摘要:
Systems and methods for reducing the power consumption of some combinations of logic gates by reducing the number of unnecessary transitions that are made by logic gates that do not affect the output of the logic. In one embodiment, a modified exclusive-OR (XOR) gate is coupled to a modified multiplexer. The XOR gate has two inputs, Ain, and Bin, and an output, XORout, which is provided as an input to the multiplexer. Another input to the multiplexer is Cin. A select signal is input to the multiplexer to select either Cin or XORout to be provided at the output of the multiplexer. If XORout is selected, the XOR gate operates in a first mode in which it functions as a normal XOR gate. If Cin is selected, the XOR gate operates in a second mode in which the XOR gate uses less power than when the XOR gate operates normally.
摘要:
Systems and methods for reducing the power consumption of some combinations of logic gates by reducing the number of unnecessary transitions that are made by logic gates that do not affect the output of the logic. In one embodiment, a modified exclusive-OR (XOR) gate is coupled to a modified multiplexer. The XOR gate has two inputs, Ain and Bin, and an output, XORout, which is provided as an input to the multiplexer. Another input to the multiplexer is Cin. A select signal is input to the multiplexer to select either Cin or XORout to be provided at the output of the multiplexer. If XORout is selected, the XOR gate operates in a first mode in which it functions as a normal XOR gate. If Cin is selected, the XOR gate operates in a second mode in which the XOR gate uses less power than when the XOR gate operates normally.
摘要:
Systems and methods for preventing the corruption of a CAM lookup result when a lookup in the CAM and a write to the CAM are concurrently executed. In one embodiment, a tag value is clocked into a tag latch simultaneously with a data value being clocked into a data latch. The tag value and initial CAM element values begin propagating through comparison logic. After a delay, the data value is written from the data latch to a CAM element. After the tag value and initial CAM values propagate through the comparison logic to produce comparison outputs, but before the newly written data value propagates through the comparison logic and changes the comparison outputs, the comparison outputs are latched. The comparison outputs can then be processed as if the data values in the CAM elements had not been changed by the write operation.
摘要:
Systems and methods for reducing the power consumption associated with the capacitance of sections of a multiplexer are disclosed. At each cycle, a timing signal is selectively sent only to sections of the multiplexer that include active logic. A plurality of control signals is received for processing by a corresponding plurality of input selection circuits. A plurality of additional inputs corresponding to the plurality of input selection circuits may also be received. In one embodiment, each input selection circuit is configured to output a corresponding input signal if a corresponding control signal is asserted and a timing signal is made available to the input selection circuit. To avoid unnecessary power consumption associated with the capacitance of various portions of the multiplexer, the timing signal is only asserted to a portion of the multiplexer at any given clock cycle according to the values of the control signals.
摘要:
A method, an apparatus, and a computer program are provided for distributing data in a high speed processing unit. Traditionally, true readout data from multiport register files are inverted multiple times when transmitting the readout to data latches, located at multiple physical layers. The inversion of the readout data can be boost the signals and provide the proper true or complement data to the data latches. To reduce the number of inverters, the register files are configured to output true and complement signals. Therefore, power consumption and area are reduced with the elimination of the inverters.
摘要:
An apparatus, a method, and a computer program product are provided for time reduction and energy conservation during address distribution in a high speed memory macro. To address these concerns, this design divides the typical data arrays into sets of paired subarrays, divides the conventional memory address latches into separate sets, and interposes one set of memory address latches between each pair of subarrays. Therefore, time is saved because the address signals have less wire length to travel and energy is saved because only one set of address latches needs to be powered on for each transmission.
摘要:
An apparatus, a method, and a computer program product are provided for time reduction and energy conservation during address distribution in a high speed memory macro. To address these concerns, this design divides the typical data arrays into sets of paired subarrays, divides the conventional memory address latches into separate sets, and interposes one set of memory address latches between each pair of subarrays. Therefore, time is saved because the address signals have less wire length to travel and energy is saved because only one set of address latches needs to be powered on for each transmission.
摘要:
A domino circuit configuration includes a precharge transistor coupled to a discharge transistor, wherein the precharge transistor and the discharge transistor are not on simultaneously.
摘要:
Systems and methods for preventing the corruption of a CAM lookup result when a lookup in the CAM and a write to the CAM are concurrently executed. In one embodiment, a tag value is clocked into a tag latch simultaneously with a data value being clocked into a data latch. The tag value and initial CAM element values begin propagating through comparison logic. After a delay, the data value is written from the data latch to a CAM element. After the tag value and initial CAM values propagate through the comparison logic to produce comparison outputs, but before the newly written data value propagates through the comparison logic and changes the comparison outputs, the comparison outputs are latched. The comparison outputs can then be processed as if the data values in the CAM elements had not been changed by the write operation.
摘要:
Systems and methods for reducing the power consumption associated with the capacitance of sections of a multiplexer are disclosed. At each cycle, a timing signal is selectively sent only to sections of the multiplexer that include active logic. A plurality of control signals is received for processing by a corresponding plurality of input selection circuits. A plurality of additional inputs corresponding to the plurality of input selection circuits may also be received. In one embodiment, each input selection circuit is configured to output a corresponding input signal if a corresponding control signal is asserted and a timing signal is made available to the input selection circuit. To avoid unnecessary power consumption associated with the capacitance of various portions of the multiplexer, the timing signal is only asserted to a portion of the multiplexer at any given clock cycle according to the values of the control signals.