Abstract: The present application presents a Uniform Memory Access (UMA) network including a cluster of UMA nodes each having at least one UMA memory unit and a server local to the at least one UMA memory unit. A respective UMA memory unit in a respective UMA node comprises persistent memory; non-persistent memory, a node control device operatively coupled to the persistent memory and the non-persistent memory, a local interface for interfacing with the local server in the respective UMA node, and a network interface for interfacing with the UMA network. The node control device is configured to translate between a local unified memory access (UMA) address space accessible by applications running on the local server and a global UMA address space that is mapped to a physical UMA address space. The physical UMA address space includes physical address spaces associated with different UMA nodes in the cluster of UMA nodes.

Abstract: Methods, systems, and apparatus, including an apparatus for determining pixel coordinates for image transformation and memory addresses for storing the transformed image data. In some implementations, a system includes a processing unit configured to perform machine learning computations for images using a machine learning model and pixel values for the images, a storage medium configured to store the pixel values for the images, and a memory address computation unit that includes one or more hardware processors. The processor(s) are configured to receive image data for an image and determine that the dimensions of the image do not match the dimensions of the machine learning model. In response, the processor(s) determine pixel coordinates for a transformed version of the input image and, for each of the pixel coordinates, memory address(es), in the storage medium, for storing pixel value(s) that will be used to generate an input to the machine learning model.

Abstract: A system and method for cache coherency within multiprocessor environments is provided. Each node controller of a plurality of nodes within a multiprocessor system receives a cache coherency protocol request from local processor sockets and other node controller(s). A ternary content addressable memory (TCAM) accelerator in the node controller determines if the cache coherency protocol request comprises a snoop request and, if it is determined to be a snoop request, searching the TCAM based on an address within the cache coherency protocol request. In response to detecting only one match between an entry of the TCAM and the received snoop request, sending a response to the requesting local processor a response without having to access a coherency directory.

Abstract: Generally, the present disclosure is directed to systems and methods for streaming processing within one or more systems of an autonomy computing system. When an update for a particular object or region of interest is received by a given system, the system can control transmission of data associated with the update as well as a determination of other aspects by the given system. For example, the system can determine based on a received update for a particular aspect and a priority classification and/or interaction classification determined for that aspect whether data associated with the update should be transmitted to a subsequent system before waiting for other updates to arrive.

Abstract: A system includes a hierarchical metadata predictor and a processing circuit. The hierarchical metadata predictor includes a first-level metadata predictor and a second-level metadata predictor. The processing circuit is configured to perform a plurality of operations including storing new or updated metadata into the first-level metadata predictor and searching the first-level metadata predictor for a metadata prediction. Responsive to finding the metadata prediction in the first-level metadata predictor, the metadata prediction is output corresponding to an entry of the first-level metadata predictor that is a hit. One or more entries of the first-level metadata predictor that are non-hits are periodically written to the second-level metadata predictor. The first-level metadata predictor is updated based on locating the metadata prediction in the second-level metadata predictor.

Abstract: A cache memory includes a data array, a directory of contents of the data array that specifies coherence state information, and snoop logic that processes operations snooped from a system fabric by reference to the data array and the directory. The snoop logic, responsive to snooping on the system fabric a request of a first flush/clean memory access operation that specifies a target address, determines whether or not the cache memory has coherence ownership of the target address. Based on determining the cache memory has coherence ownership of the target address, the snoop logic services the request and thereafter enters a referee mode. While in the referee mode, the snoop logic protects a memory block identified by the target address against conflicting memory access requests by the plurality of processor cores until conclusion of a second flush/clean memory access operation that specifies the target address.

Abstract: A cache memory includes a data array, a directory of contents of the data array that specifies coherence state information, and snoop logic that processes operations snooped from a system fabric by reference to the data array and the directory. The snoop logic, responsive to snooping on the system fabric a request of a flush or clean memory access operation of an initiating coherence participant, determines whether the directory indicates the cache memory has coherence ownership of a target address of the request. Based on determining the directory indicates the cache memory has coherence ownership of the target address, the snoop logic provides a coherence response to the request that causes coherence ownership of the target address to be transferred to the initiating coherence participant, such that the initiating coherence participant can protect the target address against conflicting requests.

