Conference Papers
Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506
Browse
1 results
Search Results
Item Optimizing Split Algorithm Performance: A Heuristic Method for Enhanced Tensor Product Matrix Computations(Institute of Electrical and Electronics Engineers Inc., 2024) Bhowmik, B.; Kumar, S.; Raju, S.R.; Prakash, A.; Mense, O.Optimizing tensor product matrix computations is critical for enhancing computational efficiency in high-performance applications. Traditional algorithms, like the Split algorithm, often struggle due to the unique properties of each matrix involved. This paper presents a novel heuristic method that optimizes the selection of cutting points and matrix ar-rangement, significantly reducing redundant calculations and minimizing memory usage. The proposed approach adapts to the varying characteristics of tensor products, improving performance across different computational scenarios. Enhancing floating-point operation efficiency and CPU utilization delivers substantial speed and efficiency gains, particularly in large-scale tensor product matrix operations, offering a robust solution for complex computational tasks. © 2024 IEEE.