Abstract: There is provided programmability, automatic error correction, flexible inter-path implementation for a bit error rate estimator. The bit error rate estimator uses an initialization configuration for a number of memory elements of a reference pattern generator in such a way as to eliminate the need for an N bit counter. In addition, the bit error rate estimator allows for two error correction options that may be pre-grammed into the system. One option allows the system administrator to monitor the errors in the system and take steps as deemed necessary. Another option generates a new reference pattern if the number of errors detected reaches a predetermined maximum amount. Thus, the bit error rate estimator provides automatic error correction, programmability, and flexible inter-path implementation.

Abstract: A high-isolation, low-power high-speed multiplexer circuit suitably includes a buffer stage and a current steering tree stage. By employing common select lines for both stages of the circuit, both the input buffer and the deselected channel provide cumulative isolation for the deselected channels.

Abstract: Fully connected multiple FCU-based architectures reduce requirements for Tag SRAM size and memory read latencies. A preferred embodiment of a symmetric multiprocessor system includes a switched fabric (switch matrix) for data transfers that provides multiple concurrent buses that enable greatly increased bandwidth between processors and shared memory. A high-speed point-to-point Channel couples command initiators and memory with the switch matrix and with I/O subsystems.

Abstract: A switch suitable for use in high-bandwidth environments is disclosed. The switch eliminates the need for inter-stage jitter compensation by determining the timing signals associated with each data input and then re-timing the data based upon the timing signals at the switch output. Bandwidth is conserved by routing timing signals through a multiplexer that preferably determines the difference between the timing signal and a reference signal, combines the difference signal with other difference signals calculated for other data inputs, and then transmits the multiplexed difference signals to a demultiplexer. Suitable multiplexing schemes include time division multiplexing, wavelength division multiplexing, code division multiple access (CDMA) multiplexing, as well as various combinations of suitable multiplexing methods.

Abstract: A computer system couples multiple function units using channels having full-duplex, low power, point-to-point interconnect. Each function unit couples to the channel via a Channel Interface Block (CIB). The CIB includes a transmitter and a receiver. The receiver includes an integrating sampling capacitor, pass-gates having particular resistive characteristics, an auto-zero inverter, and a set of inverter stages for squaring the output of the inverter. These components are used to implement sampled-data methods and structures that perform received data extraction from the full-duplex channel signal.

Abstract: A preferred embodiment of a symmetric multiprocessor system includes a switched fabric (switch matrix) for data transfers that provides multiple concurrent buses that enable greatly increased bandwidth between processors and shared memory. A high-speed point-to-point Channel couples command initiators and memory with the switch matrix and with I/O subsystems. Each end of a channel is connected to a Channel Interface Block (CIB). The CIB presents a logical interface to the Channel, providing a communication path to and from a CIB in another IC. CIB logic presents a similar interface between the CIB and the core-logic and between the CIB and the Channel transceivers. A channel transport protocol is is implemented in the CIB to reliably transfer data from one chip to another in the face of errors and limited buffering.

Abstract: Signal processing techniques are applied to data rates at state-of-the-art circuit speeds (presently 1.6 Gbit/sec) by carrying out the signal flow graph of a cannonical FIR filter algorithm using hybrid analog and digital circuit techniques. A plurality of digital to analog converters (DACs) generate analog currents that are the analogue of the tap coefficients of the FIR filter model. The DACs are used as programmable current sources for the tail current sources of respective differential pair stages. Differential delay signals that are the analogue of the FIR delay-line tap signals are connected to the inputs of respective ones of the differential pair stages. The drains of the input devices of the differential pair stages are connected in parallel to common complementary load circuits. The delay signals act to steer the tap coefficient currents to one or the other of the common load circuits.

Abstract: A preferred embodiment of a symmetric multiprocessor system includes a switched fabric (switch matrix) for data transfers that provides multiple concurrent buses that enable greatly increased bandwidth between processors and shared memory. A Transaction Controller, Transaction Bus, and Transaction Status Bus are used for serialization, centralized cache control, and highly pipelined address transfers. The shared Transaction Controller serializes transaction requests from Initiator devices that can include CPU/Cache modules and Peripheral Bus modules. The Transaction Bus of an illustrative embodiment is implemented using segmented buses, distributed muxes, point-to-point wiring, and supports transaction processing at a rate of one transaction per clock cycle.

