This inventor holds 1 USPTO granted patent. Top assignee: Microsoft Technology Licensing, LLC. Active years: 2026.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2026
Title: Torsten Hoefler: Patent Work in Spatial Computing
Introduction
Torsten Hoefler is an inventor associated with Pfaeffikon, Schwyz, Switzerland. His patent work includes a method for assigning workloads to physical resources in spatial architectures. The available information lists 1 patent and identifies Microsoft Technology Licensing, LLC as the company associated with the invention.
Latest Patents
Hoefler’s listed patent, “Assigning Workloads to Physical Resources in Spatial Architectures,” describes distributing a workload across a physical device array. The workload is represented as a graph of operations, which is partitioned into subgraphs to reduce the number of subgraphs and improve resource utilization within each one.
The method then maps each subgraph to logical processing engine units using the subgraph’s features and the units’ tiling factors. That logical mapping is assigned to physical processing engine units, with the assignment designed to minimize network traffic across the array. The selected physical units then perform the subgraph’s operations, a process intended to improve computational efficiency.
Career Highlights
The available information connects Hoefler’s patent activity with Microsoft Technology Licensing, LLC. The patent focuses on workload partitioning, resource mapping, and reducing communication overhead in spatial computing architectures.
Collaborations
Fanny Nina Paravecino and Michael Eric Davies are identified as Hoefler’s coworkers. The provided information does not specify their roles in the patent or describe particular collaborative contributions.
Conclusion
Torsten Hoefler’s listed invention addresses efficient execution of workloads on physical device arrays by combining graph partitioning, logical mapping, and traffic-conscious assignment of processing resources.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
