Advanced Micro Devices Corporation

Scott E Swanstrom

Larry David Hewitt

Douglas D Gephardt

Frank P Helms

Steven L Belt

Dale E Gulick

William Alexander Hughes

Derrick R Meyer

Drew Jon Dutton

J Andrew Lambrecht

Scott A White

David S Christie

Michael T Wisor

Amitabh Mehra

Anil Harwani

Paul C Miranda

William R Alverson

Grant E Ley

Jerry A Ahrens

Mustansir M Pratapgarhwala

Andy Steinbach

Andy Lambrecht

Ross Voigt LaFetra

Automatic part testing includes: booting a part under testing into a first operating environment; executing, via the first operating environment, one or more test patterns on the part; performing a comparison between one or more observed characteristics associated with the one or more test patterns and one or more expected characteristics; and modifying one or more operational parameters of a central processing unit of the part based on the comparison.

Automatic part testing

A serial communication system includes an integrated circuit having a master serial interface; and a processor having a slave serial interface coupled to the master serial interface through a clock signal line and a data signal line. The slave serial interface is responsive to a read temperature command issued by the master serial interface to return a first temperature value associated with the processor.

Serial interface having a read temperature command

A message based power management approach is utilized to provide power management for a multi-processor system. A power management message is received at one processor of the multi-processor system over an input/output communication link that provides input/output access for the processors of the multi-processor system. The power management message includes a power management field encoding a power management state for processors of the multi-processor system. The processor that received the power management message over the input/output communication link sends a power management message to other processors in the multi-processor system over one or more inter-processor communication links encoding the power management state.

Message based power management in a multi-processor system

An integrated circuit is coupled to a communication link and to a separate signal line and includes programmable registers specifying communication link width and frequency. The integrated circuit responds to a change in the value of the signal line by changing the width and/or frequency of at least a portion of the communication link to the programmed value in response to a change in a logical value of the signal line, without the integrated circuit entering a reset state. The width and/or frequency may be changed during a POST routine or during system operation as part of a power management or other system function while maintaining its operational state.

Use of a signal line to adjust width and/or frequency of a communication link during system operation

A message based power management system converts legacy signals used in power management, and other signals used to differentiate between power states, to messages sent over a communication link. A system message sent on a communication link includes a field encoding the level of power management for the device receiving the system message. Further, one or more additional signals, separate from the communication link, may be used to indicate when to take action after the power management message has been received.

Message based power management

A computer system optimized for real-time applications which provides increased performance over current computer architectures. The system includes a standard local system bus or expansion bus, such as the PCI bus, and may also include a dedicated real-time bus or multimedia bus. Various multimedia devices are coupled to one or more of the expansion bus and/or the multimedia bus. The computer system includes byte slicing and/or data packing logic coupled to one or more of the expansion bus and/or the multimedia bus which operates to allow different data streams to use different byte channels simultaneously. Thus the byte sliced bus allows different peripherals to share the bus simultaneously. The byte slicing logic thus may assign one data stream to a subset of the total byte lanes on the multimedia bus, and fill the unused byte lanes with another data stream. The data packing logic may optimally fill the bus with data having more or fewer bits than the bus. The computer system of the present invention thus provides much greater performance for real-time applications than prior systems.

Computer system which performs intelligent byte slicing/data packing on

A computer system optimized for real-time applications which provides increased performance over current computer architectures. The system includes a standard local system bus or expansion bus, such as the PCI bus, and may also include a dedicated real-time bus or multimedia bus. Various multimedia devices are coupled to one or more of the expansion bus and/or the multimedia bus. The computer system includes byte slicing logic coupled to one or more of the expansion bus and/or the multimedia bus which operates to allow different data streams to use different byte channels simultaneously. Thus the byte sliced multimedia bus allows different peripherals to share the bus simultaneously. The byte slicing logic thus may assign one data stream to a subset of the total byte lanes on the multimedia bus, and fill the unused byte lanes with another data stream. The computer system of the present invention thus provides much greater performance for real-time applications than prior systems.

Computer system which performs intelligent byte slicing on a multi-byte

A computer chip including multiple on-chip modules connected by an on-chip bus which provides increased performance over current computer chip architectures. The on-chip system bus is a bit sliced bus. Various transmitters/and or receivers are coupled the bit sliced bus. The transmitters and/or receivers include bus interface logic and/or bit transfer logic and/or bit receive logic operatively coupled to the on-chip bit sliced bus which operates to allow different data streams to use different bit lines substantially simultaneously. Thus the bit sliced bus allows different devices to share the bus simultaneously. The bus interface logic and/or the bit transfer logic thus may assign one data stream to a subset of the total bit lines on the bit sliced bus, and fill the unused bit lines with another data stream.