Abstract: The system and method for operating a cache-coherent shared-memory multiprocessing system is disclosed. The system includes a number of devices including processors, a main memory, and I/O devices. Each device is connected by means of a dedicated point-to-point connection or channel to a flow control unit (FCU). The FCU controls the exchange of data between each device in the system by providing a communication path between two devices connected to the FCU. The FCU includes a snoop signal path for processing transactions affecting cacheable memory and a network of signal paths that are used to transfer data between devices. Each signal path can operate concurrently thereby providing the system with the capability of processing multiple data transactions simultaneously.

Abstract: An improved flip-flop circuit exhibits a higher phase margin than conventional flip-flop circuits without a substantial increase in operating power. The flip-flop circuit includes a master latch circuit operatively coupled to a slave latch circuit. The flip-flop circuit uses any number of techniques to delay the hold-to-sample transition of the slave latch circuit relative to the sample-to-hold transition of the master latch circuit. The delay enables the flip-flop circuit to better tolerate clock/data timing alignment issues. In a first embodiment, the slave clock signal is delayed relative to the master clock signal. In a second embodiment, the master clock signal buffer is unbalanced such that its duty cycle is skewed to produce unequal sample and hold periods. In a third embodiment, the master latch circuit is unbalanced to create an unequal delay associated with the sampling and holding periods.

Abstract: In a preferred embodiment, the invention uses an 8-to-1 data serialization circuit in the transmitter to convert 80-bit parallel 200 MHz data to 10-bit parallel 1.6 Mb/s date. On the receiver side, data are captured using a forwarded clock and de-serialized. A single global DLL generates 16 master phases without reference to the word boundaries of data being transmitted. These 16 unreferenced phases are input to a phase rotator that, via a series of calibration steps, maps the unreferenced phases into named phases, and in doing so references the phases to the word boundary of the data being transmitted over the slowest data line of the parallel channel. The named phases are then input to a data interpolator in each receiver, which generates 16 local phases. The 16 local phases correspond to the data-bit centers and data-bit edges for each of the 8 bits transferred per major channel clock period.

Abstract: Methods and apparatus for processing cells in an asynchronous transfer mode (ATM) communication system. A group of bits comprises a primary scoreboard indicative of the scheduling status for cell time slots in a periodic container of cells, with each bit indicating the availability of a corresponding cell time slot. A connection identifier (ID) table is maintained with each location in the table corresponding to one of the cell time slots and thus a single primary scoreboard bit. A cell scheduling instruction specifies a connection ID for a virtual connection on an ATM transmission link. A processor searches the primary scoreboard until a bit corresponding to an available cell time slot is located, reserves the located cell slot by setting the corresponding bit, and stores the connection ID in the corresponding location in the connection ID table. A cell servicing instruction specifies an address in the connection ID table.

Abstract: An ATM system transmits different types of cells (data, forward resource management (RM) and backward RM) from station A through switch(es) to station B. Different fields in an Available Bit Rate (ABR) table provide controls over the rate of such cell transmissions. First particular field values in such table control the selection of successive ones of cell decision blocks which determine the type of cell to be transmitted. Second particular field values in such table control the selection of one of a plurality of entries in an exponent table which also provides other parameter values controlling the generation of an explicit rate. Third particular field values in the ABR table control the selection of an individual one of a plurality of rate decision blocks each indicating an individual rate of cell transmission from the station A to the station B.

Abstract: A preferred embodiment of a symmetric multiprocessor system includes a switched fabric (switch matrix) for data transfers that provides multiple concurrent buses that enable greatly increased bandwidth between processors and shared memory. A Transaction Controller, Transaction Bus, and Transaction Status Bus are used for serialization, centralized cache control, and highly pipelined address transfers. The shared Transaction Controller serializes transaction requests from Initiator devices that can include CPU/Cache modules and Peripheral Bus modules. The Transaction Bus of an illustrative embodiment is implemented using segmented buses, distributed muxes, point-to-point wiring, and supports transaction processing at a rate of one transaction per clock cycle.