International Business Machines Corporation

Arm Limited

Texas Instruments Corporation

Barry Duane Williamson

David Scott Ray

Abhishek Raja

Shih-Hsiung Stephen Tung

Miles Robert Dooley

Alexander K Spencer

Kevin Arthur Chiarot

Christopher McCall Durham

Gerard Richard Williams

Michael Filippo

Sharad Mehrotra

David Andrew Schroter

Adrian Montero

Huzefa Moiz Sanjeliwala

Muralidharan Santharaman Chinnakonda

A James Van Norstrand, Jr

Dq Nguyen

Jim E Phillips

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 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 storage apparatus for storing data is disclosed. The storage apparatus comprises: a plurality of stores having storage locations for storing data items, including a level one store and a level two store the storage apparatus having a hierarchy such that in response to an access request for accessing a data item the level one store is accessed and in response to detecting that the item is not stored in the level one store the level two store is accessed. The storage apparatus is configured to store a copy of at least some items in both of the one level one store and the level two store, the storage apparatus comprising a plurality of indicator storage elements associated with a corresponding plurality of storage locations of the level two store, a set value of an indicator stored in one of the indicator storage elements indicating that the corresponding stored data item is also stored in the level one store. The storage apparatus is configured such that in response to a request to allocate a storage location in the level two store, storage locations having a set indicator associated with them are not selected for eviction.

Line allocation in multi-level hierarchical data stores

The application describes a data processor operable to process data, and comprising: a cache in which a storage location of a data item within said cache is identified by an address, said cache comprising a plurality of storage locations and said data processor comprising a cache directory operable to store a physical address indicator for each storage location comprising stored data; a hash value generator operable to generate a generated hash value from at least some of said bits of said address said generated hash value having fewer bits than said address; a buffer operable to store a plurality of hash values relating to said plurality of storage locations within said cache; wherein in response to a request to access said data item said data processor is operable to compare said generated hash value with at least some of said plurality of hash values stored within said buffer and in response to a match to indicate a indicated storage location of said data item; and said data processor is operable to access one of said physical address indicators stored within said cache directory corresponding to said indicated storage location and in response to said accessed physical address indicator not indicating said address said data processor is operable to invalidate said indicated storage location within said cache.

Correction of incorrect cache accesses

Adjacent signal lines within the critical path of logic within an integrated circuit are checked for capacitive coupling induced signal delay variations resulting from concurrent signal transitions. When found, signal transition overlap is eliminated by delaying the clock edge (rising or falling) triggering the signal driving logic, without necessarily delaying the other clock edge. A delay circuit is incorporated into clock stages for the signal driving logic, and may be selectively actuated to delay the clock edge to particular signal driving logic circuits. Selection of signal lines in which signal transitions are to be delayed may be performed after manufacture of the integrated circuit, and iterative determinations may be required since signal adjustment may create new critical paths within the integrated circuit logic. Once the final signal adjustment configuration is determined, that configuration may be stored as a vector within a memory in the integrated circuit and read during power-up into a scan chain controlling the individual delay circuits.

Post-manufacture signal delay adjustment to solve noise-induced delay variations

A method for analyzing a design or a model of a microprocessor chip model is provided. A base chip model is generated, and it is modified with wire-only changes to produce a modified chip model. The modified chip model is compared to the base chip model to discern the wire-only changes to form a delta chip model. A set of values for RC delays and net capacitance based on the delta chip model is produced, and the modified chip model is timed using the set of values for RC delays and net capacitance on the delta chip model and RC delays and net capacitance on the base chip model.

Method and system for analyzing wire-only changes to a microprocessor design using delta model

One aspect of the invention relates to an apparatus for processing a store instruction on a superscalar processor that employs in-order completion of instructions, the processor having an instruction dispatch unit, an architected register file, a rename register file, a load store unit, a completion unit and cache memory. In one embodiment of the invention, the apparatus includes a pointer queue having an entry corresponding to the store instruction, the entry containing a pointer to the entries in the architected and rename register files that contain data required by the store instruction; and a multiplexer coupled to read ports on the architected and rename register files so that data can be passed from one of the register files into an entry in a data queue, the data queue being coupled to the cache memory.

Method and apparatus for facilitating multiple storage instruction completions in a superscalar processor during a single clock cycle

The present invention is directed towards a means to detect and reorder out of order instructions that may violate data coherency. The invention comprises a mis-queue table for holding entries of instruction data, each entry corresponding to an instruction in a computer microprocesor. The instruction data in each entry comprises: i) address information for the instruction; ii) ordering information for the instruction, indicating the order of the instruction relative to other instructions in the mis-queue table; iii) data modification information for the instruction, for indicating a possibility of modified data; and iv) out of order information, for indicating that a newer instruction has completed before the corresponding older instruction to the entry. The invention also comprises an out of order comparator for comparing an address of a completed instruction to any address information entries in the miss queue. If a completed instruction accesses the same address as another instruction, as indicated in the address information in the mis-queue table, and the completed instruction is newer than the matched instruction, the out of order field is marked indicating this condition exists. The invention comprises a modification comparator. This compares addresses from data altering events to those addresses in the entries in the mis-queue table. On a match, the modification field of the corresponding entry is marked to indicate this condition exists. When an instruction entry indicates that the corresponding instruction's data is modified, and that the instruction is out of order, all subsequent instructions are canceled.

Apparatus and method for tracking out of order load instructions to

One aspect of the invention relates to a method for processing load instructions in a superscalar processor having a data cache and a register file. In one embodiment, the method includes the steps of dispatching a misaligned load instruction to access a block of data that is misaligned in the cache; while continuing to dispatch aligned instructions: generating a first access and a final access to the cache in response to the misaligned load instruction; storing data retrieved from the first access until data from the final access is available; reassembling the data from the first and final access into the order required by the load instruction; and storing the re-assembled data to the register file.

Apparatus and method for processing misaligned load instructions in a

One aspect of the invention relates to a method for operating a superscalar processor having an instruction cache, a sequencing unit, a load/store unit, a cache, an architectural register file and a rename register file. In one particular version of the invention, the method includes the steps of forwarding an instruction from the instruction cache to the sequencing unit operable to access multiple architectural registers; generating a plurality of primitive instructions responsive to the forwarded instruction in which an individual primitive instruction is operable to access an individual architectural register; and sequentially issuing the primitive instructions to move data between the data cache and the rename register file.

System for providing high performance speculative processing of complex

A method and system for a handling multiple store instruction completions in a processing system after a stall condition is disclosed. The processing system includes an instruction unit, the instruction unit including a dispatch unit and a completion unit, a translation unit and at least one execution unit. A load store unit comprises an instruction queue for receiving a plurality of instructions from the dispatch unit; at least one effective address (EA) unit for receiving the plurality of instructions from the instruction queue, and a store queue. The store queue is coupled to the translation unit, the at least one execution unit and the at least one EA unit. The store queue receives data and real address information relating to each of the plurality of instructions from the at least one execution unit prior to completion of each of the plurality of instructions. In so doing, the bottleneck associated with conventional systems, i.e., a maximum number of instructions that can be dispatched by the instruction unit is reduced.

Growing community of inventors

Cedar Park, TX, United States of America

Barry Duane Williamson