18235632. PROCESSING COMPUTATIONAL GRAPHS simplified abstract (Google LLC)
Contents
- 1 PROCESSING COMPUTATIONAL GRAPHS
PROCESSING COMPUTATIONAL GRAPHS
Organization Name
Inventor(s)
Paul A. Tucker of Los Altos CA (US)
Jeffrey Adgate Dean of Palo Alto CA (US)
Sanjay Ghemawat of Mountain View CA (US)
PROCESSING COMPUTATIONAL GRAPHS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18235632 titled 'PROCESSING COMPUTATIONAL GRAPHS
Simplified Explanation
The patent application describes methods, systems, and apparatus for processing a computational graph by partitioning it into subgraphs and assigning operations to available devices for efficient computation.
- Receiving a request from a client to process a computational graph
- Obtaining data representing the computational graph with nodes and directed edges
- Identifying available devices for performing the operations
- Partitioning the computational graph into subgraphs
- Assigning operations to available devices for each subgraph
Potential Applications
This technology can be applied in distributed computing systems, machine learning algorithms, and parallel processing tasks.
Problems Solved
Efficient utilization of available computing resources, faster computation of complex graphs, and improved scalability of computational tasks.
Benefits
Optimized resource allocation, reduced processing time, increased system performance, and enhanced overall efficiency in computational tasks.
Potential Commercial Applications
1. Cloud computing services optimization 2. High-performance computing solutions for large-scale data processing
Possible Prior Art
One possible prior art could be the use of distributed computing frameworks like Apache Spark or TensorFlow for parallel processing of computational graphs.
Unanswered Questions
How does this technology handle failures in the assigned devices during computation?
The patent application does not mention how the system deals with failures in the assigned devices during the computation process. It would be essential to understand the fault tolerance mechanisms in place to ensure uninterrupted processing.
What is the impact of network latency on the performance of partitioned computational graphs across multiple devices?
The abstract does not address the potential impact of network latency on the performance of partitioned computational graphs. Understanding how the system mitigates latency issues and optimizes data transfer between devices would be crucial for assessing its real-world applicability.
Original Abstract Submitted
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for receiving a request from a client to process a computational graph; obtaining data representing the computational graph, the computational graph comprising a plurality of nodes and directed edges, wherein each node represents a respective operation, wherein each directed edge connects a respective first node to a respective second node that represents an operation that receives, as input, an output of an operation represented by the respective first node; identifying a plurality of available devices for performing the requested operation; partitioning the computational graph into a plurality of subgraphs, each subgraph comprising one or more nodes in the computational graph; and assigning, for each subgraph, the operations represented by the one or more nodes in the subgraph to a respective available device in the plurality of available devices for operation.
