Sun Microsystems, Inc.

Xerox Corporation

3par Inc.

Intel Corporation

Michel P Cekleov

Russell N Mirov

Mark S Young

William M Baldwin

Pradeep S Sindhu

Jean-Marc Frailong

Akhilesh Kumar

Lily Pao Looi

Robert T George

Edmund Joseph Kelly

Christopher T Cheng

Manoj Khare

Eric H Jensen

Jeffrey H Hoel

Michael L Powell

David Chu

Ken Creta

Greg L Dykema

Michel Dubois

David L Griffith

A method is provided to align clock and data signals over a source-synchronous link. The method includes sending header data and a default clock signal over the link. The header indicates a start of a training packet and the default clock signal ensures that the header is received without error. The method further includes providing a long clock pulse, phase shifting the clock signal during the long clock pulse, and thereafter sending training data and the clock signal over the link. The above steps are repeated until the training data are received with error. At that point, the phase shift of the clock signal is saved as a boundary of an optimal alignment. The above steps are then repeated with the clock signal shifted in a different direction. Once another boundary is located, the boundary midpoint is saved as the phase shift that provides the optimal alignment.

Trainable link

A memory controller includes a gate circuit gating a data strobe signal from a memory device, a delay circuit delaying the data strobe signal from the gate circuit, a read buffer capturing values of a data signal from the memory device in response to the data strobe signal, a cumulative strobe counter incrementing a detected strobe count by the number of edges detected in the data strobe signal, and a control logic controlling the gate circuit and receiving the detected strobe count from the strobe counter. The control logic enables and disables the gate circuit after the start of a preamble and before the end of a postamble in the data strobe signal, respectively. When the memory controller is not expecting the data strobe signal from the memory device, the control logic compares the detected and the expected strobe counts and reports a strobe error when they do not match.

Dual data rate memory strobe checker

A method is provided for controlling transitions between a first and second clock frequency signal in first and second components electrically coupled together and in communication with one another. The method comprises asserting a freeze signal to cause communications between the first and second components to cease. A freeze acknowledge signal is then received from the first and second components, indicating that communications therebetween have ceased. A change signal is delivered to the first and second components to cause the components to switch between the first and second clock frequency signals.

Method and apparatus for controlling transitions between a first and a second clock frequency

A method for operating a cache having a sleep mode is provided. The cache is located within a memory hierarchy of a computer system, and the method is comprised of receiving a first cache request, and servicing the first cache request. A sleep mode signal is asserted in response to completion of the servicing of the first cache request. Thereafter, a second cache request is received, and the sleep mode signal is deasserted in response to receiving the second cache request. Thereafter, the second cache request is serviced.

Method and apparatus for reducing power consumption in a cache memory system

A method of invalidating a cache line in a system having a plurality of nodes that include a processor and a cache memory. A request to invalidate a cache line that is caching a particular memory block is sent from a first node. The request is a request to invalidate a cache line in another node without returning to the first node the data stored in a cache line to be invalidated. In an embodiment, the data in the cache line to be invalidated is not returned to the first node even if the cache line is in the modified state. In a further embodiment, new data is written to a cache line in the first node that is caching the particular memory block without writing old data that was stored in that cache line back to a memory.

Method and apparatus for invalidating a cache line without data return in a multi-node architecture

A method for controlling thermal cycles in a computer system is provided. The method is comprised of receiving a request to transition the computer system from a first operating mode to a second operating mode, where less power is consumed in the second operating mode. A historical rate at which the computer system has transitioned between the first and second operating modes is determined, and the requested transition is permitted in response to the historical rate being less than a first preselected setpoint, but denied in response to the historical rate being greater than the first preselected setpoint.

Method and apparatus for reducing power consumption

A method for controlling operation of a bus and components coupled thereto is provided. The method is comprised of receiving a request for a bus transaction from one of the components coupled to the bus. Thereafter, the frequency of a clock signal supplied to at least the requesting component is increased, and the requested bus transaction is serviced. The frequency of the clock signal supplied to at least the requesting component is decreased upon completion of the requested bus transaction.

Method and apparatus for controlling a bus clock frequency in response to a signal from a requesting component

In one aspect of the present invention, a method for controlling the operation of a phase locked loop circuit is provided. The method is comprised of monitoring a frequency of a system clock, and a first signal is delivered in response to the detected frequency of the system clock being greater than a preselected setpoint. A second signal is delivered in response to the detected frequency of the system clock being less than a preselected setpoint. A first operating mode of the phase locked loop circuit is selected in response to receiving the first signal. The first mode of operation allows the phase locked loop circuit to synchronize with a clock signal in a first preselected range of frequencies. A second operating mode of the phase locked loop circuit is selected in response to receiving the second signal. The second mode of operation allows the phase locked loop circuit to synchronize with a clock signal in a second preselected range of frequencies.

