61 results about "Runtime library" patented technology

The invention provides a method and system for achieving CUDA call tracking based on an eBPF, and relates to the field of GPU call tracking of the eBPF. The method comprises the following steps: dynamically mounting an eBPF program at an API function entry and an API function exit of a user mode CUDA runtime library through an uprobe technology of the eBPF; the eBPF program captures metadata of the CUDA API calling event in a kernel mode and packages the metadata into a predefined structural body; asynchronously transmitting the structural body data to a user mode through an annular buffer area; and calling a standard function library by a user mode program to load an eBPF program, reading data in the annular buffer area in real time, and completing analysis and visual display of a CUDA calling event. According to the method, the calling parameter, the return value and the nanosecond timestamp are effectively captured, and the observability and the debugging efficiency of the GPU application program are improved.

The invention provides a heterogeneous GPU resource management scheduling method, and relates to the technical field of GPU resource allocation, heterogeneous equipment management and unified abstract modeling are carried out, GPU resources of different architectures are registered to a container arrangement platform, and a unified abstract layer is constructed to shield bottom layer hardware differences; gPU cluster optimization management based on a multi-dimensional real-time monitoring and intelligent scheduling strategy is carried out, GPU operation indexes are collected, priorities are dynamically calibrated for tasks, and task performance portraits are constructed; scheduling decision making is carried out through multi-strategy cooperation, and optimal GPU resources are distributed for tasks; carrying out fine-grained resource allocation, carrying out space or time segmentation on the GPU, and dynamically adjusting resource allocation according to a load state; aPI conversion of cross-architecture tasks is realized through a unified runtime library, and task execution data is collected to feed back an optimization scheduling model; automatic detection, isolation and task migration of GPU faults are carried out, and unified monitoring and alarm are provided.

The invention provides a runtime virtual machine system. A specific embodiment of the system comprises a user mode application program; the asynchronous IO model is used for enabling the user mode application program and the kernel mode to share an IO request memory queue; a dynamic runtime library can be loaded when a user mode application program is started, and after the dynamic runtime library is loaded, a runtime virtual environment for replacing a runtime environment of the user mode application program is generated; the dynamic runtime library comprises an IO operation library function hook; the IO operation library function hook is constructed on the basis of a native library function of the user mode application program, and the IO operation library function is used for adapting to the application program of the native library function and the asynchronous IO model in the runtime virtual environment and converting IO operation of an operation system where the user mode application program is located into asynchronous IO model processing. According to the implementation mode, the transparent running environment of the application program can be replaced with the pure user mode environment, and IO operation of the pure user mode is achieved.

The invention discloses a credential application compatibility adaptation method and system based on instruction set conversion, and relates to the technical field of computer compatibility adaptation. According to the method, related fragments of the instruction set in the source code are accurately recognized through a static analysis tool, and the fragments are converted into the target instruction set through an adapter mode and a dynamic compiling technology; meanwhile, a binary analysis tool is adopted to adapt to the runtime library and optimize the performance, and the compatibility and stability of the application are comprehensively verified through function testing, performance testing and safety testing, so that the whole scheme not only solves the compatibility problem in cross-platform migration, but also optimizes the performance and guarantees the safety; and the adaptation efficiency and the maintainability of the creative application are remarkably improved.

One embodiment provides a method comprising, at a runtime library executed by a processor of a data processing system, receiving an input frame having objects to be stylized via a style transfer network associated with the runtime library, wherein the style transfer network is a neural network model trained to apply one or more visual styles to an input frame, performing instance segmentation on the input frame to generate one or more instance masks to identify one or more objects to be stylized, generating one or more stylized frames for each style to transfer to the input frame, and merging, via the one or more instance masks, stylized objects from one or more stylized frames with un-stylized content from the input frame to generate an output frame with per-instance stylization.

The invention provides a task scheduling method and device, a storage medium and electronic device.The method is applied to a runtime library of a host and comprises the steps that in response to a received calculation task, a kernel driver is called to obtain performance data of at least two core particles, and a scheduling score corresponding to each core particle is determined according to the performance data; scheduling the calculation task to the corresponding core particle for execution based on the scheduling score; the scheduling score is determined on the basis of the performance data of the core particles, task scheduling is executed, accurate adaptation of the calculation task and the core particles is achieved, and the task execution efficiency and the resource utilization rate under the multi-core-particle architecture are effectively improved.

