Hewlett-Packard Development Company, L.p.

Hewlett-Packard Company

Advanced Micro Devices Corporation

Other

Intel Corporation

Ronny Lee Arnold

Donald Charles Soltis, Jr

Gary J Benjamin

Paul L Rogers

Mahadev S Deshpande

Kevin Safford

Thomas Albert Petersen

Christopher Allan Poirier

Cameron B McNairy

Benjamin James Patella

Alexander Edward Okpisz

Douglas R Williams

Amy May Tuvell

Josef A Dvorak

Donald Charles Soltis Jr

A prefetch device and method are disclosed that determines from which addresses to speculatively fetch data based on information collected regarding previous cache-miss addresses. A historical record showing a propensity to experience cache-misses at a particular address-offset from a prior cache-miss address within a region of memory provides an indication that data needed by future instructions has an increased likelihood to be located at a similar offset from a current cache-miss address. The prefetch device disclosed herein maintains a record of the relationship between a cache-miss address and subsequent cache-miss addresses for the most recent sixty-four unique data manipulation instructions that resulted in a cache-miss. The record includes a weighted confidence value indicative of how many cache-misses previously occurred at each of a selection of offsets from a particular cache-miss address.

Region prefetcher and methods thereof

A first address is received and is used to determine a first address range. The first address range includes a second address range and a third address range. If the first address is in the second address range, a fourth address range is determined. The fourth address range is different from the first address range. Information is retrieved from a memory in response to determining that a second address is in the first address range or the fourth address range. If the first address is in the third address range, a fifth address range is determined. The fifth address range is different from the first address range. Other information is retrieved from the memory in response to determining the second address is in the first address range or the fifth address range.

Hybrid region CAM for region prefetcher and methods thereof

Generally, the present invention provides a system and method for processing instructions of a computer program and for indicating instruction attribute and/or status information so that the efficiency of the processing system may be increased. In architecture, the system of the present invention utilizes a pipeline, a scoreboard, and hazard detection circuitry. The pipeline processes and executes instructions of a computer program. Many of the instructions include register identifiers that identify registers where data should be written when the instructions are executed. When the data produced by execution of one of the instructions has yet to be written to the register identified by the one instruction's register identifier and is unavailable for use in executing other instructions of the program, the one instruction's register identifier is transmitted to the scoreboard. The scoreboard includes a plurality of multi-bit registers, and a first bit in one of the multi-bit registers is changed based on the received register identifier. The scoreboard also may receive data associated with the one instruction and may change a second bit in the one register based on the received data. Therefore, each register in the scoreboard indicates whether a pending write to a particular register exists and indicates information associated with the instruction causing the pending write.

System and method for utilizing a scoreboard to indicate information pertaining to pending register writes

A processing system provides predicate data that indicates whether instructions processed by a processor pipeline should be executed by the pipeline. In architecture, the system of the present invention utilizes a register, a pipeline, and predicate circuitry. The pipeline includes a first stage and a second stage for processing instructions of a computer program. The predicate circuitry is configured to read a first predicate value from the register and to receive a second predicate value. The predicate circuitry may transmit the first predicate value read from the register to the first stage and then select between the first predicate value and the second predicate value. The predicate value selected by the predicate circuitry is transmitted to the second stage.

System and method for providing predicate data to multiple pipeline stages

Generally, the present invention provides a processing system and method for indicating when there is a pending write to a general register of the processing system. The processing system of the present invention utilizes a plurality of general registers, a plurality of connections, a pipeline, a scoreboard, and hazard detection circuitry. The plurality of connections corresponds respectively with the general registers. The scoreboard maintains a plurality of bits such that each bit indicates whether there is a pending write to a corresponding general register. The scoreboard transmits to the hazard detection circuitry one of the bits that is indicative of whether a pending write to the one general register exists based on a value of the one bit and based on which of the connections is used to transmit the one bit. The hazard detection circuitry then detects whether a data hazard exists based on the one bit.

Processing system and method for efficiently enabling detection of data hazards for long latency instructions

Parity and mask bit(s) are stored in a random access memory (RAM) that is coupled to a CAM. This CAM may be part of a TLB. The parity and mask bits(s) are stored in conjunction with the CAM entry write. Upon a CAM query match, the reference parity bit(s) and mask bit(s) stored at the address output by the CAM are output from the RAM. These reference parity bit(s) are compared to parity bit(s) generated from a query data value that is masked by the retrieved mask bit(s). In the absence of a CAM or RAM bit error, the reference parity bit(s) from the RAM and the parity bit(s) generated from the masked query data will match. If a CAM or RAM bit error occurred, these two sets of parity bit(s) will not match and thus an error will be detected. This error may be used as an indication that a false CAM match has occurred.

