International Business Machines Corporation

Arm Limited

Tenstorrent Inc.

Tenstorrent Usa, Inc.

Miles Robert Dooley

Huzefa Moiz Sanjeliwala

Jeffrey A Stuecheli

Abhishek Raja

Michael Filippo

Joseph Michael Pusdesris

David Allen Hrusecky

Francis Patrick O'Connell

John Steven Dodson

Anthony Saporito

Benjiman L Goodman

Jody B Joyner

Eric Eugene Retter

Sheldon Bernard Levenstein

Stephen J Powell

Klas Magnus Bruce

Bruce Joseph Ronchetti

Jason Nathaniel Dale

Barry Duane Williamson

Adrian Montero

Milos Trajkovic

Matthew Andrew Rafacz

Scott Bruce Frommer

Stanislav Sokorac

Rakesh Shaji Lal

Hung Qui Le

Alper Buyuktosunoglu

Pradip Bose

Michael Stephen Floyd

Naresh Nayar

Randal Craig Swanberg

Sundeep Chadha

Richard James Eickemeyer

Bradly George Frey

Yaoqing Gao

Venkat Rajeev Indukuru

David Scott Ray

Luca Nassi

John Barry Griswell, Jr

Alexander Cole Shulyak

Geoffray Lacourba

Guy Larri

Chris Abernathy

Dam Sunwoo

Natalya Bondarenko

Joaquin Hinojosa

Balaji Vijayan

Stefano Ghiggini

Krishnendra Nathella

Alex E Mericas

Leigang Kou

Ho-Seop Kim

Todd Rafacz

Sharmila Shridhar

Richard J Eickemeyer, Jr

Methods and systems relating to improved coprocessing architectures with static instruction caching are disclosed herein. A disclosed coprocessor architecture includes a processor, a coprocessor defined by an instruction set, a cache of static instruction blocks, and an instruction cache. The instruction set includes at least one static instruction. The instruction cache is configured to identify instructions in a series of instructions, provided by the processor for execution by the coprocessor, as either static instructions or dynamic instructions. The instruction cache is configured to, in response to identifying two or more sequential instructions in the series of instructions as static instructions, cache the two or more sequential instructions in the cache of static instruction blocks as an instruction block and store a program counter value of a program counter. The instruction cache is configured to provide the instruction block to the coprocessor from the cache of static instruction blocks when the program counter returns to the program counter value.

Static instruction caching in a coprocessor architecture

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

Methods and systems relating to improved processing architectures with pre-staged instructions are disclosed herein. A disclosed processor includes a memory, at least one functional processing unit, a bus, a set of instruction registers configured to be loaded, using the bus, with a set of pre-staged instructions from the memory, and a logic circuit configured to provide the set of pre-staged instructions from the set of instruction registers to the at least one functional processing unit in response to receiving an instruction from the instruction memory.

Pre-staged instruction registers for variable length instruction set machine

Methods and systems relating to improved processing architectures with pre-staged instructions are disclosed herein. A disclosed processor includes an instruction memory, at least one functional processing unit, a bus, a set of instruction registers configured to be loaded, using the bus, with a set of pre-staged instructions from the instruction memory, and a logic circuit configured to provide the set of pre-staged instructions from the set of instruction registers to the at least one functional processing unit in response to receiving an instruction from the instruction memory.

A request node is provided, that includes request circuitry for issuing outgoing memory access requests to a remote node. Status receiving circuitry receives statuses regarding remote memory access requests at the remote node and control circuitry controls at least one of a rate or an aggression at which the outgoing memory access requests are issued to the remote node in dependence on at least some of the statuses. The control circuitry is inhibited from controlling the rate or the aggression until multiple statuses are received.

Controlling access requests of request nodes

An apparatus is provided that includes cache circuitry that comprises a plurality of cache lines. The cache circuitry treats one or more of the cache lines as trace lines each having correlated addresses and each being tagged by a trigger address. Prefetch circuitry causes data at the correlated addresses stored in the trace lines to be prefetched.

Correlated addresses and prefetching

An apparatus is described, comprising load issuing circuitry configured to issue load operations to load data from memory, and memory ordering tracking storage circuitry configured to store memory ordering tracking information on issued load operations. The apparatus also includes control circuitry configured to access the memory ordering tracking storage circuitry to determine, using the memory ordering tracking information, whether at least one load operation has been issued in disagreement with a memory ordering requirement, and, if so, to determine whether to re-issue one or more issued load operations or to continue issuing load operations despite disagreement with the memory ordering requirement. Furthermore, the control circuitry is capable of merging the memory ordering tracking information for a plurality of issued load operations into a merged entry in the memory ordering tracking storage circuitry.

Merging memory ordering tracking information for issued load instructions

A data processing apparatus is provided that includes a plurality of storage elements. Receiving circuitry receives a plurality of incoming data beats from cache circuitry and stores the incoming data beats in the storage elements. At least one existing data beat in the storage elements is replaced by an equal number of the incoming data beats belonging to a different cache line of the cache circuitry. The existing data beats stored in said plurality of storage elements form an incomplete cache line.

Cache storage

Storage circuitry is provided, that is designed to form part of a memory hierarchy. The storage circuitry comprises receiver circuitry for receiving a request to obtain data from the memory hierarchy. Transfer circuitry causes the data to be stored at a selected destination in response to the request, wherein the selected destination is selected in dependence on at least one selection condition. Tracker circuitry tracks the request while the request is unresolved. If at least one selection condition is met then the destination is the storage circuitry and otherwise the destination is other storage circuitry in the memory hierarchy.

