Advanced Micro Devices Corporation

Stmicroelectronics Gmbh

Hewlett-packard Company

Hewlett-packard Development Company, L.p.

Anthony Jarvis

Mark Owen Homewood

Thomas A Clouqueur

Paolo Faraboschi

Gary Lee Vondran, Jr

Marius Evers

James David Dundas

Geoffrey M Brown

Aparna Thyagarajan

Alexander J Starr

Arunachalam Annamalai

Gerald David Zuraski, Jr

Ravindra Nath Bhargava

David N Suggs

Robert Bradford Cohen

Ashok Tirupathy Venkatachar

Steven R Havlir

Trivikram Krishnamurthy

Thomas Cloqueur

Aparna Mandke

Hybrid parallelized tagged geometric (TAGE) branch prediction, including: selecting, based on a branch instruction, a first plurality of counts from at least one TAGE table; selecting, based on the branch instruction, a second plurality of counts from at least one non-TAGE branch prediction table; generating, based on the first plurality of counts and a second plurality of counts; and wherein selecting the first plurality of counts and selecting the second plurality of counts are performed during a same branch prediction pipeline stage.

Hybrid parallelized tagged geometric (TAGE) branch prediction

Merging branch target buffer entries includes maintaining, in a branch target buffer, an entry corresponding to first branch instruction, where the entry identifies a first branch target address for the first branch instruction and a second branch target address for a second branch instruction; and accessing, based on the first branch instruction, the entry.

Merged branch target buffer entries

A processor includes two or more branch target buffer (BTB) tables for branch prediction, each BTB table storing entries of a different target size or width or storing entries of a different branch type. Each BTB entry includes at least a tag and a target address. For certain branch types that only require a few target address bits, the respective BTB tables are narrower thereby allowing for more BTB entries in the processor separated into respective BTB tables by branch instruction type. An increased number of available BTB entries are stored in a same or a less space in the processor thereby increasing a speed of instruction processing. BTB tables can be defined that do not store any target address and rely on a decode unit to provide it. High value BTB entries have dedicated storage and are therefore less likely to be evicted than low value BTB entries.

Multiple-table branch target buffer

A processor predicts a number of loop iterations associated with a set of loop instructions. In response to the predicted number of loop iterations exceeding a first loop iteration threshold, the set of loop instructions are executed in a loop mode that includes placing at least one component of an instruction pipeline of the processor in a low-power mode or state and executing the set of loop instructions from a loop buffer. In response to the predicted number of loop iterations being less than or equal to a second loop iteration threshold, the set of instructions are executed in a non-loop mode that includes maintaining at least one component of the instruction pipeline in a powered up state and executing the set of loop instructions from an instruction fetch unit of the instruction pipeline.

Using loop exit prediction to accelerate or suppress loop mode of a processor

A processor includes a prediction engine coupled to a training engine. The prediction engine includes a loop exit predictor. The training engine includes a loop exit branch monitor coupled to a loop detector. Based on at least one of a plurality of call return levels, the loop detector of the processor takes a snapshot of a retired predicted block during a first retirement time, compares the snapshot to a subsequent retired predicted block at a second retirement time, and based on the comparison, identifies a loop and loop exit branches within the loop for use by the loop exit branch monitor and the loop exit predictor to determine whether to override a general purpose conditional prediction.

Loop exit predictor

Systems and methods selectively filter, buffer, and process cache coherency probes. A processor includes a probe buffering unit that includes a cache coherency probe buffer. The probe buffering unit receives cache coherency probes and memory access requests for a cache. The probe buffering unit identifies and discards any of the probes that are directed to a memory block that is not cached in the cache, and buffers at least a subset of the remaining probes in the probe buffer. The probe buffering unit submits to the cache, in descending order of priority, one or more of: any buffered probes that are directed to the memory block to which a current memory access request is also directed; any current memory access requests that are directed to a memory block to which there is not a buffered probe also directed; and any buffered probes when there is not a current memory access request.

Systems and methods for selectively filtering, buffering, and processing cache coherency probes

A method and apparatus for branch prediction is disclosed. A pattern history table (PHT) is accessed based on at least one global history value to obtain a prediction value. The prediction value and the at least one global history value used to obtain the prediction value are placed in a queue. If a branch prediction is requested, the queue is accessed to obtain a prediction value. The queue may include any number of entries and the queue maintains the oldest prediction value at the head of the queue. The prediction value at the head of the queue is used when a branch prediction is needed.

