Abstract:
An oscillator provided with oscillation amplification means equipped with a plurality of oscillation inverters, for amplifying an original oscillation signal supplied from outside and outputting the amplified signal, selection means for generating a selection signal for selectively driving any of a plurality of oscillation inverters, amplification detection means for detecting output amplitude of the oscillation inverter driven in accordance with the selection signal, and generating a detection signal corresponding to the output amplitude, and oscillation indication means for designating the oscillation inverter to be selected on the basis of the detection signal. The oscillation indication means generates a selection indication signal for designating any of a plurality of oscillation inverters in accordance with the detection signal, and the selection means generates the selection signal corresponding to this selection indication signal. Since only the oscillation inverter equipped with a minimum necessary amplification capacity for continuing oscillation is always driven, excessive power consumption is much more restricted than in the prior art, and the oscillator has extremely low power consumption.
Abstract:
A phase detector circuit is provided for correction of operation of a synchronous delay line clock generator. The phase detector includes multiple edge detectors (144, 146, 148). The multiple edge detectors provide an override of any corrective action by the rest of the phase detector to the synchronous delay line output, notwithstanding presence or absence of any phase error of less than 360 DEG , if the phase position of the delay line output signal is off by an integral multiple of 360 DEG . Multiple taps (TAP 2, TAP 9, TAP 14), from daisy-chained or series-connected delay line elements are provided to the multiple edge detectors. The multiple edge detectors compare the edge produced by each such tap against one division of a divided clock signal. Apparatus and method are provided for saving a control signal for a signal-controlled system. A control signal is provided to a multiplexer, which normally produces that control signal. That control signal is digitized and stored by a storage device. The output of the storage device is provided both to the multiplexer and to a comparator. The comparator also receives the output of the multiplexer, and compares the output of the storage device and the multiplexer. The comparator provides a signal to the storage device to increment or decrement the storage device.
Abstract:
The microprocessor unit of a system implemented using metal-oxide-silicon (MOS) technology, is driven by a variable-frequency clock. The microprocessor controls the clock frequency based on the present rate of required microprocessor activity. By driving the microprocessor unit at a lower clock frequency when such activity rate is low, the energy dissipated by the microprocessor unit is reduced due to the MOS power-frequency characteristic.
Abstract:
A process and apparatus for preparing said process for reducing the power consumption of microprocessor-based devices by reducing the frequency of the oscillator (18) governing the logical operation of the microprocessor (10) during periods of use in which system performance is not critical. In one embodiment of apparatus the microprocessor (10) is controlled by a monitor circuit operable with the microprocessor and operated by the variable frequency oscillator. In another embodiment a hardware monitor circuit is utilized and which tracks microprocessor instructions to determine periods of use when performance is not critical. The shift in oscillator speed is mediated by a flip-flop latch circuit (28) connected between one or more clock oscillators and the oscillator input of the controlled microprocessor.
Abstract:
A method for updating a single ported, pipelined Z-buffer (11) where the Z-buffer is updated only after determining the beginning and ending location of a contiguous group of pixels requiring up dating in the Z-buffer. This allows parallel operations in a graphics architecture that utilizes a single ported Z-buffer memory.
Abstract:
In a data processing system (10) which includes a central processing unit (22) capable of operating at a higher speed than memory device (20) associated therewith, method and apparatus provide for sensing and correcting errors generated in the memory device, particularly as a result of operation of the memory at speeds approaching design limits wherein errors may be introduced into the data. Specifically, clock stretching circuitry includes circuitry (36) which operate the central processing unit at a preselected clock rate, circuitry (42, 44, 46) which senses for data errors and circuitry (38) which introduces delays in the clocking of the central processing unit only in the presence of indicated data errors. The clock stretching interval permits error correcting circuitry (48) to correct identified bit errors in the data before the central processing unit accesses the data.
Abstract:
A frequency-voltage conversion circuit (21) is adapted to receive clock (CLK) as an input, and provide as an output a voltage (IVdd) corresponding to the frequency of the clock (CLK). The input/output characteristics of the frequency-voltage conversion circuit (21) are so adjusted that they substantially agree with given input/output characteristics.
Abstract:
A timing signal generator (10) includes a voltage controlled oscillator (VCO) (14), a logic circuit (21), N set circuits (19), N reset circuits (20) and a bistable latch circuit (23). The VCO produces a set of N reference signals (T1-TN) frequency locked to a reference clock signal (T1) and distributed in phase so as to evenly resolve the reference clock period into N intervals. The logic circuit (21) asserts ones of N set signals (SE1-SEN) and N reset signals (RE1-REN) selected by input control words. The timing of leading and trailing edges of pulses of the output timing signal (26) may be controlled with a resolution that is 1/Nth of the period of the reference clock by supplying an appropriate control word sequence to the logic circuit.
Abstract:
Timing is set for DRAM memory access in a computer by polling the DRAM memory banks (11), calculating capacitive load by accessing a prestored table of capacitive load versus DRAM size (fig. 3), and assigning wait states according to calculated capacitive load by accessing a prestored formula. In one embodiment, wait states are assigned in increasing increments for increasing total capacitive load. In an alternative embodiment, timing is assigned bank by bank. Control routines are preferably a part of a system BIOS (14).
Abstract:
A logic circuit equipped with a scan in/out means that is provided with n+p specific scan in/out latch circuits. A first delay means which selectively generates input clocks as delay clock signals of a maximum of m = 2n steps according to selection signals of n bits, is connected in cascade with a second delay means which selectively generates input clock signals as delay clock signals of a minimum step of 1/2p times as small as the width of a minimum step of the clock signals by said first delay means, according to selection signals of p bits. By setting selection data to said n+p scan in/out latch circuits, any delay clock signal can be obtained that is delayed by a given 1/2p step relative to the input clock signal.