Abstract: Systems for prosecuting Internet messaging campaigns. Two or more data sources are determined where at least one of the data sources comprise demographic attributes corresponding to shared IDs such as recipient IDs. A first join operation is performed over matching instances of the shared IDs in the two or more data sources. The first join operation results in a personalization table comprising rows having at least recipient IDs, respective external addresses, and at least one of the demographic attributes. The personalization table is transformed into a key-value data structure that is published to a caching subsystem. The caching subsystem is used to select a first set of recipients determined without performing a second join operation. Personalized messages to at least some of the first and second set of recipients are formed using the message template and the key-value data structures.

Abstract: Systems, apparatuses, and methods for maintaining a region-based cache directory are disclosed. A system includes multiple processing nodes, with each processing node including a cache subsystem. The system also includes a cache directory to help manage cache coherency among the different cache subsystems of the system. In order to reduce the number of entries in the cache directory, the cache directory tracks coherency on a region basis rather than on a cache line basis, wherein a region includes multiple cache lines. Accordingly, the system includes a region-based cache directory to track regions which have at least one cache line cached in any cache subsystem in the system. The cache directory includes a reference count in each entry to track the aggregate number of cache lines that are cached per region. If a reference count of a given entry goes to zero, the cache directory reclaims the given entry.

Abstract: Techniques for activity tracking, data classification, and in-database archiving are described. Activity tracking refers to techniques that collect statistics related to user access patterns, such as the frequency or recency with which users access particular database elements. The statistics gathered through activity tracking can be supplied to data classification techniques to automatically classify the database elements or to assist users with manually classifying the database elements. Then, once the database elements have been classified, in-database archiving techniques can be employed to move database elements to different storage tiers based on the classifications. However, although the techniques related to activity tracking, data classification, and in-database archiving may be used together as described above; each technique may also be practiced separately.

Abstract: A processing unit includes a processor core that executes a store-conditional instruction that generates a store-conditional request specifying a store target address. The processing unit further includes a reservation register that records shared memory addresses for which the processor core has obtained reservations and a cache that services the store-conditional request by conditionally updating the shared memory with the store data based on the reservation register indicating a reservation for the store target address. The cache includes a blocking state machine configured to protect the store target address against access by any conflicting memory access request snooped on a system interconnect during a protection window extension following servicing of the store-conditional request.

Abstract: In accordance with an embodiment, described herein is a system and method for connection concentration in a database environment. A transparency engine provided between client applications and a database can include a connection pool (e.g., UCP connection pool). The transparency engine can operate as a proxy engine for the database and as a session abstraction layer for the client applications, to enable the client applications to utilize features provided by the connection pool without code changes. The transparency engine can receive application connections from the client applications, and concentrate the application connections on a smaller number of database connections maintained in the connection pool.

Abstract: Methods testing a data coherency algorithm via a simulated multi-processor environment are provided, which include implementing: (i) a transactional footprint keeping the address of each cache line used by the processor core, (ii) a reference model operating on and keeping a set of timestamps for a cache line, the set including a construction date representing a global timestamp when new data arrives at a private cache hierarchy and an expiration date representing another global timestamp when a cross-invalidation hits the private cache hierarchy, (iii) a core observed timestamp representing a global timestamp of an oldest construction date of data used before, and (iv) interface events monitoring instruction sequences guaranteed by transactional execution to ensure atomicity of a transaction. Upon detecting a transaction end event and finding a cache line of the transactional footprint having an expiration date older than or equal to a core observed time, an error is reported.

Abstract: A network processor includes a memory subsystem serving a plurality of processor cores. The memory subsystem includes a hierarchy of caches. A mid-level instruction cache provides for caching instructions for multiple processor cores. Likewise, a mid-level data cache provides for caching data for multiple cores, and can optionally serve as a point of serialization of the memory subsystem. A low-level cache is partitionable into partitions that are subsets of both ways and sets, and each partition can serve an independent process and/or processor core.

Abstract: Data management for edge architected computing systems extends current storage and memory schemes of edge resources to expose interfaces to allow a device, such as an endpoint or client device, or another edge resource, to specify criteria for managing data originating from the device and stored in an edge resource, and extends the storage and memory controllers to manage data in accordance with the criteria, including removing stored data that no longer satisfies the criteria. The criteria includes a temporal hint to specify a time after which the data can be removed, a physical hint to specify a list of edge resources outside of which the data can be removed, an event-based hint to specify an event after which the data can be removed, and a quality of service condition to modify the time specified in the temporal hint based on a condition, such as memory and storage capacity of the edge resource in which the data is managed.