The present disclosure provides a runtime virtualization method and apparatus. A specific implementation of the method includes: constructing a dynamic runtime library based on the native runtime library of an application, where the dynamic runtime library simulates threads implemented in the native runtime library in a coroutine manner; loading the dynamic runtime library when starting the application to generate a runtime virtualization environment and a coroutine scheduling thread running in the runtime virtualization environment, where the coroutine scheduling thread is used to schedule coroutines, and the runtime virtualization environment is an environment after replacing the native runtime library with a generated runtime environment; receiving a task request of the application in the runtime virtualization environment; and scheduling coroutines using the coroutine scheduling thread based on the task request and the dynamic runtime library so that the coroutines complete the tasks corresponding to the task request. This implementation improves the concurrency performance of the application, reduces resource consumption and resource scheduling pressure without making any modifications or recompiling to the application in the target operating system.

The invention discloses a task scheduling method and system for an application system, a storage medium and electronic equipment, and the method comprises the steps: recognizing the message information of a task description file, and obtaining a configuration file and a task; starting the application process, and writing the configuration file; initializing a framework library, a dynamic runtime library and a component application program; creating a task execution framework based on the configuration file and the framework library; loading a framework library and a component application program based on the task; the dynamic runtime library sets task execution time and executes tasks, the trigger module detects whether trigger information is received or not, if yes, a trigger strategy is determined, and plug-ins in the component application program are called; the process tasks are uniformly managed and controlled by the task management module and are placed in the thread pool to be processed, the problem that the application randomly opens the thread to execute the tasks is avoided, the report state of the tasks can be monitored in real time, the multiple tasks are scheduled to the dynamic runtime library during execution, and the size of a code file is reduced while the influence range of modification is reduced.

The invention relates to a hybrid accelerator card management method and device, an electronic device and a storage medium, and is applied to the field of computer hardware acceleration and data storage and transmission, and the hybrid accelerator card management method comprises the following steps: determining a runtime library interface of a hybrid accelerator card in response to a received server data transmission request; registering a memory address of the hybrid acceleration card to a preset hardware abstraction layer based on a remote direct memory access technology; a unified hardware abstraction interface is preset in the hardware abstraction layer; the runtime library interface is bound with the unified hardware abstract interface; and calling the runtime library interface to execute data transmission operation between the hybrid acceleration cards through a preset hardware abstraction layer. According to the method and the device, the GPU acceleration cards of different manufacturers are compatible, and meanwhile, the effect of improving the data processing and transmission efficiency is achieved.

The invention provides a database thread pool control method and device. A specific embodiment of the method comprises the steps of loading a dynamic runtime library capable of replacing a thread with a coroutine in an operating system of a database, and generating at least one coroutine scheduling thread; receiving a connection request sent by the client to the database; detecting whether the request of the database for creating the new thread is a request for processing the connection request; and in response to detecting that the request for creating the new thread of the database is a request for processing the connection request, creating a coroutine corresponding to the connection request, and scheduling the coroutine through at least one coroutine scheduling thread to complete an access task corresponding to the connection request. According to the implementation mode, the coroutine scheduling thread serves as a working thread of a database thread pool, the problem of task starvation of a traditional multi-thread scheduling model and the problem of related task response speed jitter are avoided through a coroutine swap-in and swap-out mechanism, the concurrency performance of a database is improved, and meanwhile resource consumption of concurrency scheduling is reduced.

The present disclosure provides a database thread pool control method and apparatus. A specific implementation of the method includes: loading a dynamic runtime library capable of replacing threads with coroutines in the operating system of the database to generate at least one coroutine scheduling thread; receiving a connection request sent by a client to the database; detecting whether a request for the database to create a new thread is a request for processing the connection request; in response to detecting that the request for the database to create a new thread is a request for processing the connection request, creating a coroutine corresponding to the connection request, and scheduling the coroutine through at least one coroutine scheduling thread to complete the access task corresponding to the connection request. In this implementation, the coroutine scheduling thread serves as the working thread of the database thread pool, and the task "starvation" problem of the traditional multithreaded scheduling model and its related task response speed jitter problem are avoided by using the coroutine swap-in and swap-out mechanism. While improving the database concurrency performance, the resource consumption of concurrent scheduling is reduced.