Storage circuitry request tracking

Apparatuses and methods for prefetch generation are disclosed. Prefetching circuitry receives addresses specified by load instructions and can cause retrieval of a data value from an address before that address is received. Stride determination circuitry determines stride values as a difference between a current address and a previously received address. Plural stride values corresponding to a sequence of received addresses are determined. Multiple stride storage circuitry stores the plurality of stride values determined by the stride determination circuitry. New address comparison circuitry determines whether a current address corresponds to a matching stride value based on the plurality of stride values stored in the multiple stride storage circuitry. Prefetch initiation circuitry can causes a data value to be retrieved from a further address, wherein the further address is the current address modified by the matching stride value of the plurality of stride values. By the use of multiple stride values, more complex load address patterns can be prefetched.

Multiple stride prefetching

An apparatus and method are provided for efficient utilisation of an address translation cache. The apparatus has an address translation cache with a plurality of entries, where each entry stores address translation data used when converting a virtual address into a corresponding physical address of a memory system. Each entry identifies whether the address translation data stored therein is coalesced or non-coalesced address translation data, and also identifies a page size for a page within the memory system that is associated with that address translation data. Control circuitry is responsive to a virtual address, to perform a lookup operation within the address translation cache to produce, for each page size supported by the address translation cache, a hit indication to indicate whether a hit has been detected for an entry storing address translation data of the associated page size. The control circuitry is further arranged to determine, from at least each hit indication for a page size that is able to be associated with coalesced address translation data, a coalesced multi-hit indication which is set when a hit is detected for both an entry containing coalesced address translation data and for an entry containing non-coalesced address translation data. The control circuitry is then arranged, when the lookup operation has completed, to determine whether multiple hits have been detected, and in that instance to reference the coalesced multi-hit indication to determine whether multiple hits have resulted from both coalesced address translation data and non-coalesced address translation data in the address translation cache. This provides an efficient and precise mechanism for distinguishing between multiple hits caused by hardware coalescing and multiple hits caused by software induced issues.

Apparatus and method for efficient utilisation of an address translation cache

An apparatus and method are provided for maintaining address translation data within an address translation cache. The address translation cache has a plurality of entries, where each entry is used to store address translation data used when converting a virtual address into a corresponding physical address of a memory system. Control circuitry is used to perform an allocation process to determine the address translation data to be stored in each entry. The address translation cache is used to store address translation data of a plurality of different types representing address translation data specified at respective different levels of address translation within a multiple-level page table walk. The plurality of different types comprises a final level type of address translation data that identifies a full translation from the virtual address to the physical address, and at least one intermediate level type of address translation data that identifies a partial translation of the virtual address. The control circuitry is arranged, when performing the allocation process, to apply an allocation policy that permits each of the entries to be used for any of the different types of address translation data, and to store type identification data in association with each entry to enable the type of the address translation data stored therein to be determined. Such an approach enables very efficient usage of the address translation cache resources, for example by allowing the proportion of the entries used for full address translation data and the proportion of the entries used for partial address translation data to be dynamically adapted to changing workload conditions.

Apparatus and method for maintaining address translation data within an address translation cache

A data processing apparatus is provided including queue circuitry to respond to control signals each associated with a memory access instruction, and to queue a plurality of requests for data, each associated with a reference to a storage location. Resolution circuitry acquires a request for data, and issues the request for data, the resolution circuitry having a resolution circuitry limit. When a current capacity of the resolution circuitry is below the resolution circuitry limit, the resolution circuitry acquires the request for data by receiving the request for data from the queue circuitry, stores the request for data in association with the storage location, issues the request for data, and causes a result of issuing the request for data to be provided to said storage location. When the current capacity of the resolution circuitry meets or exceeds the resolution circuitry limit, the resolution circuitry acquires the request for data by examining a next request for data in the queue circuitry and issues a further request for the data based on the request for data.

Queuing memory access requests

A data processing systemincludes an address translation cacheto store a plurality of address translation entries. Eviction control circuitryselects a victim entry for eviction from address translation cacheusing an eviction control parameter. The address translation cachecan store multiple different types of entry corresponding to respective different levels of address translation within a multiple-level page table walk. The different types of entry have different eviction control parameters assigned at the time of allocation. Eviction from the address translation cache is dependent upon the entry type, as well as the subsequent accesses to the entry concerned and the other entries within the address translation cache.

Eviction control for an address translation cache

A cache is provided comprising a plurality of ways, each way of the plurality of ways comprising a data array, wherein a data item stored by the cache is stored in the data array of one of the plurality of ways. A way tracker of the cache has a plurality of entries, each entry of the plurality of entries for storing a data item identifier and for storing, in association with the data item identifier, an indication of a selected way of the plurality of ways to indicate that a data item identified by the data item identifier is stored in the selected way. Each entry of the way tracker is further for storing a miss indicator in association with the data item identifier, wherein the miss indicator is set by the cache when a lookup for a data item identified by that data item identifier has resulted in a cache miss. A corresponding method of caching data is also provided.

Tracking the content of a cache using a way tracker having entries with a cache miss indicator

Growing community of inventors

Austin, TX, United States of America

Miles Robert Dooley