Abstract: An apparatus to switch a central processing unit between operational modes includes, in one embodiment, a central processing unit (“CPU”) having at least a first operation mode and a second operation mode, where the second operation mode is a higher performance operation mode than the first operation mode. The apparatus also includes a switching unit that switches a state of the CPU to the second operation mode in response to starting one of an operating system or an application program based on a user operation in a state in which the first operation mode is set, and switches the state of the CPU to the first operation mode in response to a determination that a condition is met.

Abstract: A system and method dynamically transitions the file system role of compute nodes in a distributed clustered file system for an object that includes an embedded compute engine (a storlet). Embodiments of the invention overcome prior art problems of a storlet in a distributed storage system with a storlet engine having a dynamic role module which dynamically assigns or changes a file system role served by the node to a role which is more optimally suited for a computation operation in the storlet. The role assignment is made based on a classification of the computation operation and the appropriate filesystem role that matches computation operation. For example, a role could be assigned which helps reduce storage needs, communication resources, etc.

Abstract: A symmetrical multi-processing (SMP) node, a distributed SMP (DSMP) system comprising a plurality of SMP nodes, and a method implemented in the SMP node are disclosed. The SMP node comprises: a plurality of processors, a memory coupled to the plurality of processors, and a memory coherent proxy coupled to the plurality of processors through a coherent accelerator interface. The memory coherent proxy is configured to manage statuses of cache lines in the memory.

Abstract: Embodiments are directed towards managing and tracking item identification of a plurality of items to determine if an item is a new or existing item, where an existing item has been previously processed. In some embodiments, two or more item identifiers may be generated. In one embodiment, generating the two or more item identifiers may include analyzing the item using a small item size characteristic, a compressed item, or for an identifier collision. The two or more item identifiers may be employed to determine if the item is a new or existing item. In one embodiment, the two or more item identifiers may be compared to a record about an existing item to determine if the item is a new or existing item. If the item is an existing item, then the item may be further processed to determine if the existing item has actually changed.

Abstract: An apparatus for using a dataless cache entry includes a cache memory and a cache controller configured to identify a first cache entry in cache memory as a potential cache entry to be replaced according to a cache replacement algorithm, compare a data value of the first cache entry to a predefined value, and write a memory address tag and state bits of the first cache entry to a dataless cache entry in response to the data value of the first cache entry matching the predefined value, wherein the dataless cache entry in the cache memory stores a memory address tag and state bits associated with the memory address, wherein the dataless cache entry represents the predefined value, and wherein the dataless cache entry occupies fewer bits than the first cache entry.

Abstract: In one embodiment, a cache coherent system includes one or more agents (e.g. coherent agents) that may cache data used by the system. The system may include a point of coherency in a memory controller in the system, and thus the agents may transmit read requests to the memory controller to coherently read data. The point of coherency may determine if the data is cached in another agent, and may transmit a copy back request to the other agent if the other agent has modified the data. The system may include an interconnect between the agents and the memory controller. At a point on the interconnect at which traffic from the agents converges, a copy back response may be converted to a fill for the requesting agent.

Abstract: Maintaining multiple cache areas in a storage device having multiple processors includes loading data from a specific portion of non-volatile storage into a local cache slot in response to a specific processor of a first subset of the processors performing a read operation to the specific portion of non-volatile storage, where the local cache slot is accessible to the first subset of the processors and is inaccessible to a second subset of the processors that is different than the first subset of the processors and includes converting the local cache slot into a global cache slot in response to one of the processors performing a write operation to the specific portion of non-volatile storage, wherein the global cache area is accessible to the first subset of the processors and to the second subset of the processors. Different ones of the processors may be placed on different directors.

Abstract: A memory arrangement having a memory, a first buffer memory, a first buffer memory controller which, during the storage of memory contents from the memory in the first buffer memory, is configured to invalidate the memory contents in the memory by means of a modification, a second buffer memory and a second buffer memory controller which is configured to read memory contents from the memory, to check whether the memory contents read from the memory are valid and, if the memory contents read from the memory are invalid, to apply a reversal of the modification to the read memory contents.

Abstract: A method for determining affinity domain information based on virtual memory address in a computing system where access to memory is non-uniform includes receiving a request to identify an affinity domain associated with a specified virtual memory address. The affinity domain includes a cluster of processors and memory local to the cluster of processors. A physical memory page corresponding to the specified virtual memory address is determined using a page table mapping a plurality of virtual memory addresses to a plurality of physical addresses. An affinity domain associated with the determined physical memory page is identified. Affinity domain information is provided for the identified affinity domain.

