Arm Limited

Nikos Nikoleris

Andreas Lars Sandberg

Stephan Diestelhorst

Prakash Shyamlal Ramrakhyani

Ilias Vougioukas

Wendy Arnott Elsasser

Radhika Sanjeev Jagtap

Shidhartha Das

David Hennah Mansell

Wei Wang

Frederic Claude Marie Piry

Peter Richard Greenhalgh

Ian Michael Caulfield

Albin Pierrick Tonnerre

Andreas Hansson

Ricardo Daniel Queiros Alves

Jonas S̆vedas

Apparatuses and methods for branch prediction are provided. Branch prediction circuitry generates prediction with respect to branch instructions of whether those branches will be taken or not-taken. Hypervector generation circuitry assigns an arbitrary hypervector in deterministic dependence on an address of each branch instruction, wherein the hypervectors comprises at least 500 bits. Upon the resolution of a branch a corresponding hypervector is added to a stored taken hypervector or a stored not-taken hypervector in dependence on the resolution of the branch. The branch prediction circuitry generates a prediction for a branch instructions in dependence on a mathematical distance metric of a hypervector generated for that branch instruction from the stored taken hypervector or the not-taken hypervector.

Branch prediction using hypervectors

Apparatus comprises two or more processing devices each having an associated translation lookaside buffer to store translation data defining address translations between virtual and physical memory addresses, each address translation being associated with a respective virtual address space; and control circuitry to control the transfer of at least a subset of the translation data from the translation lookaside buffer associated with a first processing device to the translation lookaside buffer associated with a second, different, processing device.

Apparatus and method

Coherency control circuitry () supports processing of a safe-speculative-read transaction received from a requesting master device (). The safe-speculative-read transaction is of a type requesting that target data is returned to a requesting cache () of the requesting master device () while prohibiting any change in coherency state associated with the target data in other caches () in response to the safe-speculative-read transaction. In response, at least when the target data is cached in a second cache associated with a second master device, at least one of the coherency control circuitry () and the second cache () is configured to return a safe-speculative-read response while maintaining the target data in the same coherency state within the second cache. This helps to mitigate against speculative side-channel attacks.

Cache control in presence of speculative read operations

A cache apparatus is provided comprising a data storage structure providing N cache ways that each store data as a plurality of cache blocks. The data storage structure is organised as a plurality of sets, where each set comprises a cache block from each way, and further the data storage structure comprises a first data array and a second data array, where at least the second data array is set associative. A set associative tag storage structure stores a tag value for each cache block, with that set associative tag storage structure being shared by the first and second data arrays. Control circuitry applies an access likelihood policy to determine, for each set, a subset of the cache blocks of that set to be stored within the first data array. Access circuitry is then responsive to an access request to perform a lookup operation within an identified set of the set associative tag storage structure overlapped with an access operation to access within the first data array the subset of the cache blocks for the identified set. In the event of a hit condition being detected that identifies a cache block present in the first data array, that access request is then processed using the cache block accessed within the first data array. If instead a hit condition is detected that identifies a cache block absent in the first data array, then a further access operation is performed to access the identified cache block within a selected way of the second data array. Such a cache structure provides a high performance and energy efficient mechanism for storing cached data.

Cache apparatus and method that facilitates a reduction in energy consumption through use of first and second data arrays

There is provided an apparatus that includes an input port to receive, from a requester, any one of: a lookup operation comprising an input address, and a maintenance operation. Maintenance queue circuitry stores a maintenance queue of at least one maintenance operation and address storage stores a translation between the input address and an output address in an output address space. In response to receiving the input address, the output address is provided in dependence on the maintenance queue. In response to storing the maintenance operation, the maintenance queue circuitry causes an acknowledgement to be sent to the requester. By providing a separate maintenance queue for performing the maintenance operation, there is no need for a requester to be blocked while maintenance is performed.

Performing maintenance operations

There is provided an apparatus that includes an input address port to receive an input address from processor circuitry. Address storage stores a translation between the input address and an output address in an output address space. An output address port outputs the output address. An input data port receives data. Data storage stores the data. An output data port outputs the data stored in the data storage and control circuitry causes the data storage to store the translation between the input address and the output address. The control circuitry issues a signal to cause a page walk to occur in response to the input address being absent from the address storage and the data storage.

Data storage for multiple data types

An apparatus comprises control circuitry to control access to a memory implemented using a memory technology providing variable access latency. The control circuitry has request handling circuitry to identify an execution context switch comprising a transition from servicing memory access requests associated with a first execution context to servicing memory access requests associated with a second execution context. At least when the execution context switch meets a predetermined condition, a delay masking action is triggered to control subsequent memory access requests associated with the second execution context, for which the required data is already stored in the memory, to be serviced with a response delay which is independent of which addresses were accessed by the memory access requests associated with the first execution context. This can help guard against attacks which aim to exploit variation in response latency to gain insight into the addresses accessed by a victim execution context.

