GPU Accelerated Third-Order Low-Rank Tensor Completion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing CPU-based third-order low-rank tensor completion methods are computationally inefficient, especially for large-scale tensors, due to high computational time and exponential increase in running time with tensor size, making them unsuitable for processing large tensors.
Innovation Solution
A GPU-based third-order low-rank tensor completion method that utilizes the GPU's parallelism and memory bandwidth to accelerate the completion process through steps involving data transmission, memory allocation, and iterative least squares methods to minimize tensor completion time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If CPU-based third-order low-rank tensor completion methods are used, then the algorithm can process tensor data, but the computational time increases exponentially with tensor size making it unsuitable for large-scale tensors
Solution Approach 1:
The patent replaces the CPU-based sequential processing system with a GPU-based parallel processing system. The GPU's architecture with thousands of cores enables simultaneous execution of multiple computational tasks, particularly suited for the tensor decomposition operations (SVD, FFT) required in low-rank tensor completion. This substitution transforms the mechanical processing approach from sequential to parallel, achieving exponential speedup for large-scale tensor data.
2Manufacturing precision
If CPU-based methods perform singular value decomposition and Fourier transforms in each iteration, then the algorithm can achieve tensor completion, but the calculation becomes more time consuming
Solution Approach 1:
The patent introduces parallelism as a new dimension for computation by utilizing the GPU's multi-core architecture. Instead of performing SVD and Fourier transforms sequentially on a single CPU core, the computation is distributed across thousands of GPU cores, executing multiple iterations of these operations simultaneously. This dimensional shift from sequential to parallel processing maintains completion accuracy while dramatically reducing total calculation time.
Data Source
AI summary
The present disclosure provides a GPU-based third-order low-rank tensor completion method. Operation steps of the method includes: (1) transmitting, by a CPU, input data DATA1 to a GPU, and initializing the loop count t=1; (2) obtaining, by the GPU, a third-order tensor Yt of a current loop t based on the least squares method; (3) obtaining, by the GPU, a third-order tensor Xt of the current loop t based on the least squares method; (4) checking, by the CPU, whether an end condition is met; and if the end condition is met, turning to (5); otherwise, increasing the loop count t by 1 and turning to (2) to continue the loop; and (5) outputting, by the GPU, output data DATA2 to the CPU. In the present disclosure, in the third-order low-rank tensor completion, a computational task with high concurrent processes is accelerated by using the GPU to improve computational efficiency.