Abstract: A processing system includes a plurality of compute units, with each compute unit having an associated first cache of a plurality of first caches, and a second cache shared by the plurality of compute units. The second cache operates to manage transfers of caches between the first caches of the plurality of first caches such that when multiple candidate first caches contain a valid copy of a requested cacheline, the second cache selects the candidate first cache having the shortest total path from the second cache to the candidate first cache and from the candidate first cache to the compute unit issuing a request for the requested cacheline.

Abstract: A mechanism for cache quota control is disclosed. A cache memory is configured to receive access requests from a plurality of agents, wherein a given request from a given agent of the plurality of agents specifies an identification value associated with the given agent of the plurality of agents. A cache controller is coupled to the cache memory, and is configured to store indications of current allocations of the cache memory to individual ones of the plurality of agents. The cache controller is further configured to track requests to the cache memory based on identification values specified in the requests and determine whether to update allocations of the cache memory to the individual ones of the plurality of agents based on the tracked requests.

Abstract: Systems, methods, and apparatuses relating to circuitry to implement a duplication resistant on-die irregular data prefetcher are described.

Abstract: A configuration, in which access processing such as data recording from an external device to the archive device can be efficiently performed, is implemented. A data processing section that performs a data recording processing control on a library, which is a data storage section of the archive device is provided. The data processing section saves record data in the local cache section in response to an input of the data recording request from the external device to the library, and outputs a recording processing completion notification to the external device. Further, an elapsed time after the saving in the local cache section is measured, and after the saving in the local cache section, the record data is transferred from the local cache section to the library and recorded after a pre-specified postponement time elapses.

Abstract: Systems, methods, and apparatuses relating to circuitry to implement a pipelined out of order page miss handler are described. In one embodiment, a hardware processor core includes an execution circuit to generate data storage requests for virtual addresses, a translation lookaside buffer to translate the virtual addresses to physical addresses, and a single page miss handler circuit comprising a plurality of pipelined page walk stages, wherein the single page miss handler circuit is to contemporaneously perform a first page walk within a first stage of the plurality of pipelined page walk stages for a first miss of a first virtual address in the translation lookaside buffer, and a second page walk within a second stage of the plurality of pipelined page walk stages for a second miss of a second virtual address in the translation lookaside buffer.

Abstract: In some examples, an electronic device includes a processor to determine a threshold magnitude. Based on the threshold magnitude, the processor is to determine a bit mapping for a partial timestamp. The partial timestamp is a sequence of bits of a timestamp that is based on a clock signal of the processor. The processor is to determine the partial timestamp based on the bit mapping and the timestamp, append the partial timestamp to a data signal, and cause transmission of the data signal.

Abstract: A cache may store critical cache lines and non-critical cache lines, and may attempt to retain critical cache lines in the cache by, for example, favoring the critical cache lines in replacement data updates, retaining the critical cache lines with a certain probability when victim cache blocks are being selected, etc. Criticality values may be retained at various levels of the cache hierarchy. Additionally, accelerated eviction may be employed if the threads previously accessing the critical cache blocks are viewed as dead.

Abstract: Systems and methods for configuring a virtual machine provided by a remote computing system based on the availability of one or more remote computing resources and respective corresponding prices of the one or more remote computing resources.

Abstract: One embodiment of a cache invalidation method includes storing an invalidation status usable by a computing node to identify, from a broadcast cache invalidation queue, a last processed invalidation that was processed with respect to an object cache used by the node. The method further comprises the node determining a set of unprocessed invalidations from the broadcast cache invalidation queue that are subsequent to the last processed invalidation determined from the invalidation status. The node processes the set of unprocessed invalidations to clear cached objects from the object cache. Based on processing the set of unprocessed invalidations to clear cached objects from the object cache, the invalidation status is updated with an identifier corresponding to a last invalidation from the set of previously unprocessed invalidations.

Abstract: Techniques for performing cache operations are provided. The techniques include tracking re-references for cache lines of a cache, detecting that eviction is to occur, and selecting a cache line for eviction from the cache based on a re-reference indication.

Abstract: A method, computer program product, and computer system for receiving, by a computing device, an I/O request. It may be identified whether the I/O request is eligible for handling via a first path without also requiring handling via a second path. If the I/O request is eligible, the I/O request may be processed via the first path on a host I/O stack without processing the I/O request via the second path on a storage array I/O stack. If the I/O request is ineligible, the I/O request may be processed via the first path on the host.

