47 results about "Shared virtual memory" patented technology

In a first embodiment of the present invention, a method for caching in a processor system having virtual memory is provided, the method comprising: monitoring slow memory in the processor system to determine frequently accessed pages; for a frequently accessed page in slow memory: copy the frequently accessed page from slow memory to a location in fast memory; and update virtual address page tables to reflect the location of the frequently accessed page in fast memory.

A method, system, and computer program product provide a shared virtual memory space via a cluster-aware virtual input / output (I / O) server (VIOS). The VIOS receives a paging file request from a first LPAR and thin-provisions a logical unit (LU) within the virtual memory space as a shared paging file of the same storage amount as the minimum required capacity. The VIOS also autonomously maintains a logical redundancy LU (redundant LU) as a real-time copy of the provisioned / allocated LU, where the redundant LU is a dynamic copy of the allocated LU that is autonomously updated responsive to any changes within the allocated LU. Responsive to a second VIOS attempting to read a LU currently utilized by a first VIOS, the read request is autonomously redirected to the logical redundancy LU. The redundant LU can be utilized to facilitate migration of a client LPAR to a different computing electronic complex (CEC).

In one embodiment, the present invention includes a memory management unit (MMU) having entries to store virtual address to physical address translations, where each entry includes a location indicator to indicate whether a memory location for the corresponding entry is present in a local or remote memory. In this way, a common virtual memory space can be shared between the two memories, which may be separated by one or more non-coherent links. Other embodiments are described and claimed.

The invention provides a Chinese short text classification method based on characteristic extension, and the method comprises the following steps that (1) a background knowledge base is established: the two-tuples of feature words which meet a certain constraint condition are dug from a long text corpus with category marks to form the background knowledge base; (2) short text which is trained in a centralized way is extended: extension words are added to the short text which is trained in a centralized way according to a certain extension rule according to the two-tuples in the background knowledge base; (3) a classification model is built: a (shared virtual memory) SVM classification model is established through an extended short text training set; (4) the short text to be classified is extended: the extension words are added to the short text to be classified according to a certain extension rule according to the two-tuples in the background knowledge base and the feature space of the classification model; and (5) a classification result is generated: the classification result is generated through the classification model and the extended short text. According to the Chinese short text classification method based on characteristic extension, the features of the short text are enriched through the long text corpus, so that the accuracy and the recall rate in the classification of the short text are improved.

Methods and apparatus are disclosed for efficient TLB (translation look-aside buffer) shoot-downs for heterogeneous devices sharing virtual memory in a multi-core system. Embodiments of an apparatus for efficient TLB shoot-downs may include a TLB to store virtual address translation entries, and a memory management unit, coupled with the TLB, to maintain PASID (process address space identifier) state entries corresponding to the virtual address translation entries. The PASID state entries may include an active reference state and a lazy-invalidation state. The memory management unit may perform atomic modification of PASID state entries responsive to receiving PASID state update requests from devices in the multi-core system and read the lazy-invalidation state of the PASID state entries. The memory management unit may send PASID state update responses to the devices to synchronize TLB entries prior to activation responsive to the respective lazy-invalidation state.

Various embodiments of the present invention provide for immediate allocation of virtual memory on behalf of processes running within a computer system. One or more bit flags within each translation indicate whether or not a corresponding virtual memory page is immediate. READ access to immediate virtual memory is satisfied by hardware-supplied or software-supplied values. WRITE access to immediate virtual memory raises an exception to allow an operating system to allocate physical memory for storing values written to the immediate virtual memory by the WRITE access.

A computer system using virtual memory provides hybrid memory access either through a conventional translation between virtual memory and physical memory using a page table possibly with a translation lookaside buffer, or a high-speed translation using a fixed offset value between virtual memory and physical memory. Selection between these modes of access may be encoded into the address space of virtual memory eliminating the need for a separate tagging operation of specific memory addresses.

A share memory service system and method of web service oriented applications, which includes the steps of declaring an object requiring to be shared of a plurality of web services as a shared object, storing the shared object to a shared virtual memory, and providing the shared object of the shared virtual memory for the access of the web services and a plurality of clients. Accordingly, a shared virtual memory may be simulated in any web service oriented applications, and the method of managing the shared virtual memory providing a way for web services of application program to access the shared object directly instead of invoking each other through the external call. Hence the complexity of service oriented applications development may be reduced.

An efficient memory management method for handling large data volumes, comprising a memory management interface between a plurality of applications and a physical memory, determining a priority list of buffers accessed by the plurality of applications, providing efficient disk paging based on the priority list, ensuring sufficient physical memory is available, sharing managed data buffers among a plurality of applications, mapping and unmapping data buffers in virtual memory efficiently to overcome the limits of virtual address space.

