MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641088390 A) filed by Vellore Institute Of Technology on July 20, 2026, for Context-Aware Autonomous Memory Layout Morphing Engine For Machine Learning Based Runtime Self-Reconfiguration Of Software Memory Hierarchies.
Inventors include Konatham Sumalatha; Shann Antony Suresh; and Krish Chirag Shah.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: The present invention discloses a context-aware autonomous memory layout morphing engine for adaptive runtime optimization of software memory hierarchies. The system comprises a live workload fingerprinting module configured to monitor hardware performance metrics and generate workload fingerprints representative of runtime execution behavior. An on-process machine learning oracle analyzes the workload fingerprints to predict optimized memory configurations, including data structure layout, memory allocation policy, software prefetch aggressiveness, NUMA memory binding, and traversal order. A runtime reconfiguration actuator dynamically applies the recommended configurations to a running process using standard operating system interfaces while maintaining application correctness through a copy-on-write shadow remapping mechanism. A closed-loop feedback mechanism continuously evaluates runtime performance and updates subsequent optimization decisions. The invention operates as a software-only framework on commodity computing platforms without requiring hardware modification or application restart, thereby improving application throughput, cache utilization, memory bandwidth efficiency, translation lookaside buffer performance, and NUMA memory locality across diverse workloads.
Disclaimer: Curated by HT Syndication.
