Abstract: A method and apparatus to accelerate variable length decode is disclosed. The system includes a logic device to receive a bit stream of variable length encoded information. The logic device outputs a fixed length value corresponding to a variable length code received as part of the bit stream of the variable length encoded information. The system also includes a processor to receive the fixed length value. The processor to performs a write of a coefficient to a system memory device, the coefficient corresponding to the fixed length value received from the logic device.

Abstract: A method and system for enhancing graphics processing through selected placement of at least one graphics operand in main memory. The system includes a graphics controller in communication with system memory through a dedicated graphics bus such as an Accelerated Graphics Port (AGP) bus. This allows texture maps, alpha blending data and other graphics information to be contained in system memory without degradation of system performance.

Abstract: A method and apparatus for arbitrating between command streams. The method unblocks high priority commands which are blocked and then selects any remaining high priority commands. Normal priority commands are selected after the high priority commands. A memory controller described includes a command queue block having a plurality of command queues, each being coupled to receive a different type of command. The memory controller also includes arbitration logic which, among other things, selects high priority read commands before high priority write commands. Memory interface logic generates memory accesses performing commands selected by the arbitration logic.

Abstract: A method and apparatus for ordering memory access commands. A command ordering circuit which is described includes a plurality of command slots which receive memory access commands. A page register stores a value indicating a last page accessed by a prior memory access command. Comparators compare the value in the page register to values stored in the command slots, and an arbiter receives outputs from the comparators and selects a command from one of the slots. According to the method described, memory accesses are reordered depending on the portion of memory accessed. A first memory access command requesting access to a first portion of memory is issued. Additional memory access commands also referencing the first portion of memory are issued until a count is reached. After the count is reached, a second memory access command which references a second portion of memory is issued.

Abstract: A method and apparatus for fencing the execution of commands. A fence command and an executable command are received in succession. The executable command is enqueued in a first queue together with an indication that the executable command succeeded the fence command. A synchronization value is enqueued in a second queue. The executable command is then delayed from being dequeued from the first queue until the synchronization value is advanced to the head of the second queue.

Abstract: A system and method for ensuring the execution of commands is visible throughout that system. A flush command is received and enqueued in a first queue together with a synchronization value enqueued in a second queue. The flush command is delayed from being dequeued from the first queue until both the flush command and the synchronization value are advanced to the head of their respective queues. Thereafter, the flush command is dequeued subsequently causing the return of a quad-word of random data as an acknowledge signal.

Abstract: A method of issuing a data retrieval command to a re-order unit in a bus bridge is described. The method requires maintaining an indication of the available (or unreserved) capacity in a data buffer into which a data package is received from a memory resource in response to a data retrieval command. Data packages are furthermore dispatched, relative to other data packages, in the order in which a corresponding data retrieval command is issued from a requesting device. The size of a data package requested by each data retrieval command is determined prior to issuance thereof to the re-order unit. The size of each data package is then compared to the then available capacity in the data buffer, and the relevant data retrieval command is only issued to the re-order unit if the available capacity in the data buffer is sufficient to accommodate the data package requested by the relevant data retrieval command.

Abstract: A high throughput memory access interface is provided. The interface includes features which provide higher data transfer rates between system memory and video/graphics or audio adapters than is possible using standard local bus architectures, such as PCI or ISA. The interface allows memory access requests to be performed in such a manner that only portions of an access request are required to be transmitted to the target device for certain bus transactions. Each access request includes command bits, address bits, and length bits. In the initiating device, each access request is separated into three segments, which are stored in separate registers in both the initiating device and the target device. Only the segment which contains the lowest order address bits and the length bits is required by the target device to initiate the bus transaction. Thus, if either of the other two segments has not changed since the previous access request, then such segment or segments are not transmitted to the target.

Abstract: A mechanism for automatically detecting whether a selected type of cache memory is implemented within a cache memory element. The mechanism features a dedicated control line coupled between the cache memory element and a system controller. Logic circuitry is coupled to the control line to force the line to a first logic level in the event that the cache memory element has no connection to support the control line. However, if the cache memory element contains the selected type of cache memory, the logic circuitry is unable to for force the control line to go from a second logic level to the first logic level. After system reset, the system controller samples the voltage on the control line to determine whether the cache memory element is implemented with the selected type of cache memory.

Abstract: Apparatus and a method for utilizing a memory bus write buffer to blend up-to-date data stored in a processor cache and being written back to memory with data in the write buffer being written to the same memory address by a bus master in order to maintain data coherency. The circuitry also utilizes the memory bus write buffer to write valid data furnished in a bus master write over up-to-date data in the write buffer being written to the same memory address from a processor cache in order to maintain data coherency. Combining the data from the two sources prior to writing it to memory eliminates at least one write operation by the write controller along with any associated ECC value generation, may eliminate a number of read/modify/write back operations with any associated ECC value generations, and can double the effective depth of the buffer.