An application programming interface (API) enables application programs in a multitasking operating environment to classify portions of their code and data in a group that the operating system loads into physical memory all at one time. Designed for operating systems that implement virtual memory, this API enables memory-intensive application programs to avoid performance degradation due to swapping of units of memory back and forth between the hard drive and physical memory. Instead of incurring the latency of a page fault whenever the application attempts to access code or data in the group that is not located in physical memory, the API makes sure that all of the code or data in a group is loaded into physical memory at one time. This increases the latency of the initial load operation, but reduces performance degradation for subsequent memory accesses to code or data in the group.

In one embodiment, a memory allocator of a memory manager can service memory allocation requests within a specific size-range from a section of pre-reserved virtual memory. The pre-reserved virtual memory allows allocation requests within a specific size range to be allocated in the pre-reserved region, such that the virtual memory address of a memory allocation serviced from the pre-reserved region can indicate elements of metadata associated with the allocations that would otherwise contribute to overhead for the allocation.

The aspects include systems and methods of managing virtual memory page shareability. A processor or memory management unit may set in a page table an indication that a virtual memory page is not shareable with an outer domain processor. The processor or memory management unit may monitor for when the outer domain processor attempts or has attempted to access the virtual memory page. In response to the outer domain processor attempting to access the virtual memory page, the processor may perform a virtual memory page operation on the virtual memory page.

The disclosure provides a virtual machine memory access method. The method comprises the following steps: acquiring a virtual machine control structure corresponding to a virtual machine, wherein the virtual machine control structure comprises a virtual memory starting address and a virtual memory size; establishing mapping relationships with a physical memory address and a virtual memory address corresponding to the virtual machine according to the virtual memory starting address and the virtual memory size; and accessing the corresponding physical memory address through a virtual memory space address based on the mapping relationships. The disclosure also provides electronic equipment in which the virtual machine is deployed, and a virtual machine memory access system.

A method, system, and computer program product provide a shared virtual memory space via a cluster-aware virtual input / output (I / O) server (VIOS). The VIOS receives a paging file request from a first LPAR and thin-provisions a logical unit (LU) within the virtual memory space as a shared paging file of the same storage amount as the minimum required capacity. The VIOS also autonomously maintains a logical redundancy LU (redundant LU) as a real-time copy of the provisioned / allocated LU, where the redundant LU is a dynamic copy of the allocated LU that is autonomously updated responsive to any changes within the allocated LU. Responsive to a second VIOS attempting to read a LU currently utilized by a first VIOS, the read request is autonomously redirected to the logical redundancy LU. The redundant LU can be utilized to facilitate migration of a client LPAR to a different computing electronic complex (CEC).

Various embodiments of the present invention provide for immediate allocation of virtual memory on behalf of processes running within a computer system. One or more bit flags within each translation indicate whether or not a corresponding virtual memory page is immediate. READ access to immediate virtual memory is satisfied by hardware-supplied or software-supplied values. WRITE access to immediate virtual memory raises an exception to allow an operating system to allocate physical memory for storing values written to the immediate virtual memory by the WRITE access.

Disclosed herein is a virtual memory system including a nonvolatile memory allowing random access, having an upper limit to a number of times of rewriting, and including a physical address space accessed via a virtual address; and a virtual memory control section configured to manage the physical address space of the nonvolatile memory in page units, map the physical address space and a virtual address space, and convert an accessed virtual address into a physical address; wherein the virtual memory control section is configured to expand a physical memory capacity allocated to a virtual page in which rewriting occurs.

A shared virtual memory management apparatus for ensuring cache coherence. When two or more cores request write permission to the same virtual memory page, the shared virtual memory management apparatus allocates a physical memory page for the cores to change data in the allocated physical memory page. Thereafter, changed data is updated in an original physical memory page, and accordingly it is feasible to achieve data coherence in a multi-core hardware environment that does not provide cache coherence.

An application programming interface (API) enables application programs in a multitasking operating environment to classify portions of their code and data in a group that the operating system loads into physical memory all at one time. Designed for operating systems that implement virtual memory, this API enables memory-intensive application programs to avoid performance degradation due to swapping of units of memory back and forth between the hard drive and physical memory. Instead of incurring the latency of a page fault whenever the application attempts to access code or data in the group that is not located in physical memory, the API makes sure that all of the code or data in a group is loaded into physical memory at one time. This increases the latency of the initial load operation, but reduces performance degradation for subsequent memory accesses to code or data in the group.

A long distance paging method and network interface card for virtual share storage system, which contains mapping the memory area in network interface card to virtual address space of user process, source node user process directly accessing network interface card, generating and sending long distance paging request message to network interface card of destination node which directly accessing the needed page and feeding back response message to source node network interface card which directly writing said page to memory area of source node, said network interface card is added with frame head analysis logic, RDMA operation logic and virtual / real address conversion, said invention can realize long distance paging operation without breaking off the current work of CPU by bidirection operation of user process and network interface card.