The invention discloses a Java application lightweight deployment system and method based on dependency stability grading, and relates to the technical field of software deployment, and the method comprises the following steps: collecting dependency metadata, and generating a dependency list description file; packaging the Java application service code and the internal runtime library into a service code file, forming a service code ciphertext file, and generating a lightweight application deployment package; constructing a basic container mirror image, pulling the basic container mirror image and a lightweight application deployment package, decrypting a service code ciphertext file, downloading missing dependency and verifying a dependency fingerprint, and starting a SpringBoot application; and collecting dependency operation data, generating and issuing an incremental updating plan, executing the incremental updating plan, and updating the dependency stability score and the dependency stability level. And the stability of the deployment carrier and the operation reliability of the Java application are improved.

The invention discloses a cross-platform model deployment framework, electronic equipment and a storage medium, and the framework is applied to the electronic equipment and specifically comprises a framework interface abstraction layer and a plurality of runtime libraries connected with the framework interface abstraction layer. The framework interface abstraction layer is used for being connected with an application framework layer and providing a uniform abstraction interface for the application framework layer; the runtime library is used for providing running support for the framework interface abstraction layer. Compared with a traditional model deployment scheme, the scheme provided by the invention can accelerate the deployment speed of the model, ensures the uniformity of evaluation standards of the model on different platforms, and improves the reasoning performance. If the running environment is inferred to be updated, deployment updating of the model on a new platform can be rapidly completed according to the framework, and manpower and material resources needed by deployment can be greatly saved.

This invention provides a method and apparatus for acquiring memory access characteristic data of a graphics processing unit (GPU), relating to the field of computer technology. The method includes: intercepting call parameters of application programming interfaces (APIs) in the GPU runtime library; constructing a memory region tree structure based on the call parameters; inserting a preset function into the execution instruction stream of the GPU kernel to acquire instruction-level execution data; and generating GPU memory access characteristic data based on the memory region tree structure and the instruction-level execution data. The GPU memory access characteristic data acquisition method and apparatus provided by this invention correlate macroscopic memory and communication semantics with microscopic instruction-level execution data to generate global, fine-grained memory access characteristic data, improving the comprehensiveness and accuracy of acquiring GPU memory access characteristic data in cloud-edge-device architectures.

The application discloses a PE virus detection method, comprising the following steps: determining target resources required by a PE file during runtime; wherein the target resources comprise a CPU, a memory management module and a runtime library; virtually executing the PE file by simulating the target resources, and recording characteristic information generated by the PE file during the virtual execution; and judging whether the PE file is a virus file according to the characteristic information. The application can reduce resource consumption of PE virus detection. The application also discloses a PE virus detection device, a storage medium and an electronic device, which have the above beneficial effects.

A runtime virtual machine system. The system comprises: a user-mode application program (101); an asynchronous IO model (102), which is used for sharing an IO request memory queue between the user-mode application program (101) and a kernel mode; and a dynamic runtime library (103), which can be loaded when the user-mode application program (101) is started, wherein the dynamic runtime library (103) is used for generating, after being loaded, a runtime virtual environment for replacing a runtime environment of the user-mode application program (101), the dynamic runtime library (103) comprises an IO operation library function hook (1031) and a monitoring thread function (1032), the IO operation library function hook (1031) is constructed on the basis of a native library function of the user-mode application program (101), and the IO operation library function is used for adapting an application program interface of the native library function to the asynchronous IO model (102) in the runtime virtual environment, such that an IO operation of an operating system where the user-mode application program (101) is located is transformed to be processed by the asynchronous IO model (102).

Systems and methods for developing one or more applications associated with a browser-based user interface within a multi-developer computing platform employ one or more processors that receive a request to build one or more applications configured to run in a browser-based user interface on a client; determine whether the one or more applications are associated with a core library and one or more runtime libraries; and in response to determining that the one or more applications are associated with the core library and the one or more runtime libraries: determine one or more version numbers associated with the one or more runtime libraries; and in response to the determined one or more version numbers being within a predetermined range associated with the core library, dynamically link the one or more runtime libraries and the one or more applications.

The invention discloses a skeletal animation rendering method and device, electronic equipment and a storage medium, and relates to the technical field of graphic processing. The method is applied to the electronic equipment, an operating system is installed in the electronic equipment, and the method comprises the steps that an application page of an application program is displayed, a runtime library with skeletal animations is imported into the application program, the runtime library is called through GLSkeletal View of the operating system to create Skeletal Drawable, the Skeletal Drawable is called through the GLSkeletal View to set animation parameters, an animation instance object is obtained, and the animation instance object is displayed on the application page. And calling the SkeleteronDrawable and the runtime library through the GLSkeleteronView, carrying out animation drawing according to the animation instance object to render a skeleton animation picture, and playing the skeleton animation page in the application page. According to the method and the device, the runtime library of the skeletal animation is imported into the application program, and the GLSkeleteronView of the operating system is called to realize rendering playing of the skeletal animation, so that a game engine does not need to be additionally introduced, and the access cost is low.

