Arm Limited

Other

Luca Nassi

Cédric Denis Robert Airaud

Albin Pierrick Tonnerre

Rémi Marius Teyssier

Geoffray Lacourba

François Donati

Xiaoyang Shen

Damien Robin Martin

Guillaume Bolbenes

Damian Maiorano

Quentin Éric Nouvel

Houdhaifa Bouzguarrou

Paolo Monti

Yohan Fernand Fargeix

Pasquale Ranone

Christophe Carbonne

Lee Evan Eisen

Abhishek Raja

Miles Robert Dooley

Thibaut Elie Lanois

Natalya Bondarenko

Huzefa Moiz Sanjeliwala

Lucas Garcia

Stefano Ghiggini

Damien Matthieu Valentin Cathrine

Ugo Castorina

Jocelyn François Orion Jaubert

Adrien Pesle

Nicola Piano

Alexander James Pane

Muhammed Enes TopÇu

Ilaria Bosco

An apparatus has cache data storage, and tagged-data prediction circuitry to generate a tagged-data prediction in response to a streaming-write request requesting that write data corresponding to a target address which missed in a previous level of cache is written to the cache data storage for the given level of cache without being allocated into the previous level of cache. The tagged-data prediction is indicative of whether a target cache data entry corresponding to the target address of the streaming-write request is predicted to be a tagged cache data entry that stores cached data associated with the target address and a valid memory safety check tag corresponding to the target address, or an untagged cache data entry that stores the cached data but does not store a valid memory safety check tag. Cache access scheduling circuitry selects, based on the tagged-data prediction generated by the tagged-data prediction circuitry for the streaming-write request, how to schedule access to the cache data storage in response to the streaming-write request.

Tagged-data prediction

Apparatuses, systems, methods, computer-readable media, and computer code are disclosed. An apparatus comprises allocation circuitry configured to allocate store data comprising N portions in a store buffer. The store buffer comprises a data buffer configured to hold data entries, each data entry configured to hold one of the N portions of the store data; and an address buffer configured to hold address entries, wherein a given address entry is indicative of a memory address at which the store data is to be stored and is associated with N indications identifying one or more of the data entries that contain the store data. The allocation circuitry is responsive to a determination that M portions of the N portions of the store data each comprise duplicate data to set M indications of the N indications to identify a given data entry holding the duplicate data.

Allocation of store buffer entries

Mode change detection circuitry detects a mode change when processing circuitry switches between first and second modes of processing in which a first set of architectural registers are designated as having different register lengths. Register mapping circuitry maps architectural registers to corresponding physical registers. For an operation specifying a given architectural register of the first set of architectural registers: in response to a determination that the operation is to be processed in the first mode of processing, the register mapping circuitry maps the given architectural register to a physical register of a first physical register file, and in response to a determination that the operation is to be processed in the second mode of processing, the register mapping circuitry maps the given architectural register to a physical register of a second physical register file separate from the first physical register file and having physical registers of different register length to physical registers of the first physical register file.

Register mapping to map architectural registers to corresponding physical registers based on a mode indicating a register length

There is provided an apparatus, method for data processing. The apparatus comprises post decode cracking circuitry responsive to receipt of decoded instructions from decode circuitry of a processing pipeline, to crack the decoded instructions into micro-operations to be processed by processing circuitry of the processing pipeline. The post decode cracking circuitry is responsive to receipt of a decoded instruction suitable for cracking into a plurality of micro-operations including at least one pair of micro-operations having a producer-consumer data dependency, to generate the plurality of micro-operations including a producer micro-operation and a consumer micro-operation, and to assign a transfer register to transfer data between the producer micro-operation and the consumer micro-operation.

Cracking instructions into a plurality of micro-operations

An apparatus is provided for controlling the operating mode of control circuitry, such that the control circuitry may change between two operating modes. In an allocation mode, data that is loaded in response to an instruction is allocated into storage circuitry from an intermediate buffer, and the data is read from the storage circuitry. In a non-allocation mode, the data is not allocated to the storage circuitry, and is read directly from intermediate buffer. The control of the operating mode may be performed by mode control circuitry, and the mode may be changed in dependence on the type of instruction that calls the data, and whether the data may be used again in the near future, or whether it is expected to be used only once.