Independent use of bits on an on-chip bus

A computing system and bus bridge in which the bus bridge includes a buffer pool wherein the storage buffers contained in the buffer pool may be allocated as post buffers or fetch buffers in response to appropriate requests from the bus bridge. In the preferred embodiment, the bus bridge includes a buffer pool control unit adapted to temporarily allocate any of the plurality of storage buffers as either a post buffer or a fetch buffer depending upon the system requirements. Broadly speaking, the present invention contemplates a computing system including a first component connected to a first bus, a second component connected to a second bus, and a bus bridge connected to a first and second busses. The bus bridge includes a buffer pool comprised of a plurality of storage buffers and a buffer pool control unit that is capable of temporarily allocating at least one of the storage buffers as either a post buffer or a fetch buffer in response to system requirement. Preferably, each storage buffer includes corresponding tag information for identifying an origin or destination location within a main memory of the data associated with the storage buffer. In one embodiment, each of the plurality of storage buffers includes corresponding allocation information used by the buffer pool control unit for the temporary allocation of the storage buffers. In a presently preferred embodiment, the allocation information includes an available bit indicative of whether the storage buffers available for allocation and a post/fetch bit indicative of whether an unavailable storage buffer is currently allocated as a post buffer or as a fetch buffer. The allocation information is accessible by the buffer pool control unit and is used by the buffer pool control unit for temporarily allocating the storage buffers.

Allocatable post and prefetch buffers for bus bridges

A computer system is provided for controlling the ownership of a bus to which a variety of both real time and non-real time resources are coupled. The bus arbiter includes a request detection unit for detecting bus request signals of a plurality of bus masters, and a grant generator for generating corresponding grant signals to indicate a grant of ownership of the bus. A set of counters referred to as 'request latency' counters is further provided wherein a separate counter unit corresponds to each bus master. Each counter is configured to generate a signal indicative of a lapse of time since a time when the peripheral requested ownership of the bus. An arbitration control unit is coupled to the request latency counters, the request detection unit and the grant generator for processing incoming bus request signals. The arbitration control unit is configured to dynamically vary the level of arbitration priority given to each peripheral device based upon the latency signal corresponding to the device. Accordingly, as the time from when a peripheral device requests the bus increases, the level of arbitration priority given to that peripheral also increases. A set of programmable registers are provided to allow software programming of the initial count value associated with each request latency counter. The request latency counter for a particular device may further be held or inhibited from counting to provide a constant level of priority for that particular peripheral device.

Bus arbiter including programmable request latency counters for varying

An interrupt polling unit included within a bus interface unit of a microprocessor is provided. The interrupt polling unit causes an interrupt acknowledge bus transaction to occur. If an interrupt controller receiving the interrupt acknowledge bus transaction returns an interrupt vector indicative of an interrupt service routine, then the microprocessor executes the interrupt service routine. The number of interrupt acknowledge bus transactions associated with the interrupt is reduced from two to one. In one embodiment, the interrupt polling unit causes an interrupt acknowledge bus transaction to occur when the microprocessor is performing a task switch. The task switch may be performed by hardware included within the microprocessor or, alternatively, by software executing upon the microprocessor.

Microprocessor including an interrupt polling unit configured to poll

A cache memory system comprising a first bus for connecting to a bus master and a second bus for connecting to a system memory. The system memory comprises a plurality of cacheable memory locations. A bus bridge provides an interface between the first bus and the second bus. A cache memory controller for caching data stored in the cacheable memory locations is connected to the system memory. The cache memory controller includes a snoop control circuit directly coupled to the first bus for snooping bus transactions upon the first bus and further coupled to the second bus for snooping bus transactions on said second bus.

Architecture and method for controlling a cache memory

A computer system optimized for real-time applications which provides increased performance over current computer architectures. The system includes a standard local system bus or expansion bus, such as the PCI bus, and also includes a dedicated real-time bus or multimedia bus. Various multimedia devices are coupled to one or more of the expansion bus and/or the multimedia bus. The computer system includes byte slicing logic coupled to one or more of the expansion bus and/or the multimedia bus which operates to allow different data streams to use different byte channels simultaneously. Thus the byte sliced multimedia bus allows different peripherals to share the bus simultaneously. The byte slicing logic thus may assign one data stream to a subset of the total byte lanes on the multimedia bus, and fill the unused byte lanes with another data stream. The computer system of the present invention thus provides much greater performance for real-time applications than prior systems.

Computer system having a multimedia bus and comprising a centralized I/O

A computer system comprising an improved DMA controller for performing DMA transfers between a peripheral device and system memory and receiving and servicing interrupts generated by the peripheral device. The system comprises one or more buses for transferring data. A CPU, system memory and a plurality of peripheral devices are interconnected by the buses. Each of the peripheral devices comprises one or more peripheral interrupt request outputs. The system further comprises a programmable DMA controller coupled to the bus which receives the peripheral interrupt request outputs. The DMA controller is configured to perform a DMA transfer on the one or more buses between two or more devices, including the system memory and the plurality of peripheral devices. The CPU programs the DMA controller to start the DMA transfer in response to one of the plurality of peripheral devices generating an interrupt request on its interrupt request output or to start the DMA transfer immediately. The system further comprises a peripheral interrupt controller (PIC) coupled to the CPU. The PIC includes a plurality of PIC interrupt request inputs coupled to the DMA controller. The DMA controller is configured to selectively couple the peripheral interrupt request outputs from the peripheral devices to the plurality of PIC interrupt request inputs of the PIC, to selectively decouple the peripheral interrupt request outputs from the peripheral devices from the plurality of PIC interrupt request inputs of the PIC, and to selectively generate a plurality of interrupt requests on the plurality of PIC interrupt request inputs of the PIC when the peripheral interrupt request outputs from the peripheral devices are decoupled from the plurality of PIC interrupt request inputs. Thus, the DMA controller renders the CPU more responsive to real-time events by servicing a portion of the peripheral interrupts.

Growing community of inventors

San Jose, CA, United States of America

Scott E Swanstrom