Embodiments of the present application provide a file storage method, a file access method, an electronic device and a storage medium, and relate to the field of information technology. The method comprises: obtaining an installation package file; decompressing a runtime library compressed file in the installation package file to obtain a runtime library file corresponding to the runtime library compressed file; recording a correspondence relationship and a first offset, and releasing one or more data blocks of the runtime library compressed file in the installation package file; wherein the correspondence relationship is used to represent the correspondence relationship between the one or more data blocks of the runtime library compressed file in the installation package file and the runtime library file, and the first offset is used to represent the offset of the one or more data blocks of the runtime library compressed file in the installation package file. The method provided by the embodiments of the present application can save the space occupied by the installation package file, thereby improving the storage efficiency.

The invention discloses a method and device for operating an SPEC CPU in a simulator user mode. The method comprises the steps that a system operation environment is configured, and an SPEC CPU test suite, a cross compiling tool chain and a simulator are installed; compiling the SPEC CPU test suite through the cross compiling tool to obtain an executable file of a given instruction set architecture; obtaining an execution command and a verification command of the executable file, and summarizing to generate an execution script of the simulator in a user mode; and running the execution script. According to the method, the SPEC CPU can be completely operated in a simulator user mode, so that function verification and performance evaluation can still be performed on an instruction set, a compiler, a runtime library and the like in the absence of hardware, the verification and evaluation process is simplified, and meanwhile, the occupation of software resources is reduced.

The invention discloses an application system security protection method and system based on an execution virtual machine, and the method comprises the steps: reading a CLASS file and related metadata in a JAR package, and converting the code of the CLASS file into a pseudo code of an EVM line by line; encrypting the generated pseudo code by adopting a symmetric encryption algorithm, packaging the encrypted pseudo code and related metadata, and generating a program package protected by EVM (Error Vector Magnitude); dependency on EVM is added into the converted pseudo-code, and meanwhile, the pseudo-code is integrated with an EVM runtime environment; the JRE can load a runtime library of the EVM and start the EVM, and the EVM initializes an internal data structure; when the code is executed to be required to call a function, access a variable or a constant, searching a corresponding element from the HashMap of the initialized internal data structure by the EVM; judging whether the EVM searches a corresponding element in the HashMap or not; cache acquisition is carried out, and the data is directly delivered to a just-in-time compiler for final local code generation; and performing pre-compilation release on the EVM, and returning to the step S6 again.

The invention discloses a space target part segmentation development implementation method based on edge end equipment, and the method comprises the following steps: 1, constructing a space target radar image data set, and training a constructed network model; step 2, configuring an MLU server development environment; 3, calling a quantification interface, and converting the weight of the deep learning network trained in the step 1 into a weight with int8 precision; 4, designing online reasoning software, operating a layer-by-layer reasoning mode by using MLU270 equipment, and verifying whether the weight of the int8 quantization precision meets the precision requirement or not; step 5, running the fusion reasoning mode by using MLU270 equipment, and generating an offline model of the MLU270 version and an offline model of the MLU220 version; 6, off-line reasoning software design is carried out through a CNRT runtime library, and an MLU270 version off-line model is loaded to test an off-line reasoning result; and step 7, deploying an offline reasoning program in the intelligent micro server, loading an MLU220 version offline model, and testing an offline reasoning result under the edge end equipment. According to the invention, high-efficiency real-time detection and segmentation can be carried out on the components of the inverse synthetic aperture radar image of the space target on the embedded edge end equipment, and the method has the remarkable advantages of high recognition rate, high processing speed, strong real-time performance and the like.

A method and apparatus for offloading compute-intensive workloads is provided. A database system compiles an execution plan to generate an offload-enabled plan by identifying a candidate offloading region in the execution plan, generating and adding an offloading branch in the offload-enabled plan, corresponding to the candidate offloading region, for execution by a compute offload runtime, wherein the compute offload runtime comprises a compute offload runtime library executing on the database system and on each node of a compute offload server, and adding the candidate offloading region as a fallback branch in the offload-enabled plan. The database system executes the offload-enabled plan by executing the offloading branch using one or more compute nodes in the database server or the compute offload server using the offload runtime or by executing the fallback branch using one or more compute nodes in the database server.