Controlling data allocation to storage circuitry

A data processing apparatus is provided. Instruction send circuitry sends an instruction to an external processor to be executed by the external processor. Allocation circuitry allocates a specified one of several registers for a result of the instruction having been executed on the external processor and data receive circuitry receives the result of the instruction having been executed on the external processor and stores the result in the specified one of the several registers. In response to a condition being met: the specified one of the several registers is dereserved prior to the result being received by the data receive circuitry, and the result is discarded by the data receive circuitry when the result is received by the data receive circuitry.

Register freeing latency

A data processing apparatus comprises: execution circuitry to execute instructions in order to perform data processing operations specified by those instructions; a plurality of registers to store data values for access by the execution circuitry when performing the data processing operations, each register having an associated physical register identifier; register rename circuitry to select physical register identifiers to associate with architectural register identifiers specified by the instructions; and rename storage having a plurality of entries, each entry being associated with one of the architectural register identifiers and used by the register rename circuitry to indicate a physical register identifier selected for association with that one of the architectural register identifiers; the register rename circuitry comprising an execute unit, and being responsive to detection of an early execute condition for a given instruction, the early execute condition requiring at least detection that each source value required to execute the given instruction is available to the register rename circuitry without accessing the plurality of registers, to cause the execute unit to perform the data processing operation specified by the given instruction in order to generate a result value, and to cause the generated result value to be stored in an entry of the rename storage associated with a destination architectural register identifier specified by the given instruction.

Technique for improving efficiency of data processing operations in an apparatus that employs register renaming

Processing circuitry performs processing operations in response to micro-operations. Front end circuitry supplies the micro-operations to be processed by the processing circuitry. Prediction circuitry generates a prediction of a number of loop iterations for which one or more micro-operations per loop iteration are to be supplied by the front end circuitry, where an actual number of loop iterations to be processed by the processing circuitry is resolvable by the processing circuitry based on at least one operand corresponding to a first loop iteration to be processed by the processing circuitry. The front end circuitry varies, based on a level of confidence in the prediction of the number of loop iterations, a supply rate with which the one or more micro-operations for at least a subset of the loop iterations are supplied to the processing circuitry.

Micro-operation supply rate variation

An apparatus comprises prefetch circuitry, and a cache having a plurality of entries to store data for access by processing circuitry and blocks of metadata for reference by the prefetch circuitry. The prefetch circuitry can detect one or more access sequences in dependence on training inputs derived from demand accesses processed by the cache in response to memory access operations performed by the processing circuitry. On detecting a given access sequence, this causes an associated given block of metadata providing information indicative of the given access sequence to be stored in a selected entry of the cache. Eviction control circuitry, responsive to a victimisation event, performs an operation to select a victim entry in the cache, the victim entry being selected from one or more candidate victim entries. Each entry has an associated age indication value used to determine whether that entry is allowed to be a candidate victim entry, and the eviction control circuitry is arranged to perform a dynamic ageing operation to determine an ageing control value used to control updating of the associated age indication value for any entry storing a block of metadata. The dynamic ageing operation is arranged to determine the ageing control value in dependence on at least a training rate indication for the prefetch circuitry, where the training rate indication is indicative of a number of training inputs per memory access operation performed by the processing circuitry.

Technique for controlling use of a cache to store prefetcher metadata

Apparatuses and methods relating to controlling cache evictions are disclosed. Processing circuitry which execute instructions out-of-order is provided with a private cache into which blocks of data are copied from a shared storage location to which the processing circuitry shares access. The processing circuitry also has a read-after-read buffer, into which an entry is allocated when out-of-order execution of a load instruction occurs comprising an address accessed by the load instruction. The address remains as a valid entry in the read-after-read buffer until the load instruction is committed. Eviction of an eviction candidate block of data from the private cache to the shared storage location is controlled in dependence on whether the eviction candidate block of data has a corresponding valid entry in the read-after-read buffer.

