Austin, TX, United States of America

James Michael O'Connor

This inventor holds 67 USPTO granted patents and 23 published patent applications and 2 EPO patents, primarily in Memory Systems (CPC class G06F). Top assignees: Advanced Micro Devices Corporation, Nvidia Corporation. Active years: 2011-2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: James Michael O'Connor. Retrieved from https://idiyas.com/inventor/james-michael-o-connor

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

USPTO Granted Patents = 67 

% Patents Active = 95.5

 

Average Co-Inventor Count = 3.2

ph-index = 9

Forward Citations = 352(Granted Patents)

Forward Citations (Not Self Cited) = 308(Dec 10, 2025)


Inventors with similar research interests:


Company Filing History:


Years Active: 2011-2026

Loading Chart...
Loading Chart...
Areas of Expertise:
Stacked Memory
Sparse Matrices
Power Reduction
Memory Hierarchy
Data Bus Inversion
Graphics Processing Unit
Dynamic Random Access Memory
Prefetch Kernels
Memory Interface
Data Cache
Page Migration
Heterogeneous Scheduling
67 patents (USPTO):Explore Patents

Title: A Profile of James Michael O'Connor: Innovator in Memory Systems

Introduction: James Michael O'Connor, based in Austin, TX, is a prolific inventor renowned for his significant contributions to the field of memory systems. With an impressive portfolio of 65 patents, O'Connor has made remarkable advancements that optimize memory performance and computational efficiency.

Latest Patents: O'Connor's latest innovative patents include:

1. **Hierarchical Network for Stacked Memory System**

This invention outlines a hierarchical network that facilitates access for a stacked memory system comprising multiple memory dies, each containing various memory tiles. The processor die integrates several processing tiles stacked with the memory dies. Memory tiles aligned with a processing tile are directly linked to it, forming the local memory block. This architecture allows each processing tile to efficiently access not just its local memory, but also memory blocks in different processing tiles, other stacked dies, and external devices. The hierarchical network significantly enhances memory bandwidth by a factor of 50 and reduces energy consumption for bit transfer by a remarkable 10 times compared to conventional systems.

2. **Techniques for Performing Matrix Computations Using Hierarchical Representations of Sparse Matrices**

This patent introduces a method for executing matrix operations that involves traversing a tree structure to access non-empty regions in a matrix. The innovative approach organizes nodes to represent regions based on the density of non-zero values, allowing for efficient processing of matrix calculations and yielding improved results in matrix operations.

Career Highlights: Throughout his career, James O'Connor has held significant roles in esteemed technology companies, including Advanced Micro Devices Corporation and Nvidia Corporation. His work in these organizations has underscored his expertise in memory design and computational technology, positioning him as a key innovator in the industry.

Collaborations: O'Connor has collaborated with notable colleagues, including Gabriel H Loh and Nuwan S Jayasena. Together, they have pushed the boundaries of innovation in the field of memory technologies, contributing to the development of cutting-edge solutions.

Conclusion: With a strong foundation in memory systems and a robust portfolio of patents, James Michael O'Connor continues to be a formidable force in the innovation landscape. His work not only enhances memory performance but also sets the stage for future advancements in computational efficiency. As technology evolves, O'Connor's contributions are sure to play a critical role in shaping the industry.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…