Abstract: Embodiments of methods and apparatuses for defending against speculative side-channel analysis on a computer system are disclosed. In an embodiment, a processor includes a decoder, a cache, address translation circuitry, a cache controller, and a memory controller. The decoder is to decode an instruction. The instruction is to specify a first address associated with a data object, the first address having a first memory tag. The address translation circuitry is to translate the first address to a second address, the second address to identify a memory location of the data object. The comparator is to compare the first memory tag and a second memory tag associated with the second address. The cache controller is to detect a cache miss associated with the memory location.

Abstract: In various embodiments, an isolation application determines processor assignment(s) based on a performance cost estimate. The performance cost estimate is associated with an estimated level of cache interference arising from executing a set of workloads on a set of processors. Subsequently, the isolation application configures at least one processor included in the set of processors to execute at least a portion of a first workload that is included in the set of workloads based on the processor assignment(s). Advantageously, because the isolation application generates the processor assignment(s) based on the performance cost estimate, the isolation application can reduce interference in a non-uniform memory access (NUMA) microprocessor instance.

Abstract: A network interface device comprises a programmable interface configured to provide a device interface with at least one bus between the network interface device and a host device. The programmable interface is programmable to support a plurality of different types of a device interface.

Abstract: This disclosure describes a process for securely instantiating a virtual machine on a server cluster. The virtual machine just after instantiation has access to persistent storage that includes an encrypted region and lacks access to an encryption key configured to provide access to data stored within the encrypted region. The virtual machine receives a communication from a management server associated with the server cluster that includes the encryption key configured to provide access to the data stored within the encrypted region. After the virtual machine receives the encryption key, the server cluster runs services that depend upon the data stored within the encrypted region to operate after receiving the communication from the management server.

Abstract: A method, non-transitory computer readable medium, and device that assists with reducing memory fragmentation in solid state devices includes identifying an allocation area within an address range to write data from a cache. Next, the identified allocation area is determined for including previously stored data. The previously stored data is read from the identified allocation area when it is determined that the identified allocation area comprises previously stored data. Next, both the write data from the cache and the read previously stored data are written back into the identified allocation area sequentially through the address range.

Abstract: In accordance with various aspects of the invention, a recall transaction is issued if a tag filter entry needs to be freed up for an incoming transaction. Directory entries chosen for a recall transaction are pushed into a fully associative structure called victim buffer. If this structure gets full, then an entry is selected from entries inside the victim buffer for the recall.

Abstract: The application discloses a memory controller coupled between a memory module and a host controller to control accesses of the host controller to the memory module. The memory controller comprises a central buffer coupled between the memory module and the host controller via a command/address channel, wherein the central buffer is configured to receive a command/address signal from the host controller and provide the command/address signal to the memory module.

Abstract: Control systems and methods for securing software images to be executed by the control system. In some examples the control system includes a control node, a concentrator node, and a security module. The security module may include a secured memory area that can hold software images. The security module may load a first executable image from the secured memory area to the control node. The security module may also load a second executable image from the secured memory area to the concentrator node. In some examples, rather than having a security module load the software images, the control node and concentrator node each include a secured area of memory where their respective software images reside. Each of the control node and concentrator node may load the software images from their respective secured areas of memory.

Abstract: In accordance with various aspects of the invention, a recall transaction is issued if a tag filter entry needs to be freed up for an incoming transaction. Directory entries chosen for a recall transaction are pushed into a fully associative structure called victim buffer. If this structure gets full, then an entry is selected from entries inside the victim buffer for the recall.

Abstract: Data transfer between processors is efficiently performed in a multiprocessor including a shared cache memory. Each entry in a tag storage section 220 of a cache memory holds a reference number field 224 in addition to a tag address field 221, a valid field 222, and a dirty field 223. The reference number field 224 is set in a data write, and the value thereof is decremented after each read access. When the value of the reference number field 224 is changed from “1” to “0”, the entry is invalidated without performing a write-back operation. When the cache memory is used for communication between processors in the multiprocessor system, the cache memory functions as a shared FIFO, and used data is automatically deleted.

Abstract: Various embodiments for data management across a multiple-tiered storage organization by a processor. Data operations performed across the multiple-tiered storage organization are analyzed over a period of time sufficient to determine usage patterns of the data. Predictive analytics is applied to the usage patterns. Based on the predictive analytics, segments of the data are moved between the multiple-tiered storage organization according to a determined priority account for available system resources, to optimize storage characteristics of the data in the multiple-tiered storage organization.