Cache eviction control for a private cache in an out-of-order data processing apparatus

Data processing circuitry comprises out-of-order instruction execution circuitry; register mapping circuitry to map zero or more architectural processor registers relating to execution of that program instruction to respective ones of a set of physical processor registers; commit circuitry to commit, in a program code order, the results of executed program instructions, the commit circuitry being configured to access a data store which stores register tag data to indicate which physical registers mapped by the register mapping circuitry relate to a given program instruction; fault detection circuitry to detect a memory access fault in respect of a vector memory access operation and to generate fault indication data indicative of an element earliest in the element order for which a memory access fault was detected; a fault indication register to store the fault indication data, in which the register mapping circuitry is configured to generate a register mapping for a program instruction for any architectural processor registers relating to execution of that program instruction other than the fault indication register; and control circuitry to encode the fault indication data, applicable to a program instruction not yet committed by the commit circuitry, to register tag data associated with that program instruction.

Data processing

Data processing apparatus comprises a processing element configured to access an architectural register representing a given system register; mapping circuitry to map the architectural register representing the given system register to a physical register selected from a set of physical registers; a register bank having a set of two or more respective banked versions of the given system register, in which a respective one of the banked versions of the system register is associated with each of a plurality of current operating states of the processing element; in which, when the processing element changes operating state from a first operating state associated with a first one of the banked versions of the system register to a second operating state associated with a second, different, one of the banked versions of the system register, the processing element is configured to store the current contents of the architectural register representing the given system register to the first one of the banked versions of the system register and to copy the contents of the second one of the banked versions of the system register to the architectural register representing the given system register.

Data processing apparatus with respective banked registers for exception levels

An apparatus comprises execution circuitry to perform operations on source data values and to generate result data values; issue circuitry comprising one or more issue queues identifying pending operations awaiting performance by the execution circuitry, and selection circuitry to select pending operations to issue to the execution circuitry; data value cache storage comprising first and second cache regions; and cache control circuitry to control the storing to the first cache region of result data values generated by the execution circuitry and the eviction of stored result data values from the first cache region in response to newly generated result data values being stored in the first cache region; the cache control circuitry being configured to store to the second cache region result data values required as source data values for one or more oldest pending operations identified by the one or more issue queues and to inhibit eviction of a given result data value stored in the second cache region until initiation of execution of a pending operation which requires that given result data value as a source data value.

Cache control circuitry and methods

An apparatus has a processing pipeline, and first and second register files. A temporary-register-using instruction is supported which controls the pipeline to perform an operation using a temporary variable derived from an operand stored in the first register file. In response to the instruction, when a predetermined condition is not satisfied, the pipeline processes at least one register move micro-operation to transfer data from the at least one source register of the first register file to at least one newly allocated temporary register of the second register file. When the condition is satisfied, the operation can be performed using a temporary variable already stored in the temporary register of the second register file used by an earlier temporary-register-using instruction specifying the same source register for determining the temporary variable, in the absence of an intervening instruction for rewriting the source register.

Processing of a temporary-register-using instruction including determining whether to process a register move micro-operation for transferring data from a first register file to a second register file based on whether a temporary variable is still available in the second register file

Apparatus comprises counter and bit-shift circuitry to provide a succession of processing stages each comprising a count operation stage and a corresponding bit-shift stage, each processing stage operating with respect to a set of contiguous n-bit groups of bit positions, where n is 1 for a first processing stage and n doubles from one processing stage in the succession of processing stages to a next processing stage in the succession of processing stages; each count operation stage being configured to generate, for a first set of alternate instances of the n-bit groups of bit positions, count values indicating a respective number of bits of a predetermined bit value in a mask data word; and each bit-shift stage being configured to generate a bit-shifted data word by bit-shifting bits of a data word to be processed, for a second set of alternate instances of the n-bit groups of bit positions complementary to the first set, by respective numbers of bit positions dependent upon the count values generated by the respective count operation stage, in which the bit-shifted data word for one bit-shift stage in the succession of processing stages is used as the data word to be processed by the next bit-shift stage in the succession of processing stages.

Growing community of inventors

Antibes, France

Luca Nassi