Abstract: Apparatus and a method for testing each write to memory to determine whether it is addressed to an address identical to that of another write to memory waiting to be processed and merging the valid data in any subsequent writes to the same address until a memory write occurs.

Abstract: A bridge for coupling a host bus to a peripheral component interconnect (PCI) bus. A controller is used to transfer an address from the host bus while a datapath is used to transfer data from the host bus. The address and data is then transferred to the PCI bus over a set of signal lines coupled to the PCI bus such that each signal line transfers at least a portion of the address as well as at least a portion of data.

Abstract: A computer system includes an apparatus for conserving power in a tag static random access memory (SRAM). The computer system includes circuitry for placing the tags of the tag SRAM in a reduced power consumption state. The computer system also includes circuitry to power up the tag SRAM out of the reduced power consumption state while maintaining the integrity of the data stored in the tags. The computer system includes a bus, a processor, a cache memory and a memory controller. The memory controller is comprised of a tag static random access memory (SRAM) which includes sense amplifier circuitry and control logic for activating the tag SRAM in response to an address strobe signal (ADS#) from the processor initiating access to the tag SRAM.

Abstract: An apparatus for synchronously detecting phase relationships between a high-frequency clock and a low-frequency clock that are substantially synchronous. The low-frequency clock is frequency divided-by-two to generate a LFdiv2 signal. The LFdiv2 signal is synchronously delayed by one phase of the high-frequency clock to generate a dLFdiv2 signal. The LFdiv2 and dLFdiv2 signals are compared using an XOR gate to generate a PH1 signal. A rising-edge of the PH1 signal indicates that a rising-edge of the high-frequency clock corresponds to a rising-edge of the low-frequency clock. This phase information allows enhanced communication between state machines or buses that are operating at different frequencies.

Abstract: An arbitration circuit which controls arbitration for a resource by a first plurality of agents including a latency sensitive agent. The arbitration circuit comprises a mapping circuit and an arbiter. The mapping circuit is coupled to the first plurality of agents in order to receive a resource request signal from the latency sensitive agent and thereafter produce a plurality of request signals identical to the resource request signal. These request signals are input into at least a first and second I/O ports of the arbiter. The arbiter, which is coupled to the mapping circuit, including a second plurality of I/O ports and a second plurality of control ports each corresponding to one of the I/O ports. The arbiter is configured to arbitrate request signals input into the second plurality of I/O ports including the plurality of request signals, to monitor which I/O port was last activated, and to deactivate a control port associated with the I/O port thereby producing a control signal.

Abstract: A method of improving computer system performance during memory reads. Prior art computer systems experience a considerable time penalty during microprocessor reads from system memory. This time penalty is mitigated by the method of the present invention, wherein data is speculatively retrieved from system memory upon receipt of a microprocessor read request. A microprocessor initiates a read request which is decoded by a memory controller. Before the decoding has completed, the memory controller speculatively begins to retrieve data from the system memory device. Thus if the decode step determines that the requested data is in system memory, the time required to retrieve the data is decreased.

Abstract: An apparatus for synchronously detecting phase relationships between a high-frequency clock and a low-frequency clock that are substantially synchronous. The low-frequency clock is frequency divided-by-two generate a Lfdiv2 signal. The Lfdiv2 signal is synchronously delayed by one phase of the high-frequency clock to generate a dLFdiv2 signal. The LFdiv2 and dLFdiv2 signals are compared an XOR gate to generate a PH1 signal. A rising-edge of the PH1 signal indicates that a rising-edge of the high-frequency clock corresponds to a rising-edge of the low-frequency clock. This phase information allows enhanced communication between state machines or buses that are operating at different frequencies.

Abstract: A method and apparatus for streamlined handshaking between state machines. Under the handshake protocol of the present invention, an initiator state machine sends a request for execution of a task to a target state machine. The target state machine sends an ARM signal to the initiator state machine. When high, the ARM signal indicates to the initiator that the target state machine will begin executing a requested task at the beginning of the next clock cycle. When low, the ARM signal indicates to the initiator that the target state machine will not execute a requested task.

Abstract: A method and apparatus for conserving power in a tag SRAM. The present invention includes circuitry for placing the tags of the tag SRAM in a reduced power consumption state. The present invention also includes circuitry to power up the tag SRAM out of the reduced power consumption state while maintaining the integrity of the data stored in the tags.

Abstract: A secondary clock generator circuit is described in which a ratio is programmed into circuitry which will be embedded on all components. The programmable circuitry takes the system's master clock signal multiplies it by the programmed ratio and yields a secondary clock signal equal to the ratio times the master clock signal for driving components and interfaces. The invention is particularly useful in computer systems operating at a first clock frequency which have replaceable components that run at different clock frequencies and where it is desired to provide a synchronization between the system clock and the various component clocks.