Abstract: Data transfer between processors is efficiently performed in a multiprocessor including a shared cache memory. Each entry in a tag storage section 220 of a cache memory holds a reference number field 224 in addition to a tag address field 221, a valid field 222, and a dirty field 223. The reference number field 224 is set in a data write, and the value thereof is decremented after each read access. When the value of the reference number field 224 is changed from “1” to “0”, the entry is invalidated without performing a write-back operation. When the cache memory is used for communication between processors in the multiprocessor system, the cache memory functions as a shared FIFO, and used data is automatically deleted.

Abstract: For each access request received at a shared cache of the data processing device, a memory access pattern (MAP) monitor predicts which of the memory banks, and corresponding row buffers, would be accessed by the access request if the requesting thread were the only thread executing at the data processing device. By recording predicted accesses over time for a number of access requests, the MAP monitor develops a pattern of predicted memory accesses by executing threads. The pattern can be employed to assign resources at the shared cache, thereby managing memory more efficiently.

Abstract: A technique for operating a processor includes translating, using an associated translation lookaside buffer, a first virtual address into a first physical address through a first entry number in the translation lookaside buffer. The technique also includes translating, using the translation lookaside buffer, a second virtual address into a second physical address through a second entry number in the translation lookaside buffer. The technique further includes, in response to the first entry number being the same as the second entry number, determining that the first and second virtual addresses point to the same physical address in memory and reference the same data.

Abstract: A system reserves and manages a hidden service partition through components of the hardware platform of a computing device. The hidden partition is not accessible by way of a host operating system on the computing device. A hardware platform controller provisions a portion of nonvolatile storage through configuration settings of the hardware platform controller. When the host system requests settings related to storage in the system, the request is routed through the interfaces of the hardware platform, and the hardware platform controller reports in accordance with the configuration settings, hiding the service partition. The hidden partition is dynamically modifiable through secure remote access to the hardware platform controller, not through the host system such as operating system or BIOS.

Abstract: The present disclosure relates to a cache memory controller for controlling a set-associative cache memory, in which two or more blocks are arranged in the same set, the cache memory controller including a content modification status monitoring unit for monitoring whether some of the blocks arranged in the same set of the cache memory have been modified in contents, and a cache block replacing unit for replacing a block, which has not been modified in contents, if some of the blocks arranged in the same set have been modified in contents.

Abstract: A storage device includes a non-volatile memory, a cache memory and a memory controller. The non-volatile memory stores a logical-to-physical address translation table for managing partitioned data and storage locations thereof. The cache memory stores a data cache and a logical-to-physical address translation table cache which holds a portion of the logical-to-physical address translation table. When the memory controller receives a data read-out request from outside, in the case no empty entry is found in the data cache, among the partitioned data in the data cache, it creates an empty entry to read out the data thereto by evacuating partitioned data of which entries in the logical-to-physical address translation table exist in the logical-to-physical address translation table cache into the non-volatile memory prior to other partitioned data.

Abstract: Embodiments of the invention provide a solution to optimize/minimize the total capacity of Gold Image within the entire datacenter which utilizes a scale-out type of storage systems. A method of server image provisioning comprises checking whether a gold image exists in a first storage system, the gold image being one of a real gold image or a virtual gold image; if no gold image exists in the first storage system, searching a remainder of the storage systems until a real gold image is found in a second storage system; after finding the real gold image in the second storage system, creating a virtual gold image in the first storage system, the virtual gold image in the first storage system being associated with the real gold image in the second storage system; and creating a snapshot volume in the first storage system based on the virtual gold image.

Abstract: A secure demand paging (SDP) system includes a dynamic random access memory (DRAM), a microprocessor having a secure internal memory and coupled to said DRAM, and a non-volatile memory storing a representation of operations accessible by the microprocessor. The stored representation of operations includes a coded physical representation of operations to configure an SDP space in the DRAM, to organize the SDP space into virtual machine contexts, to organize at least one of the virtual machine contexts into block book keeping blocks and book keeping spaces in the block book keeping blocks, and to execute a secure demand paging process between said secure internal memory and said DRAM.

Abstract: According to one general aspect, a method is provided for managing memory when counting unique items, the method using a pattern of bits in a unique estimator mask. The method may create a unique estimator mask based on fingerprints calculated for previously encountered items, and determine a number with the highest probability for creating the pattern of bits in the mask. When the number with the highest probability is determined, it may be returned as the estimated count of unique items.