Detection of bit errors in maskable content addressable memories

A method of rapidly detecting data errors includes generating a reference syndrome from data bits in a data word, the reference syndrome having at least a first portion and a second portion, generating a partial error code from the data bits and the first portion of the reference syndrome, and initiating an error recovery procedure if the partial error code contains any true values.

Early error detection using ECC

The present invention generally relates to a processing system and method for coalescing instruction data to efficiently detect data hazards between instructions of a computer program. In architecture, the system of the present invention utilizes a plurality of pipelines, coalescing circuitry, and hazard detection circuitry. The plurality of pipelines is configured to process instructions of a computer program, and the coalescing circuitry is configured to receive, from the pipelines, a plurality of register identifiers identifying a plurality of registers. The coalescing circuitry is configured to coalesce said register identifiers thereby generating a coalesced register identifier identifying each of said plurality of registers. The hazard detection circuitry is configured to receive the coalesced register identifier and to perform a comparison of the coalesced register identifier with other information received from the pipelines. The hazard detection circuitry is further configured to detect a data hazard based on the comparison.

System and method for coalescing data utilized to detect data hazards

Utilizing a scoreboard with multi-bit registers to indicate a progression status of an instruction that retrieves data

A superscalar processing system that detects data hazards within instruction groups utilizes a memory, a plurality of pipelines, an instruction dispersal unit (IDU), and a control mechanism. The memory includes a plurality of entries that respectively correspond with a plurality of registers. The IDU receives an instruction group that includes a plurality of instructions and transmits the instructions of the instruction group to the plurality of pipelines. The control mechanism analyzes one of the instructions and identifies an entry in the memory that corresponds with a register associated with the one instruction. The control mechanism then analyzes the entry and transmits a warning signal in response to a determination that the entry indicates that another instruction within the instruction group is associated with the register.

System and method for detecting data hazards within an instruction group of a compiled computer program

A superscalar processing system that detects data hazards within instruction groups transmitted to the processing system utilizes a content-addressable memory, a plurality of pipelines, an instruction dispersal unit (IDU), and a control mechanism. The IDU receives an instruction group that includes a plurality of instructions and transmits the instructions of the instruction group to the plurality of pipelines. The control mechanism stores register identifiers of the instructions in the content-addressable memory and determines whether a register identifier of one of the instructions is stored in the content-addressable memory. When the register identifier of the one instruction is stored in the content-addressable memory, the control mechanism transmits a warning signal indicating that one of the instruction groups contained a data hazard.

System and method for detecting an erroneous data hazard between instructions of an instruction group and resulting from a compiler grouping error

Processing system and method utilizing a scoreboard to detect data hazards between instructions of computer programs

System and method for providing predicate data

A processing system receives instructions from a computer program. Each instruction is included within an issue group such that each issue group only includes instructions that may be simultaneously processed. The issue groups are then sequentially transmitted to a plurality of pipelines that simultaneously processes and executes the instructions within the issue groups in program order. During execution, the instructions within an issue group are analyzed to determine whether any of the instructions in the issue group is dependent on unavailable data. Any of the instructions in the issue group determined to be dependent on unavailable data are independently stalled, while execution of other instructions in the issue group is allowed to continue.

Superscalar processing system and method for selectively stalling instructions within an issue group

A computer system utilizing a processing system capable of efficiently comparing register identifiers and instruction attribute data to detect data hazards between instructions of a computer program is used to execute the computer program. The processing system utilizes at least one pipeline, a first decoder, a second decoder, and comparison logic. The pipeline receives and simultaneously processes instructions of a computer program. The first and second decoders are coupled to the pipeline and decode register identifiers associated with instructions being processed by the pipeline. The comparison logic is interfaced with the first and second decoders and receives the decoded register identifiers along with attribute data indicating the status and/or type of instructions being processed by the pipeline. The comparison logic compares the decoded register identifiers and the attribute data to other decoded register identifiers and attribute data to detect data hazards.

Growing community of inventors

Ft. Collins, CO, United States of America

Ronny Lee Arnold