Abstract: Managing the timing of publication of new webpages and in particular new versions of existing webpages. New webpages are uploaded into a data repository for storing them before they are made available for access externally. A dependency processor is provided to process these new webpages to assess their readiness for publication by checking for dependencies on other webpages, the dependencies including a mutual dependency; locating any of the other webpages; and ascertaining whether each such dependency is satisfied. If dependencies are satisfied, then the new webpage is deemed ready for publication and is published. The satisfied dependencies include the content being accessible. In the case that the new webpage is a new version of an existing webpage, it replaces the old version. If dependencies are not satisfied, then the new webpage is held back until such time as they are met.

Abstract: According to various aspects, a memory system may include: a memory having a memory address space associated therewith to access the memory; a cache memory assigned to the memory; one or more processors configured to generate a dummy address space in addition to the memory address space, each address of the dummy address space being distinct from any address of the memory address space, and generate one or more invalid cache entries in the cache memory, the one or more invalid cache entries referencing one or more dummy addresses of the dummy address space.

Abstract: A transaction is detected. The transaction has a begin-transaction indication and an end-transaction indication. If it is determined that the begin-transaction indication is not a no-speculation indication, then the transaction is processed.

Abstract: Address-based filtering for load/store speculation includes maintaining a filtering table including table entries associated with ranges of addresses; in response to receiving an ordering check triggering transaction, querying the filtering table using a target address of the ordering check triggering transaction to determine if an instruction dependent upon the ordering check triggering transaction has previously been generated a physical address; and in response to determining that the filtering table lacks an indication that the instruction dependent upon the ordering check triggering transaction has previously been generated a physical address, bypassing a lookup operation in an ordering violation memory structure to determine whether the instruction dependent upon the ordering check triggering transaction is currently in-flight.

Abstract: Embodiments of the invention are directed to methods for handling cache. The method includes retrieving a plurality of instructions from a cache. The method further includes placing the plurality of instructions into an instruction fetch buffer. The method includes retrieving a first instruction of the plurality of instructions from the instruction fetch buffer. The method includes executing the first instruction. The method includes retrieving a second instruction from the plurality of instructions from the instruction fetch buffer unless a back invalidate is received from the cache. Thereafter executing the second instruction without refreshing the instruction fetch buffer from the cache.

Abstract: A method for object classification in a decision tree based adaptive boosting (AdaBoost) classifier implemented on a single-instruction multiple-data (SIMD) processor is provided that includes receiving feature vectors extracted from N consecutive window positions in an image in a memory coupled to the SIMD processor and evaluating the N consecutive window positions concurrently by the AdaBoost classifier using the feature vectors and vector instructions of the SIMD processor, in which the AdaBoost classifier concurrently traverses decision trees for the N consecutive window positions until classification is complete for the N consecutive window positions.

Abstract: Embodiments of the invention are directed to methods for handling cache. The method includes retrieving a plurality of instructions from a cache. The method further includes placing the plurality of instructions into an instruction fetch buffer. The method includes retrieving a first instruction of the plurality of instructions from the instruction fetch buffer. The method includes executing the first instruction. The method includes retrieving a second instruction from the plurality of instructions from the instruction fetch buffer unless a back invalidate is received from the cache. Thereafter executing the second instruction without refreshing the instruction fetch buffer from the cache.

Abstract: Methods, systems and computer program products for heuristically invalidating non-useful entries in an array are provided. Aspects include receiving an instruction that is associated with an operand store compare (OSC) prediction for at least one of a store function and a load function. The OSC prediction is stored in an entry of an OSC history table (OHT). Aspects also include executing the instruction. Responsive to determining, based on the execution of the instruction, that data forwarding did not occur, aspects include incrementing a useless OSC prediction counter. Responsive to determining that the useless OSC prediction counter is equal to a predetermined value, aspects also include invalidating the entry of the OHT associated with the instruction.

Abstract: An apparatus, system, and method of enforcing cache coherency in a multiprocessor shared memory system are disclosed. A request is received from a node controller, to process a cache coherent operation on a memory block in a shared memory. Based on the information included in the request, a determination is made as to whether the request was transmitted from a processor that is remote relative to the memory that includes the memory block referenced in the request. If the request is from a remote processor, a hardware-based cache coherency of the system is disabled, and request is processed according to software-based cache coherency protocols and mechanisms. A coherent read request may be translated to a non-coherent request, such as an immediate read request, which does not trigger tracking or storing state and ownership information of the requested memory block, or trigger communications with processors other than those involved with request.