Abstract: A system and method for managing data within a cache is described. In sonic example embodiments, the system identifies and/or tracks consumers of data located within a cache, and maintains the data within the cache based on determining whether there is an active consumer of the data.

Abstract: A processor includes multiple processor core units, each including a processor core and a cache memory. Victim lines evicted from a first processor core unit's cache may be stored in another processor core unit's cache, rather than written back to system memory. If the victim line is later requested by the first processor core unit, the victim line is retrieved from the other processor core unit's cache. The processor has low latency data transfers between processor core units. The processor transfers victim lines directly between processor core units' caches or utilizes a victim cache to temporarily store victim lines while searching for their destinations. The processor evaluates cache priority rules to determine whether victim lines are discarded, written back to system memory, or stored in other processor core units' caches. Cache priority rules can be based on cache coherency data, load balancing schemes, and architectural characteristics of the processor.

Abstract: Implementations described and claimed herein provide a system and methods for prioritizing data in a cache. In one implementation, a priority level, such as critical, high, and normal, is assigned to cached data. The priority level dictates how long the data is cached and consequently, the order in which the data is evicted from the cache memory. Data assigned a priority level of critical will be resident in cache memory unless heavy memory pressure causes the system to reclaim memory and all data assigned a priority state of high or normal has been evicted. High priority data is cached longer than normal priority data, with normal priority data being evicted first. Accordingly, important data assigned a priority level of critical, such as a deduplication table, is kept resident in cache memory at the expense of other data, regardless of the frequency or recency of use of the data.

Abstract: A technique for operating a processor includes translating, using an associated transaction lookaside buffer, a first virtual address into a first physical address through a first entry number in the transaction lookaside buffer. The technique also includes translating, using the transaction lookaside buffer, a second virtual address into a second physical address through a second entry number in the translation lookaside buffer. The technique further includes, in response to the first entry number being the same as the second entry number, determining that the first and second virtual addresses point to the same physical address in memory and reference the same data.

Abstract: Techniques and mechanisms for adaptively changing between replacement policies for selecting lines of a cache for eviction. In an embodiment, evaluation logic determines a value of a performance metric which is for writes to a non-volatile memory. Based on the determined value of the performance metric, a parameter value of a replacement policy is determined. In another embodiment, cache replacement logic performs a selection of a line of cache for data eviction, where the selection is in response to the policy unit providing an indication of the determined parameter value.

Abstract: Methods and apparatuses for processing speculative read requests in a system cache within a memory controller. To expedite a speculative read request, the request is sent on parallel paths through the system cache. A first path goes through a speculative read engine to determine if the speculative read request meets the conditions for accessing memory. A second path involves performing a tag lookup to determine if the data referenced by the request is already in the system cache. If the speculative read request meets the conditions for accessing memory, the request is sent to a miss queue where it is held until a confirm or cancel signal is received from the tag lookup mechanism.

Abstract: A processor, operable in a computing storage environment, allocates portions of a Scatter Index Table (SIT) disproportionately between a larger portion dedicated for meta data tracks, and a smaller portion dedicated for user data tracks, and processes a storage operation through the disproportionately allocated portions of the SIT using an allocated number of Task Control Blocks (TCB).

Abstract: A method for protecting page-level metadata in a storage system is provided. The method includes providing in a page table first protection data, receiving a command to read data from a page of the storage system corresponding to the page table, and comparing first protection data to second protection data. If the first protection data is different than the second protection data, then the method includes identifying third protection data in the storage system and comparing the third protection data to the first protection data. If the third protection data is different than the first protection data, then the method includes determining that the page-level metadata is inconsistent.

Abstract: In an embodiment, a page miss handler includes paging caches and a first walker to receive a first linear address portion and to obtain a corresponding portion of a physical address from a paging structure, a second walker to operate concurrently with the first walker, and a logic to prevent the first walker from storing the obtained physical address portion in a paging cache responsive to the first linear address portion matching a corresponding linear address portion of a concurrent paging structure access by the second walker. Other embodiments are described and claimed.

Abstract: Data storage and retrieval methods and apparatus are provided for facilitating data de-duplication for serial-access storage media such as tape. During data storage, input data is divided into a succession of chunks and, for each chunk, a corresponding data item is written to the storage media. The data item comprises the chunk data itself where it is the first occurrence of that data, and otherwise comprises a chunk-data identifier identifying that chunk of subject data. To facilitate reconstruction of the original data on read-back from the storage media a cache (50) is used together with a database (35R), stored on the media, that includes for each duplicated chunk, the location of the corresponding chunk of subject data.

