A client obtains resource requirement information of an application APP running on a terminal device, where the resource requirement information is used to indicate a resource requirement of the APP; the client sends the resource requirement information to a cloud resource scheduling system; the client receives resource allocation information from the cloud resource scheduling system, where the resource allocation information includes information about at least one compute instance allocated by the cloud resource scheduling system based on the resource requirement information, and the at least one compute instance is deployed on at least one far-side site managed by the cloud resource scheduling system; and during running of the APP, the client sends some processing tasks of the APP to the at least one compute instance, and receives results of processing the some processing tasks of the APP by the at least one compute instance.

An electronic device includes a first graphics processing subsystem, a second graphics processing subsystem, and a screen. The first graphics processing subsystem includes a first application processor, a first graphics processing unit, and a first memory. The second graphics processing subsystem includes a second application processor, a second graphics processing unit, and a second memory. The first graphics processing unit renders a first GUI. The screen displays the first GUI. The second graphics processing unit renders a second GUI, and the second GUI and the first GUI belong to different interface types. The screen displays the second GUI. A display processing method applied to the electronic device is also provided, wherein the first graphics processing subsystem can be switched to the second graphics processing subsystem based on complexity of a to-be-displayed GUI.

Systems, methods, and devices are described coordinating a distributed accelerator. A command that includes instructions for performing a task is received. One or more sub-tasks of the task are determined to generate a set of sub-tasks. For each sub-task of the set of sub-tasks, an accelerator slice of a plurality of accelerator slices of a distributed accelerator is allocated, sub-task instructions for performing the sub-task are determined. Sub-task instructions are transmitted to the allocated accelerator slice for each sub-task. Each allocated accelerator slice is configured to generate a corresponding response indicative of the allocated accelerator slice having completed a respective sub-task. In a further example aspect, corresponding responses are received from each allocated accelerator slice and a coordinated response indicative of the corresponding responses is generated.

A system that stores functional information indicating a capability of each of a plurality of elements located remotely from the system; identifies a function capable of being performed by linking a plurality of the elements based on the stored functional information; and transmits information corresponding to the identified function capable of being performed by linking the plurality of elements to a first device remote from the system.

A method for data provisioning a serverless computing cluster. A plurality of user defined functions (UDFs) are received for execution on worker nodes of the serverless computing cluster. For a first UDF, one or more data locations of UDF data needed to execute the first UDF are determined. At a master node of the serverless computing cluster, a plurality of worker node tickets are received, each ticket indicating a resource availability of a corresponding worker node. The one or more data locations and the plurality of worker node tickets are analyzed to determine eligible worker nodes capable of executing the first UDF. The master node transmits a pre-fetch command to one or more of the eligible worker nodes, causing the eligible worker nodes to become a provisioned worker node for the first UDF by storing a pre-fetched first UDF data before the first UDF is assigned for execution.

A deep learning heterogeneous computing method based on layer-wide memory allocation, at least comprises steps of: traversing a neural network model so as to acquire a training operational sequence and a number of layers L thereof; calculating a memory room R.sub.1 required by data involved in operation at the i.sup.th layer of the neural network model under a double-buffer configuration, where 1≤i≤L; altering a layer structure of the i.sup.th layer and updating the training operational sequence; distributing all the data across a memory room of the CPU and the memory room of the GPU according to a data placement method; performing iterative computation at each said layer successively based on the training operational sequence so as to complete neural network training.

Techniques described herein relate to a method for allocating graphics processing unit partitions for a computer vision environment. The method includes obtaining, by a computer vision (CV) manager, an initial graphics processing unit (GPU) partition allocation request associated with a CV workload; in response to obtaining the initial GPU partition allocation request: obtaining CV workload information associated with the CV workload; obtaining first CV environment configuration information associated with the GPU partition allocation request; generating an optimal GPU partition allocation based on the first CV environment configuration information and the CV workload information using a GPU partition model; and initiating performance of the CV workload in a CV environment based on the optimal GPU partition allocation.

Safe critical section operations for virtual machines with virtual central processing unit overcommit are provided by: in response to identifying a preempting task to run on a first physical central processing unit (PCPU) from a second PCPU, setting a status of a flag in a virtual memory used by a first virtual central processing unit (VCPU) running on the first PCPU to indicate that the preempting task will interrupt the first VCPU; in response to initiating execution of a read-side critical section operation scheduled by the first VCPU to run on the first PCPU, checking the status of the flag in the virtual memory; and in response to the status of the flag being positive: exiting the first VCPU to a hypervisor; executing, by the hypervisor, the preempting task on the first PCPU; and after completing the preempting task, continuing execution of the read-side critical section operation.

Apparatuses, systems, and techniques to perform a fast Fourier transform operation. In at least one embodiment, a fast Fourier transform operation is performed based on one or more parameters, wherein the one or more parameters indicate information about one or more operands of the fast Fourier transform.

A task scheduler is connected between a central processing unit (CPU) and each hardware accelerator. The task scheduler first obtains a target task (for example, obtains the target task from a memory), and obtains a dependency relationship between the target task and an associated task. When it is determined, based on the dependency relationship, that a first associated task (for example, a prerequisite for executing the target task is that both a task 1 and a task 2 are executed) in the associated task has been executed, it indicates that the target task meets an execution condition, and the task scheduler schedules related hardware accelerators to execute the target task. Based on a dependency relationship between tasks, the task scheduler schedules, through hardware scheduling, each hardware accelerator to execute each task, and delivery of each task is performed through direct hardware access.