Abstract: A parallel search engine able to receive commands via a search instruction input and data words via a search data input. The commands received, which are optionally programmable, control operation of a data dispatch unit and a result dispatch unit. The data words received are sent by the data dispatch unit as search data to a CAM module array made up of CAM modules interconnected by a cascade information bus for comparison against pre-stored comparand databases. The CAM modules of the CAM module array provide search results to the result dispatch unit which generates results, typically multiple in parallel, at a result output. Optionally, multiple of the parallel search engines may be cascaded by connection to an expansion bus to form a mega search engine.

Abstract: A search engine (100) having a controller (112), a memory (114), and a hash pointer unit (110). The memory (114) includes a database of search data and associate content, and the controller (112) uses individual search values to access the memory (114) to obtain individual search results. The controller (112) includes a hash function (116) that generates a hash value from a, typically large, search value into a, typically smaller, hash value that may be a hash collision. The controller (112) converts the hash value into a hash address which is communicated to the hash pointer unit (110), which receives the hash address and provides a hash pointer that is communicated to and used by the memory to look up respective search results. In this manner hash collisions are avoided and the size of the memory (114) is not a function of the degree of multi-way set-associativity used.

Abstract: A search engine having a controller, a memory, and at least one hash-CAM (H-CAM). The memory includes a database of search values and associate content or just associate content. The controller uses search values to access the memory to obtain the search results. The H-CAM includes at least one set of paired hash units and CAM units and at least one logic unit. The CAM units hold values known to cause hash collisions in the paired hash units, and the logic unit prioritizes the hash and CAM unit outputs to address values usable to access the memory and obtain a search result at the controller that is not the result of a hash collision. The H-CAM may optionally include a search data storage to store the search values, so that they need not be stored in the memory, and a comparator to determine and handle new search data based hash collisions. The H-CAM may optionally also be cascaded.

Abstract: A testing station tests integrated circuits and determines if the integrated circuits pass or fail predefined tests. The integrated circuits are placed in a pass bin if the integrated circuits passed the tests, or a fail bin if the integrated circuits failed the tests. A marking station marks identification information on the integrated circuits in the pass bin. The testing and marking stations are both included in a single, integrated tester-marker system.

Abstract: A method is provided for fabricating an array of memory cells for a dynamic random access memory. Each memory cell has an associated capacitor. An array of memory cell transistors is formed and each memory cell transistor has a source, a drain and a gate. The source is coupled to a bit line, and the gate is coupled to a word line. A lower conductive layer is formed over the array of memory cell transistors. The lower conductive layer is electrically coupled to the source of the memory cell transistors. A temporary insulation layer is formed over the lower conductive layer. A portion of the temporary insulation layer and the lower conductive layer are removed to form an electrically separate capacitor bottom plate for each memory cell and an inter-capacitor isolation region. A lateral portion of the temporary insulation layer is removed to form a capacitor sidewall spacing region. A protective layer is formed to fill the inter-capacitor isolation region and the capacitor sidewall spacing region.

Abstract: Each device of a family of removable digital media devices (310, 320, 330, 340, 350 and 360) may be plugged into a host to permits the host to store data in it or to retrieve data from it. The form factors of the digital media devices in the family and the connector system used by the digital media devices are compact for minimizing the volume of space occupied in portable devices and for easy storage. Some embodiments (310, 320, 330, 350 and 360) provide an elongated compact form factor that provides easy and firm grasping for insertion and removal. The digital media devices of the family have respective body portions (312, 322, 332, 342, 352 and 362) preferably of a rigid or semi-rigid material. Preferably, the digital media devices of the family use serial memory requiring few power and signal lines, so that few electrical contacts are required.

Abstract: A Random Access Memory (RAM) with improved memory access time supporting simultaneous transitions of an address signal and a write enable signal while preventing accidental writes. The RAM includes a memory array, an address transition detector and a race detector. Operation of the memory array is controlled by the address signal and a write clock signal. In response to the write clock's read state the memory array reads data from an address represented by the address signal, while the write clock's write state causes the memory array to write data at the address represented by the address signal. The address transition detector and race detector work together to generate the write clock signal. The address transition detector generates an address transition signal when it detects a transition of the address signal from a representation of a first address of the memory array to a representation of a second address of the memory array.

