Embodiments of the present disclosure include techniques for processing neural networks. Various forms of parallelism may be implemented using topology that combines sequences of processors. In one embodiment, the present disclosure includes a computer system comprising a plurality of processor groups, the processor groups each comprising a plurality of processors. A plurality of network switches are coupled to subsets of the plurality of processor groups. A subset of the processors in the processor groups may be configurable to form sequences, and the network switches are configurable to form at least one sequence across one or more of the plurality of processor groups to perform neural network computations. Various alternative configurations for creating Hamiltonian cycles are disclosed to support data parallelism, pipeline parallelism, layer parallelism, or combinations thereof.

Embodiments of the present disclosure include techniques for processing neural networks. Various forms of parallelism may be implemented using topology that combines sequences of processors. In one embodiment, the present disclosure includes a computer system comprising a plurality of processor groups, the processor groups each comprising a plurality of processors. A plurality of network switches are coupled to subsets of the plurality of processor groups. A subset of the processors in the processor groups may be configurable to form sequences, and the network switches are configurable to form at least one sequence across one or more of the plurality of processor groups to perform neural network computations. Various alternative configurations for creating Hamiltonian cycles are disclosed to support data parallelism, pipeline parallelism, layer parallelism, or combinations thereof.

Embodiments of the present application provide an acceleration management node. The acceleration management node separately receives acceleration device information of all acceleration devices. The acceleration device information includes an algorithm type, an acceleration bandwidth or non-uniform memory access architecture (NUMA). The acceleration management node obtains an invocation request from a client. The acceleration management node queries the acceleration device information to determine, from all the acceleration devices of the at least one acceleration node, a target acceleration device matching the invocation request. The acceleration management node further instructs a target acceleration node to respond to the invocation request.

Embodiments of the present application provide an acceleration management node. The acceleration management node separately receives acceleration device information of all acceleration devices. The acceleration device information includes an algorithm type, an acceleration bandwidth or non-uniform memory access architecture (NUMA). The acceleration management node obtains an invocation request from a client. The acceleration management node queries the acceleration device information to determine, from all the acceleration devices of the at least one acceleration node, a target acceleration device matching the invocation request. The acceleration management node further instructs a target acceleration node to respond to the invocation request.

A networking system may include a switch coupled to a computing resource. A resource management system may control the computing resource. A controller may be coupled to the switch. The controller may include a resource management interface that is coupled to the resource management system via a communications link. The resource management interface may receive computing resource information for the computing resource via the communications link. The controller may provide control data to the switch to update a cloud network for the computing resource based on the received computing resource information.

A networking system may include a switch coupled to a computing resource. A resource management system may control the computing resource. A controller may be coupled to the switch. The controller may include a resource management interface that is coupled to the resource management system via a communications link. The resource management interface may receive computing resource information for the computing resource via the communications link. The controller may provide control data to the switch to update a cloud network for the computing resource based on the received computing resource information.

The invention enables high-availability, high-scale, high security and disaster recovery for API computing, including in terms of capture of data traffic passing through proxies, routing communications between clients and servers, and load balancing and/or forwarding functions. The invention inter alia provides (i) a scalable cluster of proxies configured to route communications between clients and servers, without any single point of failure, (ii) proxy nodes configured for implementing the scalable cluster (iii) efficient methods of configuring the proxy cluster, (iv) natural resiliency of clusters and/or proxy nodes within a cluster, (v) methods for scaling of clusters, (vi) configurability of clusters to span multiple servers, multiple racks and multiple datacenters, thereby ensuring high availability and disaster recovery (vii) switching between proxies or between servers without loss of session.

The invention enables high-availability, high-scale, high security and disaster recovery for API computing, including in terms of capture of data traffic passing through proxies, routing communications between clients and servers, and load balancing and/or forwarding functions. The invention inter alia provides (i) a scalable cluster of proxies configured to route communications between clients and servers, without any single point of failure, (ii) proxy nodes configured for implementing the scalable cluster (iii) efficient methods of configuring the proxy cluster, (iv) natural resiliency of clusters and/or proxy nodes within a cluster, (v) methods for scaling of clusters, (vi) configurability of clusters to span multiple servers, multiple racks and multiple datacenters, thereby ensuring high availability and disaster recovery (vii) switching between proxies or between servers without loss of session.

A device may receive data that identifies applications utilized by users, databases utilized by the applications, and the users, and may process the received data, with first models, to determine context data that matches the users and events associated with the applications, and task data that identifies tasks to be performed by the users in response to the events. The device may process the received data and the context data, with a second model, to generate role data that identifies user interfaces utilized by the users to access the applications, and credentials of the users, and may process the context data, the task data, and the role data, with a third model, to generate persona data that identifies personas, and assignment data that assigns each of the users to one of the personas. The device may perform actions based on the persona data and the assignment data.

An image forming apparatus includes a communication interface and a processor. The communication interface is configured to transmit data to and receive data from a cloud server that provides a cloud service. The processor is configured to receive a token from the cloud server via the communication interface, transmit a request including the token to the cloud server via the communication interface, receive a response including user information from the cloud server via the communication interface, and shift to a login state based on the user information.