Power efficient pattern history table fetch in branch predictor

A branch predictor for use in a processor includes a Level 1 branch predictor, a Level 2 branch predictor, a match determining circuit, and an override determining circuit. The Level 1 branch predictor generates a Level 1 branch prediction. The Level 2 branch predictor generates a Level 2 branch prediction. The match determining circuit determines whether the Level 1 and Level 2 branch predictions match. The override determining circuit determines whether to override the Level 1 branch prediction with the Level 2 branch prediction. The Level 1 branch prediction is used when the Level 1 and Level 2 branch predictions match or when the Level 1 and Level 2 branch predictions do not match and the Level 1 branch prediction is not overridden. The Level 2 branch prediction is used when the Level 1 and Level 2 branch predictions do not match and the Level 1 branch prediction is overridden.

Combined level 1 and level 2 branch predictor

Techniques are disclosed relating to improving the performance of branch prediction in processors. In one embodiment, a processor is disclosed that includes a branch prediction unit configured to predict a sequence of instructions to be issued by the processor for execution. The processor also includes a pattern detection unit configured to detect a pattern in the predicted sequence of instructions, where the pattern includes a plurality of predicted instructions. In response to the pattern detection unit detecting the pattern, the processor is configured to switch from issuing instructions predicted by the branch prediction unit to issuing the plurality of instructions. In some embodiments, the processor includes a replay unit that is configured to replay fetch addresses to an instruction fetch unit to cause the plurality of predicted instructions to be issued.

Detecting branch direction and target address pattern and supplying fetch address by replay unit instead of branch prediction unit

A loop predictor and a method for instruction fetching using a loop predictor. A processor may include a loop predictor in addition to a primary branch predictor. A relatively common scenario in program execution is that a set of branches repeat over and over forming a loop. The loop may be detected based on a repeated pattern of access to a data structure used for branch prediction. Once a loop is detected and it may be determined whether the codes would stay in the loop for at least a duration sufficient to disable the branch prediction. On a determination that the detected loop is locked, a sequence of instruction addresses in one iteration of the detected loop may be captured in a buffer and the branch predictor may be turned off and a sequence of fetch instructions may be played from the buffer.

Loop predictor and method for instruction fetching using a loop predictor

A system and method for branch prediction in a microprocessor. A hybrid device stores branch prediction information in a sparse cache for no more than a common smaller number of branches within each entry of the instruction cache. For the less common case wherein an i-cache line comprises additional branches, the device stores the corresponding branch prediction information in a dense cache. Each entry of the sparse cache stores a bit vector indicating whether or not a corresponding instruction cache line includes additional branch instructions. This indication may also be used to select an entry in the dense cache for storage. A second sparse cache stores entire evicted entries from the first sparse cache.

Hybrid branch prediction device with sparse and dense prediction caches

There is disclosed a data processor that executes variable latency load operations using bypass circuitry that allows load word operations to avoid stalls caused by shifting circuitry. The processor comprises: 1) an instruction execution pipeline comprising N processing stages, each of the N processing stages for performing one of a plurality of execution steps associated with a pending instruction being executed by the instruction execution pipeline; 2) a data cache for storing data values used by the pending instruction; 3) a plurality of registers for receiving the data values from the data cache; 4) a load store unit for transferring a first one of the data values from the data cache to a target one of the plurality of registers during execution of a load operation; 5) a shifter circuit associated with the load store unit for shifting the first data value prior to loading the first data value into the target register; and 6) bypass circuitry associated with the load store unit for transferring the first data value from the data cache directly to the target register without processing the first data value in the shifter circuit.

System and method for executing variable latency load operations in a date processor

A data processor includes execution clusters, an instruction cache, an instruction issue unit, and alignment and dispersal circuitry. Each execution cluster includes an instruction execution pipeline having a number of processing stages, and each execution pipeline is a number of lanes wide. The processing stages execute instruction bundles, where each instruction bundle has one or more syllables. Each lane is capable of receiving one of the syllables of an instruction bundle. The instruction cache includes a number of cache lines. The instruction issue unit receives fetched cache lines and issues complete instruction bundles toward the execution clusters. The alignment and dispersal circuitry receives the complete instruction bundles from the instruction issue unit and routes each received complete instruction bundle to a correct one of the execution clusters. The complete instruction bundles are routed as a function of at least one address bit associated with each complete instruction bundle.

Growing community of inventors

Acton, MA, United States of America

Anthony Jarvis