Disclosed are a method and a relevant device for determining a mode of managing a shared virtual memory page. The method for determining a mode of managing a shared virtual memory page comprises: monitoring an access characteristic of a process to the shared virtual memory page; and if the monitored access characteristic conforms to the first set condition and a current mode of managing the shared virtual memory page is a distributed shared memory mode, modifying the mode of managing the shared virtual memory page to a shared physical memory mode. The technical solutions provided by the embodiments of the present invention facilitate the improvement of access performance of the shared virtual memory.

A multiprocessor using a shared virtual memory (SVM) is provided. The multiprocessor includes a plurality of processing cores and a memory manager configured to transform a virtual address into a physical address to allow a processing core to access a memory region corresponding to the physical address.

According to one embodiment, a data storage control method, which is applied to a virtual memory that controls access to the data stored in each of the physical memory regions by the corresponding one of the virtual addresses on the basis of an address management table that manages the correspondence relationship between a plurality of virtual addresses corresponding to a plurality of virtual memory regions and a plurality of physical addresses corresponding to a plurality of physical memory regions of a first memory, includes writing the data stored in a specific number of nonconsecutive physical memory regions made to correspond to a specific number of virtual memory regions on the basis of the address management table to a specific number of consecutive physical memory regions.

Methods and apparatus for arbitration and access to hardware request ring structures in a concurrent environment. A request ring mechanism is provided including an arbiter, ring overflow guard, request ring, and request ring metadata, each of which is implemented in shared virtual memory (SVM) on a computing platform including a multi-core processor coupled to an offload device having one or more SVM-capable accelerators. Worker threads request to access the request ring to provide job descriptors to be processed by the accelerator(s). A lockless arbiter returns either an index of a slot in which to write a descriptor or information indicating the ring is full to each worker thread. The scheme enables worker threads to write descriptors to slots in the request ring corresponding to the returned indexes without contention from other worker threads. The ring overflow guard prevents valid descriptors from being overwritten before they are taken off the ring by the accelerator(s). The request ring metadata is used indicate a valid / invalid status of the ring entries.

The invention discloses a polarized SAR (Synthetic Aperture Radar) image classifying method based on cooperative representation and deep learning, and mainly solves the problems that an existing method is high in computation complexity and low in classification precision. The method comprises the realizing steps: 1, inputting a polarized SAR image, and extracting the polarization characteristics of the image; 2, selecting a training sample set according to practical ground features, and selecting pixel points of the entire image as a test sample set; 3, taking the characteristics of the training sample set as an initial dictionary, and learning the initial dictionary to obtain a learning dictionary by K-SVD (Singular Value Decomposition); 4, synergically representing the training sample set and the testing sample set to obtain the representation coefficients of the training sample set and the testing sample set by the learning dictionary; 5, deeply learning the representation coefficients of the training sample set and the testing sample set so as to obtain more essential characteristic representing; and 6, carrying out the polarized SAR image classification on the representation coefficients by an libSVM (Shared Virtual Memory) classifier after the deep learning. The SAR image classifying method provided by the utility model is low in computation complexity and high in classification accuracy, and is applicable to the polarized SAR image classification.

An access request for a data processing system having no system memory is disclosed. The data processing system includes multiple processing units. The processing units have volatile cache memories operating in a virtual address space that is greater than a real address space. The processing units and the respective volatile memories are coupled to a storage controller operating in a physical address space that is equal to the virtual address space. The processing units and the storage controller are coupled to a hard disk via an interconnect. The storage controller, which is coupled to a physical memory cache, allows the mapping of a virtual address from one of the volatile cache memories to a physical disk address directed to a storage location within the hard disk without transitioning through a real address. The physical memory cache contains a subset of information within the hard disk. When a specific set of data is needed, a processing unit generates a virtual memory access request to be received by the storage controller. The storage controller then fetches the data for the requesting processor. The virtual memory access request includes a group of hint bits regarding data prefetch associated with the fetched data.

The invention provides a virtual memory dynamic allocation method of a virtual machine. The virtual memory dynamic allocation method of a virtual machine realizes the dynamic allocation of the virtual memory by developing a control program with the memory balloon technology. Firstly, setting two memory adjustment threshold values T1 (if the memory utilization rate is larger than T1, the memory is extended) and T2 (if the memory utilization rate is less than T2, the memory is recycled) through the control program; then, collecting the memory utilization rate U1 of each virtual machine and storing in the database. In the second collection cycle, the collected virtual machine memory utilization rate U2 is also stored in the database, and the virtual memory analysis is performed at the same time. If (U1+U2) / 2>T1, the virtual memory is extended dynamically; on the contrary, if (U1+U2) / 2<T2 and U1<T1 and U2<T1, the virtual memory is recycled dynamically. According to the virtual memory dynamic allocation method of a virtual machine, the values of T1 and T2 are set according to the actual application, therefore, the virtual machine memory can be dynamically allocated effectively, and the memory utilization efficiency is improved.