Delay masking action for memory access requests

Storage circuitry comprises an array of storage locations arranged in rows and columns, a row buffer comprising a plurality of entries each to store information from a storage location at a corresponding column of an active row of the array, and comparison circuitry responsive to a tag-matching command specifying a tag value to compare the tag value with information stored in each of a subset of two or more entries of the row buffer. The comparison circuitry identifies which of the subset of entries, if any, is a matching entry storing information matching the tag value. This allows memory technologies such as DRAM to be used more efficiently as a set-associative cache.

Storage circuitry responsive to a tag-matching command

Memory address translation apparatus comprises page table access circuitry to access a page table to retrieve translation data; a translation data buffer to store one or more instances of the translation data, comprising: an array of storage locations arranged in rows and columns; a row buffer comprising a plurality of entries and comparison circuitry responsive to a key value dependent upon at least the initial memory address, to compare the key value with information stored in each of at least one key entry and an associated value entry for storing at least a representation of a corresponding output memory address, and to identify which of the at least one key entry, if any, is a matching key entry storing information matching the key value; and output circuitry to output, when there is a matching key entry, at least the representation of the output memory address.

Memory address translation using stored key entries

Memory address translation apparatus comprises page table access circuitry to access page table data to retrieve translation data defining an address translation between an initial memory address in an initial memory address space, and a corresponding output memory address in an output address space; a translation data buffer to store, for a subset of the virtual address space, one or more instances of the translation data; and control circuitry, responsive to an input initial memory address to be translated, to request retrieval of translation data for the input initial memory address from the translation data buffer and, before completion of processing of the request for retrieval from the translation data buffer, to initiate retrieval of translation data for the input initial memory address by the page table access circuitry.

Memory address translation

Aspects of the present disclosure relate to an apparatus comprising a data array having locality-dependent latency characteristics such that an access to an open unit of the data array has a lower latency than an access to a closed unit of the data array. Set associative cache indexing circuitry determines, in response to a request for data associated with a target address, a cache set index. Mapping circuitry identifies, in response to the index, a set of data array locations corresponding to the index, according to a mapping in which a given unit of the data array comprises locations corresponding to a plurality of consecutive indices, and at least two locations of the set of locations corresponding to the same index are in different units of the data array. Cache access circuitry accesses said data from one of the set of data array locations.

Methods and apparatus of cache access to a data array with locality-dependent latency characteristics

An apparatus and method are provided for transferring data between address ranges in memory. The apparatus comprises a data transfer controller, that is responsive to a data transfer request received by the apparatus from a processing element, to perform a transfer operation to transfer data from at least one source address range in memory to at least one destination address range in the memory. A redirect controller is then arranged, whilst the transfer operation is being performed, to intercept an access request that specifies a target address within a target address range, and to perform a memory redirection operation so as to cause the access request to be processed without awaiting completion of the transfer operation. Via such an approach, the apparatus can effectively hide from the source of the access request the fact that the transfer operation is in progress, and hence the transfer operation can be arranged to occur in the background, and in a manner that is transparent to the software executing on the source that has issued the access request.

Apparatus and method for transferring data between address ranges in memory

A TAGE branch predictor has, as its fallback predictor, a perceptron predictor. This provides a branch predictor which reduces the penalty of context switches and branch prediction state flushes.

TAGE branch predictor with perceptron predictor as fallback predictor

An apparatus and method are provided for handling caching of persistent data. The apparatus comprises cache storage having a plurality of entries to cache data items associated with memory address in a non-volatile memory. The data items may comprise persistent data items and non-persistent data items. Write back control circuitry is used to control write back of the data items from the cache storage to the non-volatile memory. In addition, cache usage determination circuitry is used to determine, in dependence on information indicative of capacity of a backup energy source, a subset of the plurality of entries to be used to store persistent data items. In response to an event causing the backup energy source to be used, the write back control circuitry is then arranged to initiate write back to the non-volatile memory of the persistent data items cached in the subset of the plurality of entries. By constraining the extent to which the cache storage is allowed to store persistent data items, taking into account the capacity of the backup energy source, the persistence of those data items can then be guaranteed in the event of the backup energy source being triggered, for example due to removal of the primary energy source for the apparatus.

Apparatus and method of handling caching of persistent data

A system cache and method of operating a system cache are provided. The system cache provides data caching in response to data access requests from plural system components. The system cache has data caching storage with plural entries, each entry storing a block of data items and each block of data items comprising plural sectors of data items. Sector use prediction circuitry is provided which stores a set of sector use pattern entries. In response to a data access request received from a system component specifying one or more data items, a pattern entry is selected and a sector use prediction is generated in dependence on a sector use pattern in the selected pattern entry. Further data items may then be retrieved which are not specified in the data access request but are indicated by the sector use prediction.

Growing community of inventors

London, United Kingdom

Nikos Nikoleris