18734123. DISTRIBUTING MATRIX MULTIPLICATION PROCESSING AMONG PROCESSING NODES simplified abstract (Hewlett Packard Enterprise Development LP)

DISTRIBUTING MATRIX MULTIPLICATION PROCESSING AMONG PROCESSING NODES

Organization Name

Hewlett Packard Enterprise Development LP

Inventor(s)

Aaron M. Collier of Bloomington MN (US)

DISTRIBUTING MATRIX MULTIPLICATION PROCESSING AMONG PROCESSING NODES - A simplified explanation of the abstract

This abstract first appeared for US patent application 18734123 titled 'DISTRIBUTING MATRIX MULTIPLICATION PROCESSING AMONG PROCESSING NODES

Simplified Explanation: The patent application involves selecting a candidate matrix decomposition for matrix multiplication based on the sizes of the decompositions along two dimensions. The multiplication process is then distributed among processor sockets according to this selected decomposition.

• Candidate matrix decompositions identified based on available processor sockets
• Selection of a candidate matrix decomposition based on comparative relationships of sizes along two dimensions
• Distribution of matrix multiplication processing among processor sockets based on the selected decomposition

Potential Applications: This technology can be applied in various fields such as scientific computing, data analysis, machine learning, and image processing where matrix multiplication is a common operation.

Problems Solved: This technology addresses the challenge of efficiently distributing matrix multiplication tasks among multiple processor sockets to optimize performance and resource utilization.

Benefits: - Improved efficiency in matrix multiplication tasks - Enhanced utilization of processor sockets - Faster processing of large matrices

Commercial Applications: The technology can be utilized in high-performance computing systems, cloud computing services, and data centers to accelerate matrix multiplication operations, leading to faster data processing and analysis.

Prior Art: Readers can explore prior research on distributed computing, parallel processing, and matrix decomposition techniques to understand the background of this technology.

Frequently Updated Research: Researchers are continually exploring new algorithms and techniques for optimizing distributed matrix multiplication tasks, so staying updated on the latest advancements in this field is crucial.

Questions about Matrix Decomposition Technology: 1. How does this technology improve the efficiency of matrix multiplication tasks? 2. What are the potential limitations of using candidate matrix decompositions for distributed processing?

Original Abstract Submitted

Based on a predetermined number of available processor sockets, a plurality of candidate matrix decompositions are identified, which correspond to a multiplication of matrices. Based on a first comparative relationship of a variation of first sizes of the plurality of candidate matrix decompositions along a first dimension and a second comparative relationship of a variation of second sizes of the plurality of candidate matrix decomposition sizes along a second dimension, a given candidate matrix decomposition is selected. Processing of the multiplication among the processor sockets is distributed based on the given candidate matrix decomposition.