摘要:
A processor including a first execution core section clocked to perform execution operations at a first clock frequency, and a second execution core section clocked to perform execution operations at a second clock frequency which is different than the first clock frequency. The second execution core section runs faster and includes a data cache and critical ALU functions, while the first execution core section includes latency-tolerant functions such as instruction fetch and decode units and non-critical ALU functions. The processor may further include an I/O ring which may be still slower than the first execution core section. Optionally, the first execution core section may include a third execution core section whose clock rate is between that of the first and second execution core sections. Clock multipliers/dividers may be used between the various sections to derive their clocks from a single source, such as the I/O clock.
摘要:
A voltage translator containing a bipolar transistor (Q1), a rectifier (10), a resistor (R1), and a first clamp (12) converts an input voltage (V.sub.I) into one or more output voltages of restricted voltage swing. The first clamp clamps the emitter voltage of the transistor when it is turned on. In one version, the translator includes a second clamp (14) that clamps the collector voltage of the translator when it is turned off. The translator then provides an output voltage (V.sub.O) inverse to the input voltage. In another version, the first clamp is connected between a voltage supply (V.sub.EE) and the emitter of the transistor. Its collector is connected directly to another voltage supply (V.sub.CC) so that the translator only makes non-inverting translations.
摘要:
A switching device (22) responsive to an input voltage V.sub.A is powered by low and high internal supply voltages V.sub.L and V.sub.H. The device changes state as V.sub.A -V.sub.L passes a threshold voltage V.sub.T. After the device makes a desired change of state in response to rising V.sub.A, a hysteresis circuit (24) temporarily decreases V.sub.T below that which would otherwise be present. Likewise, after the device makes a desired change of state in the opposite direction when V.sub.A is falling, the hysteresis circuit temporarily decreases V.sub.T. In both cases, V.sub.T later automatically returns to its original value. This dynamic hysteresis prevents spikes in V.sub.L and V.sub.H from causing undesired changes in state.
摘要:
A logic circuit in which (1) a first bipolar transistor has a base, an emitter, and a collector coupled to a voltage/current source, and (2) a second bipolar transistor has a base coupled to the emitter of the first transistor, an emitter coupled to a constant voltage source, and a collector coupled to the voltage/current source contains operational control circuitry for preventing the second transistor from either turning off or normally going into deep saturation. Each transistor is typically an NPN device. The operational control circuitry may then comprise (1) first circuitry for providing current from the voltage/current source in a single current-flow direction to the collector of the second transistor and (2) second circuitry for providing current from the first circuitry in a single current-flow direction to the base of the second transistor. Optimally, the first circuitry prevents the second transistor from ever going into deep saturation.
摘要:
A processing core is described having execution unit logic circuitry having a first register to store a first vector input operand, a second register to a store a second vector input operand and a third register to store a packed data structure containing scalar input operands a, b, c. The execution unit logic circuitry further include a multiplier to perform the operation (a*(first vector input operand))+(b*(second vector operand))+c.
摘要:
A look-ahead carry adder circuit has multiple stages that are grouped into a carry generation blocks. The size of one of the carry generation blocks is three stages. There may be other carry generation blocks that are of a size that is a whole number multiple of three stages. In an embodiment, the look-ahead carry adder has only one critical path. In a further embodiment, the load on the critical path is minimized by using buffers.
摘要:
A clock duty cycle correction circuit. The duty cycle correction circuit is located at a receiver in a clock distribution network to correct a duty cycle of a distributed clock signal.
摘要:
A signal transfer clocking circuit is disclosed which features a first stage including a first latch and a first, non-clocking circuit in series therewith and a second stage including a second, dynamic latch and at least a second circuit in series therewith. The first latch has a data input side and is opened in response to a first level of a pulse clock signal applied thereto to effect transfer of incoming data through the first stage in a first phase of operation of the signal transfer clocking circuit. The dynamic latch has a data input coupled to an output of the first circuit and is opened in response to a second level of the same or different pulse clock signal applied thereto as that applied to the first latch to effect transfer of data generated by the first stage, through the second stage, to at least one succeeding circuit, in a second, successive phase of operation of the signal transfer clocking circuit, the succeeding circuit being opened and closed by a clock signal having phase and frequency characteristics linked to that applied to the dynamic latch. The dynamic latch clocking methodology used in the signal transfer clocking circuit is particularly applicable to any traditional static latch circuits, such as CMOS static latch circuitry, although not limited thereto, to minimize min-delay race violations and to effect a skew-tolerant operation.
摘要:
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.
摘要:
A processor including a first execution core section clocked to perform execution operations at a first clock frequency, and a second execution core section clocked to perform execution operations at a second clock frequency which is different than the first clock frequency. The second execution core section runs faster and includes a data cache and critical ALU functions, while the first execution core section includes latency-tolerant functions such as instruction fetch and decode units and non-critical ALU functions. The processor may further include an I/O ring which may be still slower than the first execution core section. Optionally, the first execution core section may include a third execution core section whose clock rate is between that of the first and second execution core sections. Clock multipliers/dividers may be used between the various sections to derive their clocks from a single source, such as the I/O clock.