Intel corporation (20240354162). GRAPH ORCHESTRATOR FOR NEURAL NETWORK EXECUTION simplified abstract
Contents
GRAPH ORCHESTRATOR FOR NEURAL NETWORK EXECUTION
Organization Name
Inventor(s)
Srikanth Vasuki of Castleknock (IE)
GRAPH ORCHESTRATOR FOR NEURAL NETWORK EXECUTION - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240354162 titled 'GRAPH ORCHESTRATOR FOR NEURAL NETWORK EXECUTION
Simplified Explanation: The patent application describes a system where a barrier is inserted into a graph representing workloads in a neural network execution to control the flow of data between producing and consuming workloads.
- The system includes a graph orchestrator that updates the status of the barrier based on the completion of producing workloads.
- When a producing workload is complete, the graph orchestrator sends a barrier lift message to the consumer workload to start processing the data.
- This system ensures efficient coordination and data flow in a neural network execution.
Key Features and Innovation:
- Inserting a barrier into a graph to control data flow in a neural network execution.
- Graph orchestrator updating barrier status based on producing workload completion.
- Sending a barrier lift message to the consumer workload to initiate processing.
Potential Applications:
- Neural network training and inference processes.
- Distributed computing systems.
- Data processing pipelines in machine learning applications.
Problems Solved:
- Efficient coordination of producing and consuming workloads.
- Controlled data flow in neural network executions.
- Ensuring timely processing of data generated by producing workloads.
Benefits:
- Improved efficiency in neural network executions.
- Better resource utilization in distributed computing systems.
- Enhanced performance in data processing pipelines.
Commercial Applications: The technology can be applied in various industries such as healthcare, finance, and e-commerce for optimizing neural network operations, improving data processing efficiency, and enhancing machine learning applications.
Prior Art: Prior research may exist in the field of distributed computing systems, neural network orchestration, and data flow control in machine learning frameworks.
Frequently Updated Research: Stay updated on advancements in neural network orchestration, distributed computing systems, and data flow control in machine learning applications.
Questions about the Technology: 1. How does the system ensure efficient coordination between producing and consuming workloads? 2. What are the potential implications of this technology in optimizing neural network operations?
Original Abstract Submitted
a barrier may be inserted into a graph representing workloads in an execution of a neural network and placed between a producing workload performed by a producer and a consuming workload performed by a consumer. the consuming workload is to be performed using data generated from the producing workload. a graph orchestrator may modify status information of the barrier in response to receiving a message from the producer. the status information indicates whether one or more producing workloads associated with the barrier are complete. the message indicates that the producing workload is complete. the graph orchestrator may determine whether the one or more producing workloads are complete based on the modified status information. in response to determining that the one or more producing workloads are complete, the graph orchestrator may provide a barrier lift message to the consumer. the barrier lift message causing the consumer to start the consuming workload.