Abstract: An invention is provided for protecting the data integrity of a cached storage device in an alternate operating system (OS) environment. The invention includes replacing an actual partition table for a disk with a dummy partition table. The dummy partition table is designed to render data on the disk inaccessible when the dummy partition table is used by an OS to access the data. During operation, the data on the disk can be accessed using information based on the actual partition table. In response to receiving a request to disable caching, the dummy partition table on the disk is replaced with the actual partition table, thus rendering the data on the formally cached disk accessible in an alternate OS environment where appropriate caching software is not present.

Abstract: A method for minimizing soft error rates within caches by controlling a memory scrubbing rate selectively for a cache memory at an individual bank level. More specifically, the disclosure relates to maintaining a predetermined sequence and process of storing all modified information of a cache in a subset of ways of the cache, based upon for example, a state of a modified indication within status information of a cache line. A cache controller includes a memory scrubbing controller which is programmed to scrub the subset of the ways with the modified information at a smaller interval (i.e., more frequently) compared to the rest of the ways with clean information (i.e., information where the information stored within the main memory is coherent with the information stored within the cache).

Abstract: A controller receives a request to perform a release space operation. A determination is made that a new discard scan has to be performed on a cache, in response to the received request to perform the release space operation. A determination is made as to how many task control blocks are to be allocated to the perform the new discard scan, based on how many task control blocks have already been allocated for performing one or more discard scans that are already in progress.

Abstract: A method of accessing data in a shared-memory, parallel-processing computing system, comprises, on a first processing unit, receiving a reference for a data structure stored in a memory and a first value of a generation attribute associated with the data structure, waiting to receive an exclusive lock on the data structure, obtaining an exclusive lock on the data structure, receiving a second value of a second generation attribute associated with the data structure; and accessing the data structure only if the first generation attribute value and the second generation attribute value are identical.

Abstract: Embodiments relate to automatic pattern-based operand prefetching. An aspect includes receiving, by prefetch logic in a processor, an operand cache miss from a pipeline of the processor. Another aspect includes determining that an entry in a history table corresponding to the operand cache miss exists based on an instruction address of the operand cache miss. Yet another aspect includes, based on determining that the entry corresponding to the operand cache miss exists in the history table, issuing a prefetch instruction for a second operand based on the determined entry in the history table, and writing the determined entry into a miss buffer.

Abstract: A method for reducing memory latency in a processor includes identifying an independent instruction (or cache miss instruction) and corresponding dependent instructions from a re-circulating issue window (RIW) when a cache miss is encountered. The cache miss instruction and corresponding dependent instructions are moved to a re-circulating issue buffer (RIB) and moved back to the RIW from the RIB for processing when the cache miss is resolved.

Abstract: A method for memory management is provided for a memory including a plurality of pages. The method comprises assigning in-use pages to in-use buckets according to use counts. The in-use buckets include a low in-use bucket for a lowest range of use counts, and a high in-use bucket for a highest range of use counts. The method comprises assigning free pages to free buckets according to use counts. The free buckets include a low free bucket for a lowest range of use counts, and a high free bucket for a highest range of use counts. The method maintains use counts for in-use pages. On a triggering event for a current in-use page, the method determines whether the use count of the current in-use page exceeds a hot swap threshold, and if so moves data in the current in-use page to a lead page in the low free bucket.

Abstract: A mechanism is provided for managing memory of a runtime environment executing on a virtual machine. The mechanism includes an elastic cache made of objects within heap memory of the runtime environment. When the runtime environment and virtual machine are not experiencing memory pressure from a hypervisor, the objects of the elastic cache may be used to temporarily store application-level cache data from applications running within the runtime environment. When memory pressure from the hypervisor is exerted, the objects of the elastic cache are re-purposed to inflate a memory balloon within heap memory of the runtime environment.

Abstract: A storage control apparatus of the present invention is able to duplicatively manage data in a cache memory even during maintenance work. When a memory package CMPK3 specified by a user is removed from the apparatus 1 (S2), a microprocessor 2 changes a pair that has been configured using CMPK2 and CMPK3 to a pair of CMPK2 and a free area of a CMPK1. As a result, received data (S5) is respectively written to multiple cache memories (S6, S7), and duplicatively managed.