Abstract: The present invention is an antifuse structure comprising an insulation layer between a top conductor and a bottom conductor. The insulation layer has a via. A resistive layer is adjacent the via and a plug is adjacent the resistive layer. The plug is in the via and is also adjacent the top conductor.The present invention also provides a method for fabricating the antifuse on a base. A bottom conductor is deposited on the base. An insulation layer are deposited adjacent the bottom conductor. An antifuse via is etched into the insulation layer. A resistive layer is deposited in the antifuse via. A plug is deposited. The plug extends into the antifuse via. A top conductor is deposited and patterned adjacent the plug.

Abstract: A method is provided for fabricating a stacked capacitor in a storage node (memory cell) of a dynamic random access memory (DRAM) that exceeds the photolithography limit. A DRAM has an array of memory cells and each memory cell has an associated capacitor. An array of memory cell transistors is formed and each memory cell transistor has a source, drain and gate. The drain is coupled to a bit line, and the gate coupled to a word line. A lower conductive layer is formed over the array of memory cell transistors. The lower conductive layer is electrically coupled to the source of each of the memory cell transistors. A protective layer is patterned and formed over a predetermined portion of the lower conductive layer for defining an inter-capacitor isolation region. A portion of the lower conductive layer is removed to form a bottom plate of the capacitor associated with each memory cell, such that a protected portion of the lower conductive layer under the protective layer is removed.

Abstract: A distribution charge pump is disclosed that reduces leakage from a VPP node where a programming voltage (VPP) is provided. The distribution charge pump includes a pump section and a biasing network. The pump section, in response to input signals at 0V or VCC, generates corresponding output signals at 0V or VPP, respectively. Typically, VCC can be between 2V and 5V and VPP can be between 11V and 15V. The pump section includes two n-channel transistors that bootstrap each other to cooperatively pull up the output node to VPP in response to an input signal of VCC. When the charge pump is active, one of the transistors, a native-mode device, transfers charge from the VPP node to an internal node where charge is stored by a capacitor. The biasing network reduces leakage current from the VPP node through the native-mode transistor when the charge pump is inactive.

Abstract: A low voltage charge pump system with a large output voltage range is described. The charge pump system comprises eight charge pump stages, an output stage, and a four phase clock generator. The clock generator generates two sets of four phase shifted signals. The first set of four clock signals are coupled to the first four charge pump stages and have a logic high level of VCC. The second set of clock signals are coupled to the second four charge pump stages and have a logic high level of 2 VCC. Due to the body effect, the negative voltages at the charge pump output stages increases the threshold voltage of a pass transistor which couples the input and output in each charge pump. The larger high voltage level of the second set of clock signals enables the signals to overcome the body effect increased threshold voltages of the pass transistors. The pass transistors are then used to couple negative charge to the next charge pump stage, and positive charge to the preceding charge pump stage.

Abstract: An apparatus and method for minimizing the access time incurred when accessing redundant columns of a dynamic random access memory (DRAM) is herein disclosed. A pair of redundant columns is associated with a defective column. Each pair of redundant columns has a single redundant column decoder that provides access to the column data in the pair of redundant columns. The redundant column decoder is enabled by the column repair circuitry when it receives a column address signal indicating that a defective cell is to be accessed. When a defective column is accessed, the column data from the pair of associated redundant columns is read onto the IO lines as well as the data from the defective column. The three voltages are combined forming an IO signal and the complements of the three voltages are combined forming an IO-BAR signal. The sense amplifier determines the column data value based on the differential between the IO and IO-BAR signals.

Abstract: A high speed memory cell current measurement circuit uses an on-chip reference current circuit that generates a reference current Iref. The reference current circuit includes a first current source transistor. An on-chip current comparison circuit has a second current source transistor that is coupled to the first current source transistor so as to mirror the reference current Iref at a fixed current ratio WR. The current comparison circuit has a current connection path connecting the second current source transistor to a memory cell in the semiconductor memory device whose current is to be compared with Iref/WR. The memory cell is selected from the cells in a memory array using the device's on-chip address decoder circuitry. An on-chip result generation subcircuit, coupled to the current connection path between the second current source transistor and the memory cell, produces a Result signal that indicates whether current flowing through the memory cell is more or less than Iref/WR.