A phase locked loop circuit is provided. The phase locked loop circuit is comprised of a first and second divide-by-N counter, a phase comparator, a voltage controlled oscillator, a clock tree, and a feedback path. The first divide-by-N counter is adapted to receive a first clock signal and provide a second clock signal. The phase comparator has a first and second input terminal and an output terminal. The phase comparator is adapted to compare the phase of signals applied to the first and second input terminals and deliver a signal at the output terminal having a magnitude indicative of a difference in the phases of the signals. The first input terminal is coupled to receive the second clock signal. The voltage controlled oscillator is coupled to receive the phase difference signal and deliver a third clock signal having a frequency responsive thereto. The second divide-by-N counter is coupled to receive the third clock signal and deliver a fourth clock signal. The clock tree is coupled to receive the third clock signal and deliver at least one fourth clock signal. The feedback path is coupled to deliver the fourth clock signal to the second input terminal of the phase comparator.

A memory system is provided. The memory system is comprised of a memory, a clock signal generator, a phase locked loop circuit, and a bypass circuit. The clock signal generator produces a first clock signal. The clock signal generator has a first mode of operation in which the first clock signal has a first frequency and a second mode of operation in which the first clock signal has a second frequency. The phase locked loop circuit is associated with the memory and adapted for receiving the first clock signal and providing a synchronized second clock signal to the memory. The bypass circuit is adapted to deliver the first clock signal to the memory in the second mode of operation.

A method is provided for operating an electronic device by monitoring operating characteristics of the electronic device, and determining from the monitored operating characteristics to operate at least a portion of components within the electronic device in a first, second, or third mode of operation. The first, second, and third modes of operation consume power at first, second, and third different rates. At least a portion of the components are instructed to switch between the first, second, and third modes.

A power supply is provided. The power supply is comprised of a first power module, a second power module, and a controller. The first power module is capable of delivering a first preselected amount of power. The second power module is capable of delivering a second preselected amount of power. The first and second power modules are coupled together to be capable of delivering an aggregate amount of power related to the first and second preselected amounts of power. The controller is adapted to selectively enable the first and second power modules to make power available in one of the first preselected amount, the second preselected amount, and the aggregate amount.

The bus of the present invention advantageously utilizes high-speed, source synchronized data transfers and lower-speed, globally synchronized transfers of arbitration and consistency information. In a first embodiment, a high speed clock signal and slower speed clock enable signal are globally distributed from a central arbiter to agents coupled to the bus. A sending agent utilizes the high speed clock signal for source synchronized data transfers by forwarding the high speed clock signal, along with the data, to one or more receiving agents. Thus, the globally distributed clock signal is used to accomplish source synchronized data transfers. Arbitration requests, by contrast, are processed at the slower clock enable signal rate in a globally synchronous fashion. In addition, by communicating data cycles information from the central arbiter to the receiving agent at the slower clock enable signal rate, the present invention avoids resynchronization and the possibility of metastability. Dead time between packets of data is minimized in the present invention by placing the central arbiter in the center of the bus. An alternative embodiment is disclosed wherein a slower speed clock signal is globally distributed to a plurality of agents and a central arbiter. Each agent then generates a high speed clock signal that is divided down and phase locked to the slower speed clock signal, and this high speed clock signal is then utilized for high-speed, source synchronized data transfers. Arbitration and consistency information are handled at the slower speed clock signal rate. In the alternative embodiment, resynchronization is avoided through the use of a header signal.

Source synchronized metastable free bus

A computer system includes first and second processors each having a virtual cache memory, a main memory, a bus coupled to the main memory and the processors, and apparatus for addressing the cache associated with each processor for providing that the data in each virtual cache stores data from the same physical location in main memory at a same index position in each virtual cache, a memory management unit (MMU) coupled to each processor such that addressing information is transferred to each memory management unit to indicate the virtual address of data to be written to the virtual cache, the memory management unit generating from the virtual address a physical address, and determining whether any other virtual cache includes data from the same physical memory positions.

Apparatus for maintaining consistency in a multiprocessor computer

An improved memory model and implementation is disclosed. The memory model includes a Total Store Ordering (TSO) and Partial Store Ordering (PSO) memory model to provide a partial order for the memory operations which are issued by multiple processors. The TSO memory model includes a FIFO Store Buffer for Store, and Atomic Load-Store operations. The Load operations are not placed in the FIFO Store Buffer. The Load operation checks for a value stored in the same location in the FIFO Store Buffer; if no such value is found, then requested value is returned from memory. The PSO model also includes a Store Buffer for Store, and Atomic Load-Store operations. However, unlike the TSO model, the Store Buffer in the PSO model is not FIFO. The processors in the PSO model may issue the Store and Atomic Load-Store in a certain order; however, such operations may be executed by memory out of the order issued by the processors. The execution order is assured only by address matching and the STBAR operation. Two Store operations separated by a STBAR operations guarantees memory will execute the operations in an order issued by the processors. Load operations in the PSO model are not placed in the Store Buffer. The Load operation first checks for a value stored in the same location in the Store Buffer; if no such value is found, then the requested value is returned from memory.

Growing community of inventors

Mountain View, CA, United States of America

Michel P Cekleov