Abstract: According to one embodiment of the invention, a processor includes a power control unit, an interface to software during runtime that permits the software to set a plurality of power management constraint parameters for the power control unit during runtime of the processor without a reboot of the processor, and a storage element to store a respective lock bit for each of the plurality of power management constraint parameters to disable the interface from changing a respective constraint parameter when set.

Abstract: One embodiment provides for a general-purpose graphics processing device comprising a general-purpose graphics processing compute block to process a workload including graphics or compute operations, a first cache memory, and a coherency module enable the first cache memory to coherently cache data for the workload, the data stored in memory within a virtual address space, wherein the virtual address space shared with a separate general-purpose processor including a second cache memory that is coherent with the first cache memory.

Abstract: A data writing method, a memory storage device, and a memory control circuit unit are provided. The method includes: receiving a first data and writing the first data to at least one first physical programming unit of a first physical erasing unit; receiving a second data; temporarily storing the second data to a temporary storage area if a data length of the second data is less than a predefined value; receiving a third data; writing the third data to at least one second physical programming unit of the first physical erasing unit if a logical address storing the first data is consecutive with a logical address storing the third data; and moving the second data from the temporary storage area to at least one second physical programming unit of the first physical erasing unit if the logical address storing the first data is not consecutive with the logical address storing the third data.

Abstract: A method, computer program product, and computer system are disclosed that in one or more embodiments includes issuing, from an issuing processor in the computer system, an address translation invalidation instruction with a return marker, wherein the address translation invalidation instruction is to invalidate one or more address translation entries in one or more storage locations in the computer system and wherein the return marker comprises an instruction to return information to the issuing processor indicating the identity of each processor where an invalidated entry was located.

Abstract: Security key identifier remapping includes associating a system-level security key identifier to a local-level identifier requiring fewer bits of storage space. The remapped security key identifiers are used to receive, at a first compute complex of a processing system, a memory access request including a memory address value and a system-level security key identifier. The compute complex responds to the memory access request based on a determination of whether a security key identifier map of the first compute complex includes a mapping of the system-level security key identifier to a local-level security key identifier. In response to determining that the security key identifier map of the first compute complex does not include a mapping of the system-level security key identifier to the local-level security key identifier, a cache miss message may be returned without probing caches of the first compute complex.

Abstract: Methods and systems for collision handling during an asynchronous replication are provided. A system includes a cache memory system comprising a number of cache memory pages. A collision detector detects when a host is attempting to overwrite a cache memory page that has not been completely replicated. A revision page tagger copies the cache memory page to a free page and tags the copied page as protected.

Abstract: Systems, apparatuses, and methods for accelerating accesses to private regions in a region-based cache directory scheme are disclosed. A system includes multiple processing nodes, one or more memory devices, and one or more region-based cache directories to manage cache coherence among the nodes' cache subsystems. Region-based cache directories track coherence on a region basis rather than on a cache line basis, wherein a region includes multiple cache lines. The cache directory entries for regions that are only accessed by a single node are cached locally at the node. Updates to the reference count for these entries are made locally rather than sending updates to the cache directory. When a second node accesses a first node's private region, the region is now considered shared, and the entry for this region is transferred from the first node back to the cache directory.

Abstract: A cache system, having a first cache, a second cache, and a logic circuit coupled to control the first cache and the second cache according to an execution type of a processor. When an execution type of a processor is a first type indicating non-speculative execution of instructions and the first cache is configured to service commands from a command bus for accessing a memory system, the logic circuit is configured to copy a portion of content cached in the first cache to the second cache. The cache system can include a configurable data bit. The logic circuit can be coupled to control the caches according to the bit. Alternatively, the caches can include cache sets. The caches can also include registers associated with the cache sets respectively. The logic circuit can be coupled to control the cache sets according to the registers.

Abstract: In accordance with an embodiment, described herein is a system and method for row buffering in a database environment. A transparency engine can be provided between client applications and a database, and can operate as a proxy engine for the database and as a session abstraction layer for the client applications, to enable the client applications to utilize database features provided by the connection pool without code changes to the client applications. The transparency engine can maintain a plurality of local row buffers to store rows fetched from a database. The local buffers can be filled by rows pre-fetched from the database. When a client application requests rows from the database, the transparency engine can first check whether the rows exist in a local buffer. If the rows are present in the local buffer, the transparency engine sends the rows to the requesting client application, without going to the database.