Abstract: The present invention is dynamic pulse generator for generating an output pulse from a first input pulse and a second input pulse, where the output pulse is guaranteed to have a pulse width of at least the pulse width of whichever of the two input pulses has a delayed leading edge with respect to the other. The first input pulse has a first leading edge and a first trailing edge. The second input pulse has a second leading edge and a second trailing edge. The second leading edge is delayed from the first leading edge. An edge detector detects the second leading edge, and outputs a first predetermined level when the second leading edge is detected. The edge detector also detects the first trailing edge and the second trailing edge and outputs a second predetermined level. A latch is responsive to the edge detector and generates a signal indicating that the second leading edge has been detected.

Abstract: Each device of a family of removable digital media devices (310, 320, 330, 340, 350 and 360) may be plugged into a host to permits the host to store data in it or to retrieve data from it. The form factors of the digital media devices in the family and the connector system used by the digital media devices are compact for minimizing the volume of space occupied in portable devices and for easy storage. Some embodiments (310, 320, 330, 350 and 360) provide an elongated compact form factor that provides easy and firm grasping for insertion and removal. The digital media devices of the family have respective body portions (312, 322, 332, 342, 352 and 362) preferably of a rigid or semi-rigid material. Preferably, the digital media devices of the family use serial memory requiring few power and signal lines, so that few electrical contacts are required.

Abstract: A multibank DRAM memory is described having individual row address strobe bar (RASB) and column address strobe bar (CASB) signals. Logically, only one row can be activated in each memory bank at a time and column access can be performed on one memory bank at a time. A token state machine is used to coordinate column access. In a first embodiment, two banks are utilized having respective asynchronous RASB signals transmitted from an external source. In a second embodiment, N DRAM memory banks are utilized having respective asynchronous internal RASB (IRASB) and internal CASB (ICASB) signals. A global RASB signal and a RASB identifier signal (RID) is used to generate the N IRASB and ICASB signals. The RID signal identifies a particular IRASB signal that is to be generated. The token state machine is operated in a round robin manner. In a third embodiment, the N DRAM memory banks are operated in a synchronous manner.

Abstract: A byte-programmable/byte-erasable flash memory system having on-chip counters and secondary storage for word line and bit line disturbance control during program and erase operations. The counters count the numbers of program/erase cycles and compare them with empirically pre-determined counter limits; when the program/erase count exceeds the counter limit, the data then carried in the system are temporarily transferred onto the secondary storage while the memory array is refreshed and the counters are reset. The lifetime of the resulting flash memory system is improved because of decreased erase and program stresses in the memory array.

Abstract: A 5V generator circuit is disclosed that generates a reliable 5V signal for use in integrated circuits from the available VPP or VCC supplies for a wide range of VPP and VCC voltages. When VCC is greater than approximately 4V, the generator circuit generates the 5V signal directly from VCC. When VCC is less than approximately 4V and VPP is greater than approximately 4V but less than about 9V, the generator circuit generates the 5V signal directly from VPP. When VCC is less than approximately 4V and VPP is greater than approximately 9V, the generator produces the 5V signal at approximately half of the voltage level of the VPP signal. When both VCC and VPP are less than about 4V, the 5V signal is generated at the too-low VCC level.

Abstract: A row decoder for a nonvolatile memory having a low-voltage power supply that minimizes the load capacitance presented to a high voltage source without requiring additional circuitry. The row decoder accomplishes this by providing a local decoder having only one input requiring a boosted voltage higher than the power supply voltage. Further, predecoders are used to reduce the number of local decoders that receive the boosted voltage.

Abstract: A voltage pump circuit includes a native MOS device coupled as a charge transfer device (M1) between input and output stage nodes. A parallel-coupled MOS pair (M2, M3) is coupled between drain (input node) and source (output node) of the charge transfer device, in which M3 is configured as a diode. A clock generator outputs at least three non-overlapping phase signals: .phi.1 (which goes high at t1 and low at t6), .phi.2 (which goes high at t3 and low at t4), .phi.3 (which goes low qt t2 and high at t5). The t1 .phi.1 positive transient is AC-coupled to M1's drain, and a smaller fraction of the transient is coupled to M1's gate, precharging M1, which begins to turn-on. The .phi.3 t2 negative transient is AC-coupled to M1's source, increasing M1 gate-source potential, which more fully turns-on M1. The .phi.2 t3 positive transient is coupled to M1's gate, turning-on M1 very hard. A phase clock generator outputting square-wave, same-frequency signals having respective 90.degree.