This application discloses a program call stack creation method and apparatus, and an unwind stack method and apparatus. The call stack creation method includes: obtaining a first program, where the first program includes a first function, in a process of calling a runtime library to run the first program, adding a second function and creating a call stack, where the call stack includes a first stack frame, a second stack frame, and a third stack frame that are respectively corresponding to the first function, the second function, and a third function. An unwind stack performance can be improved to a maximum extent without affecting a non-stack trace scenario.

The present application relates to a hybrid acceleration card management method, apparatus, electronic device, and storage medium, which are applied to the fields of computer hardware acceleration, data storage, and transmission. Among them, the hybrid acceleration card management method includes: in response to a received server data transmission request, determining a runtime library interface of the hybrid acceleration card; based on the Remote Direct Memory Access (RDMA) technology, registering the memory address of the hybrid acceleration card to a preset hardware abstraction layer; a unified hardware abstraction interface is preset in the hardware abstraction layer; the runtime library interface is bound to the unified hardware abstraction interface; through the preset hardware abstraction layer, the runtime library interface is called to execute data transmission operations between hybrid acceleration cards. Through the present application, while being compatible with GPU acceleration cards of different manufacturers, the effect of improving data processing and transmission efficiency is achieved.

This invention proposes a heterogeneous GPU resource management and scheduling method, relating to the field of GPU resource allocation technology. It addresses heterogeneous device management and unified abstract modeling, registering GPU resources of different architectures to a container orchestration platform and constructing a unified abstraction layer to shield underlying hardware differences. The method optimizes GPU cluster management based on multi-dimensional real-time monitoring and intelligent scheduling strategies, collecting GPU performance metrics, dynamically prioritizing tasks, and constructing task performance profiles. It makes scheduling decisions through multi-strategy collaboration to allocate optimal GPU resources to tasks. Fine-grained resource allocation is performed, partitioning GPUs spatially or temporally and dynamically adjusting resource allocation based on load status. A unified runtime library enables cross-architecture task API conversion and collects task execution data to optimize the scheduling model. Finally, it performs automatic GPU fault detection, isolation, and task migration, providing unified monitoring and alerting.

The present disclosure provides a database thread pool control method and apparatus. A specific implementation of the method comprises: loading, in an operating system of a database, a dynamic runtime library allowing for replacement of a thread with a coroutine, and generating at least one coroutine scheduling thread; receiving a connection request sent by a client to the database; detecting whether a request for creating a new thread of the database is a request for processing the connection request; and in response to detecting that the request for creating the new thread of the database is the request for processing the connection request, creating a coroutine corresponding to the connection request, and scheduling the coroutine by means of the at least one coroutine scheduling thread to complete an access task corresponding to the connection request.

According to one embodiment, the system receives a request from an application to call a predetermined function at a runtime library executed in a trusted execution environment (TEE) of a host system to perform a predetermined operation. In response to the request, the system identifies a kernel object associated with the predetermined function. The system uses a public key corresponding to a private key to confirm an executable image of the kernel object, wherein the private key is used to sign the executable image of the kernel object. In response to successfully confirming the executable image of the kernel object, the system transmits the executable image confirmed by the kernel object to a data processing (DP) accelerator via a bus for execution by the DP accelerator to perform the predetermined operation.

The present application discloses a system log query method and system in EVTX format based on byte stream, which mainly relates to the field of log query technology, and is used to solve the problem that when operating log files, the interface provided by Microsoft on the Windows platform does not support additional files and cross-Linux platforms. In addition, the use of third-party tools to operate log files requires secondary development based on the third-party tools, and the process is relatively cumbersome. It includes: determining the open file address or disk device address based on the data source; obtaining the platform type corresponding to the parsed data, and registering the runtime library corresponding to the system type; opening the parsed data, sequentially reading the byte stream corresponding to the parsed data, determining that the byte stream with the identification header of the evtx file is successfully matched, calling the evtx file storage format parsing program, and parsing and obtaining several log files from the successfully matched byte stream; and returning several log files to the user end through the callback function.

The invention belongs to the technical field of vehicle software development, and particularly relates to a user interface dynamic extensible development method and system based on a software platform. The interface manager dynamically extracts type information related to a window in a software platform runtime library and inherits the type information to form a dynamically derived basic window, meanwhile, the interface manager generates software platform import codes, codes for dynamically creating a primary window are generated according to the type information in the software platform import codes, and the primary window is dynamically created according to the type information in the software platform import codes. And storing software platform import codes and codes for primary window dynamic creation into a primary derived code file.