Abstract: In one example, a method of data localization in a hyperconverged virtual computing platform is described, which includes, determining whether a logical block address (LBA) associated with a storage request received by a node maps to another one of the plurality of nodes. The page associated with the storage request is then migrated from the other one of the plurality of nodes to the node based on a recent page hit count associated with the storage request when the LBA associated with the storage request is from another one of the plurality of nodes. Mapping layers residing in each of the plurality of nodes including the remapped LBA associated with the storage request are then updated. The storage request is resolved at the node if the LBA associated with the storage is found in the updated mapping layer associated with the node.

Abstract: Facilitating page table entry (PTE) maintenance in processor-based devices is disclosed. In this regard, a processor-based device includes processing elements (PEs) configured to support two new coherence states: walker-readable (W) and modified walker accessible (MW). The W coherence state indicates that read access to a corresponding coherence granule by hardware table walkers (HTWs) is permitted, but all write operations and all read operations by non-HTW agents are disallowed. The MW coherence state indicates that cached copies of the coherence granule visible only to HTWs may exist in other caches. In some embodiments, each PE is also configured to support a special page table entry (SP-PTE) field store instruction for modifying SP-PTE fields of a PTE, indicating to the PE's local cache that the corresponding coherence granule should transition to the MW state, and indicating to remote local caches that copies of the coherence granule should update their coherence state.

Abstract: A processing unit for a data processing system includes a cache memory having reservation logic and a processor core coupled to the cache memory. The processor includes an execution unit that executes instructions in a plurality of concurrent hardware threads of execution including at least first and second hardware threads. The instructions include, within the first hardware thread, a first load-reserve instruction that identifies a target address for which a reservation is requested. The processor core additionally includes a load unit that records the target address of the first load-reserve instruction and that, responsive to detecting, in the second hardware thread, a second load-reserve instruction identifying the target address recorded by the load unit, blocks the second load-reserve instruction from establishing a reservation for the target address in the reservation logic.

Abstract: The disclosure relates to embodiments, implemented at least partially in microcode, that use cache misses to trigger logging to a processor trace. One embodiment relies on tracking bits in a processor cache. During a transition from a non-logged context to a logged context, this embodiment invalidates or evicts cache lines whose tracking bits are not set. When logging, this first embodiment logs during cache misses, and sets tracking bits for logged cache lines. Another embodiment relies on way-locking. This second embodiment assigns first ways to a logged entity and second ways to a non-logged entity. The second embodiment ensures the logged entity cannot read cache lines from the second logging ways by flushing the second way during transitions from non-logging to logging, ensures the logged entity cannot read non-logged cache lines from the first ways, and logs based on cache misses into the first ways while executing a logged context.

Abstract: Techniques perform data recovery. The techniques involve: in response to receiving to-be-written data at a first cache module, storing metadata in the data into a first non-volatile cache of the first cache module. The techniques further involve storing user data in the data into a first volatile cache of the first cache module. The techniques further involve sending the metadata and the user data to a second cache module for performing data recovery on the user data. Accordingly, a larger and better guaranteed data storage space may be provided to a cache data backup/recovery system without a need to increase the battery supply capacity and even without a battery.

Abstract: A technology for implementing a method for distributed memory operations. A method of the disclosure includes obtaining distributed channel information for an algorithm to be executed by a plurality of spatially distributed processing elements. For each distributed channel in the distributed channel information, the method further associates one or more of the plurality of spatially distributed processing elements with the distributed channel based on the algorithm.

Abstract: Salvaging renderable content includes providing a set of salvaging instructions including a digital pattern associated with digital content to be salvaged, and a predetermined minimum threshold of usefulness of the digital content. A digital data source includes digital content to be salvaged. The digital content is simultaneously read by reviewing the multiple types of digital content independently of one another using separate software salvaging modules to review each specific type of digital content. The digital content is filtered by identifying potentially recoverable digital content. The digital pattern is compared to the filtered digital content to indicate matches between the filtered digital content and the digital pattern. The digital content is reassembled and/or repaired. The matched digital content is validated by determining whether the salvaged digital content is in a form that meets the predetermined minimum threshold of usefulness. The validated digital content is displayed/rendered.

Abstract: An integrated circuit for a coherent data processing system includes a first communication interface for communicatively coupling the integrated circuit with the coherent data processing system, a second communication interface for communicatively coupling the integrated circuit with an accelerator unit including an effective address-based accelerator cache for buffering copies of data from a system memory of the coherent data processing system, and a real address-based directory inclusive of contents of the accelerator cache. The real address-based directory assigns entries based on real addresses utilized to identify storage locations in the system memory. The integrated circuit further includes request logic that communicates memory access requests and request responses with the accelerator unit via the second communication interface.