14 results about "Parallel routing" patented technology

The invention provides a routing control method, which is applied to network equipment with multi-link access capability, and comprises the following steps: carrying out service identification on a to-be-forwarded data flow of the network equipment to obtain service type information of the data flow; obtaining evaluation indexes of a plurality of candidate links of the network equipment; evaluating the plurality of candidate links based on the evaluation index and the service type information to obtain a comprehensive quality score; selecting a target link from the plurality of candidate links according to the service type information, the evaluation index and the comprehensive quality score; and controlling the network equipment to transmit the data flow through the target link. The invention further provides a multi-service parallel routing method, electronic equipment, a computer readable medium and a computer program product. According to the invention, differentiated routing control of multiple service types is realized, quality fluctuation caused by switching of operators is reduced, and the stability and the intelligent level of network transmission are remarkably improved.

Disclosed is a parallel routing method for two-stage partitions using genetic algorithms, which includes: partitioning and grouping elements constituting a circuit while maintaining a connectivity between the elements; calculating a sum of weights which a connection between any two partitioned groups has; deriving a grouping in which the calculated sum of the weights has a smallest value; generating boundary boxes for respective nets constituted by elements of the derived grouping; identifying a net having a dependency with respect to the respective nets; determining a work amount required for routing each identified net; generating a partition line; partitioning two partitioned partitions into a first partition and a second partition, and placing the other partition on any one partition; and routing the first partition and the second partition in parallel by different threads.

The application provides a workflow code automatic generation method driven by a large model, comprising: through multi-level agent cooperation, sequentially converting original requirements into a subtask set, a node relationship graph and structured workflow description data, wherein the last level agent forcibly checks the data closed loop between nodes; defining tool nodes and content generation nodes in the workflow description data to realize decoupling of content generation and tool calling; parsing the structured workflow description data, dynamically generating a strongly typed state class mapped with the description data, and automatically filling node calling logic and parallel routing rules based on a preset code template to generate workflow code; when code running errors and the number of retries exceeds a preset threshold, generating an execution track through an error positioning agent, generating candidate solutions in parallel by multiple repair agents, and after adaptive scoring by a scoring agent based on error types, verifying and executing the repair solutions in sequence.

This invention discloses a heterogeneous hybrid expert large model architecture and alignment method for high-concurrency online medical dialogue. First, the method proposes a dual-pool heterogeneous routing strategy of "4 general experts + 2 departmental experts." Through physically isolated parallel routing channels, it forces the model to simultaneously activate general language capabilities and specialized medical reasoning when processing input, fundamentally ensuring the focus and professionalism of responses. Second, it adopts a structural evolution strategy based on parameter expansion and selective freezing. By completely freezing the pre-trained basic parameters and training only newly added departmental experts, it completely eliminates catastrophic forgetting. Finally, in the model alignment stage, it employs the Group Relative Policy Optimization (GRPO) algorithm for valueless models to reduce memory overhead and designs a dual-track hybrid reward function combining low-level semantic similarity rewards and high-level structured large model referee scoring, effectively solving the "reward hacking" problem and guiding the model to generate refined responses that combine clinical accuracy and humanized interaction. This invention significantly improves the performance of medical dialogue systems. Its sparse activation and parameter freezing characteristics further ensure low latency and high throughput in online high-concurrency scenarios, providing a complete solution for the deployment of reliable medical dialogue systems.

A field-programmable gate array (FPGA) routing tool within a computer-aided design system. The tool includes an input device for receiving a netlist containing nets with source nodes, sink nodes, and intermediate nodes at fixed positions. The tool further includes a processing circuitry configured with a design router responsible for constructing non-overlapping routing trees for all nets, ensuring connections from source nodes to sink nodes without exceeding fixed routing resource capacity of the FPGA. The design router utilizes incremental routing, which applies deterministic parallel routing to a window of initial iterations with high routing workload and sequential routing to subsequent iterations. Additionally, a display device is provided to continuously exhibit interconnections and routing utilization during the determination of routing trees.

The invention relates to the technical field of model routing, and discloses a multi-model request-oriented routing method and device, equipment and a medium. The method comprises the steps of obtaining a user input request; judging whether the user input request carries a strong rule expression capable of covering a specific task scene or not; if yes, the user input request is routed to the task model of the corresponding task scene, and otherwise, the semantic similarity between the user input request and the task label of each task model is obtained; taking the task model with the highest corresponding semantic similarity as a main selection task model, and judging whether the semantic similarity corresponding to the main selection task model is greater than a preset threshold value or not; if yes, the user input request is routed to the main selection task model, otherwise, the user input request is routed in parallel to the candidate task models with the corresponding semantic similarity larger than the dynamic threshold value, and the processing result of the candidate task model which firstly processes the user input request serves as the target processing result of the user input request.

The invention discloses a supervision method and system supporting dynamic hot update and efficient matching of rules, and relates to the field of medical information, and the method comprises the steps: constructing a network during rule operation; when a rule updating instruction is received, executing a hot updating step: analyzing the rule updating instruction, and generating or modifying a corresponding target rule logic execution package; and in the atomic transaction, updating the routing states of all dynamic routes related to the target rule logic execution packet to be conducted, and updating the state of the target rule logic execution packet to be activated. When a to-be-processed event is received, executing an efficient matching step: extracting a feature key; querying the Hash reverse index table to obtain a first rule candidate set; performing logical operation on all the first rule candidate sets to obtain a second rule candidate set; routing the to-be-processed event to each rule logic execution packet in the second rule candidate set in parallel; and performing final matching verification. The millisecond-level continuous operation of the supervision service and the order of magnitude improvement of the system throughput are realized.

