Parallel Processing Processor for Binary Variable Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing processing techniques for optimizing binary variables under a one-hot constraint face limitations in both speed and accuracy, particularly in implementing efficient parallelization and improving solving accuracy for problems like the Ising model, where traditional methods such as GPGPU-based annealers struggle with high performance and versatility.
Innovation Solution
A processing system utilizing a parallel processing processor with multiple threads per block optimizes binary variables by defining group variables with combination patterns satisfying the one-hot constraint, assigning solution candidates to threads, and searching for output values based on energy evaluation values, enabling parallel and accurate processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional GPGPU-based annealing methods are used to optimize binary variables under one-hot constraint, then the processing can be performed in parallel, but the solving accuracy is limited and performance is insufficient
Solution Approach 1:
The patent divides the optimization problem into multiple blocks, where each block contains multiple threads that independently search for solution candidates. This segmentation enables parallel processing while maintaining search diversity, thereby improving both solving speed and accuracy by exploring multiple solution spaces simultaneously without the accuracy limitations of traditional GPGPU-based annealing methods.
2Productivity
If the optimization problem is divided into multiple partial problems to improve solving performance, then the processing speed increases, but the solving accuracy is limited by the division
Solution Approach 1:
The patent implements a hierarchical structure where threads are nested within blocks, and multiple blocks work simultaneously on different portions of the search space. Each block contains multiple threads that explore solution candidates, and the results from all blocks are aggregated to find the global optimum. This nested organization allows parallel processing to speed up solving while maintaining comprehensive search coverage to ensure accuracy.
3Productivity
If more threads are used for parallel processing to increase speed, then the solving performance improves, but the device complexity increases
Solution Approach 1:
The patent organizes threads into discrete blocks, where each block is an independent processing unit containing multiple threads. This segmentation allows the system to scale parallel processing capacity by adding or configuring blocks without requiring complex inter-thread communication infrastructure. Each block can be independently managed, reducing overall system complexity while enabling high-speed parallel solving through coordinated block execution.
Data Source
AI summary
A processing system includes a parallel processing processor in which threads is constructed for each of blocks, and that optimizes a combination of binary variables under a one-hot constraint. A group variable is defined with a combination pattern satisfying the one-hot constraint as a solution candidate for each of groups of the binary variables. The parallel processing processor executes assigning the solution candidate of the group variable for each of the threads in each of the blocks, searching for an output value of the group variable in each of the blocks, and outputting the output value of all the group variables having been searched.