagating through the chip at any given time and wherein the cycle time is limited by sub-array cycles wherein the cycle time is less than the access time for a memory chip having cycle times greater than access times for accesses through the same sub-array. The memory chip also incorporates dynamic storage techniques for achieving very fast access and precharge times.

Abstract: A high performance decoder/driver circuit for a semiconductor memory having A1 to AN (true) and A1 to AN (complement) address lines for receiving A1 to AN address bit signals thereon from internal address buffers. A .phi.PC line is included for receiving a .phi.PC precharge clock signal thereon and a .phi.R line is provided for receiving a .phi.R reset clock signal thereon. The decoder/driver circuit includes an OR decoder means having a plurality of transistor switching devices connected to A1 to AN-1 or A1 to AN-1 of the true and complement address lines for the AN to AN-1 address bits for producing a high or low level signal on an OR decoder output node depending on the address bits state.

Abstract: A circuit means for interfacing between small emitter-coupled-logic (ECL) circuit voltage levels and larger field effect transistor (FET) circuits voltage level. The circuit interface means includes a source-follower stage wherein a first transistor device is ratioed relative to a second transistor device, so that a high percentage of an input voltage level signal to the first transistor device appears at a node between the first and second transistor devices, having been level shifted downward by greater than or equal to an n-channel threshold voltage. The percentage is enhanced by applying the complement of the input voltage level signal to the gate of the second transistor device. A gain stage is connected to the source-follower stage and includes third and fourth transistor devices wherein gain is developed by applying the level shifted input signal to the source of the fourth transistor device and the complement of the input signal directly to the gate of the fourth transistor device.

Abstract: A heterojunction semiconductor is provided where the carrier transport dimension is governed by a layer thickness and where the characteristics of the materials self-limit process steps. A field effect transistor is provided wherein the work function is matched across regions to reduce limits on the channel dimension. A vertical transistor is provided wherein a vertical web is formed with precise thickness governed by electrolytic etching using photogenerated carrier current.

Abstract: An energy discriminator is provided wherein energy, entering through a receiving surface into a multilayer semiconductor monocrystalline body is converted into hole-electron pair carriers in different particular energy responsive layers and the electrons thereof are collected in potential wells that are asymmetric to electron flow associated with the particular layer.

Abstract: Optical-to-electrical conversion is accomplished using an undoped region bounded by a tunneling junction of the order of the mean free path of an electron. A number of regions are assembled in series with larger thickness away from the light incident surface. The thickness and doping of the regions for maximum effectiveness in monochromatic light are tailored to produce similar quantities of carriers from the light. A nine section GaAs structure with 50 .ANG. n.sup.+ and p.sup.+ tunneling bounding regions has a 90% quantum efficiency and delivers a 5 volt output with a 0.35 picosecond transit time.

Abstract: Optical-to-electrical logic operations may be performed employing as each logic variable a different light wavelength and providing an optical-to-electrical semiconductor converter such that each particular wavelength responsive optical energy receiving region is an updoped region bounded by a thin tunneling junction having a thickness of the order of the mean free path of a carrier in the tunneling region.

Abstract: A highly efficient monolithic direct bandgap lattice accommodated semiconductor structure in which an input signal is converted to photons on one side of an insulating region and the photons are reconverted on the opposite side of the insulating region to an output signal in a multi-oblique segment region. The structure converts AC to DC, AC to AC and DC to DC signals and is adapted to the efficient V-groove multijunction solar cell.