The application discloses a deep learning driven high-density standard cell global router, and relates to integrated circuit design automation technology. The router comprises a preprocessing module, a model inference module, a parallel routing scheduling module and a post-processing module. The preprocessing module is used for extracting relevant features of a layout from an intermediate result of the high-density standard cell global router and constructing a heterogeneous hypergraph. The model inference module is used for inputting the heterogeneous hypergraph into a multi-modal congestion prediction model for prediction. The parallel routing scheduling module is used for scheduling the first maze routing stage to perform early routing on the predicted overflow network. The post-processing module is used for screening and replacing the routing result of the first maze routing stage, and submitting the updated three-dimensional congestion graph and overflow network to the second maze routing stage for final routing. The application solves the problems of long global routing running time, serious repeated maze routing and insufficient practicality of traditional congestion prediction model in the prior art.

The invention discloses a high-density standard cell global wiring device driven by deep learning, and relates to an integrated circuit design automation technology, and the wiring device comprises a preprocessing module which is used for extracting related features of a layout from an intermediate result of the high-density standard cell global wiring device and constructing a heterogeneous hypergraph; the model reasoning module is used for inputting the heterogeneous hypergraph into a multi-modal congestion prediction model for prediction; the parallel route scheduling module is used for scheduling the first labyrinth wiring stage to carry out advanced wiring on the predicted overflow network; and the post-processing module is used for screening and replacing the wiring result of the first labyrinth wiring stage, and submitting the updated three-dimensional congestion map and the overflow network to the second labyrinth wiring stage for final wiring. According to the method, the problems of long global wiring operation time, serious repeated labyrinth wiring and insufficient practicability of a traditional congestion prediction model under a high-density standard unit in the prior art are solved.

The application relates to the technical field of quantum computer control and belongs to a quantum device instruction distribution system and method based on multi-thread parallel routing, which comprises a host computer, an instruction board, an instruction routing board and a device end, and the instruction distribution method comprises the following steps: S1, inputting program instructions and control instructions through the host computer and delivering the program instructions and the control instructions to the instruction board through a local bus interface; S2, the program instructions are subjected to multi-thread parallel execution in an FPGA chip of the instruction board, control instructions are generated, the control instructions are output to the routing board through a high-speed radio frequency interface, and the control instructions comprise target address codes; S3, the instruction routing board dynamically selects an output port according to the target address codes generated in S2, and distributes the instructions to specified devices; and S4, each device in the device end receives the instructions through an SSMA sequence interface and executes the instructions. Through the multi-thread parallel execution and the dynamic routing mechanism, efficient synchronous control of quantum devices is realized.

This application discloses a method and system for end-to-end order-preserving scaling and scaling of stream processing based on fixed logical partitions. This method introduces a fixed logical partition layer between business keys and physical resources, establishing a two-level mapping architecture of "business key → logical partition → physical partition," thus decoupling business identity from physical location. Logical partition identifiers are encoded into message key-value pairs and transmitted transparently throughout the entire chain, enabling each layer to work collaboratively based on a unified logical view. During scaling, a three-phase consistency switching protocol is executed, and end-to-end data order preservation is ensured through dual-version parallel routing, cross-partition consumption merging, and a dual-condition confirmation mechanism. This invention resolves the core contradiction between "order preservation" and "elastic scaling" in stream processing systems, achieving strict order preservation, smooth switching, and automatic disaster recovery during dynamic adjustment of physical resources. It possesses good scalability and versatility, and is applicable to various distributed stream processing scenarios requiring strict order preservation.

The application provides a method and system for fast searching a routing table based on FPGA, wherein routing table data is transmitted to routing FPGA software according to a specific planning length by a routing CPU software in a satellite routing device system, the routing table data is stored in the RAM resource in the routing FPGA in a RAM resource saving mode, and up to 4096 routing IP data information can be supported.When the routing FPGA software receives a routing IP search request, a specific parallel routing table searching mode is adopted to obtain the routing port number corresponding to the routing IP in the shortest time and output the same. The method of the application supports routing table updating in units of one routing table, and obtains the searching result of the routing table in the shortest time on the premise of saving the RAM resource. The function of fast searching the routing table of the networking star routing device is realized, the number of networking stars that can be supported is greatly expanded, and the efficiency of searching the routing IP is greatly improved.

The invention relates to a modular cable tray (1) for routing cable bundles in an aircraft wing structure, this modular cable tray (1) comprising: - at least one flexible routing module (2), each routing module (2) extending linearly and having two ends; - at least two parallel routing paths (2b) in each routing module (2); - a separation gap between two adjacent routing paths (2b); - a rigid connecting module (3) at each end of a routing module (2); and - at least one retaining clip (4) for each connecting module (3) on one end (2a) of each adjacent routing module (2). Abstract figure: [Fig. 2]