CPU-GPU Joint Dose Calculation for Faster Particle Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dose calculation systems for radiation therapy, particularly in boron neutron capture therapy, rely on single types of computer hardware (CPU or GPU) for acceleration, leading to waste of computing resources and suboptimal calculation speed.
Innovation Solution
A control method and apparatus that utilizes both CPU and GPU for joint dose calculation, determining a suitable joint dose calculation mode based on their computing speeds to maximize resource utilization and improve calculation speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If only a single type of computer hardware (CPU or GPU) is used for dose calculation, then the system is simple to implement, but computing resources are wasted and dose calculation speed is not maximized
Solution Approach 1:
The patent combines CPU and GPU into a unified dose calculation system that works together. The system divides calculation tasks between CPU and GPU based on their respective strengths, with CPU handling complex logic and GPU handling parallel computation, thereby merging their capabilities to achieve maximum computing efficiency while resolving the contradiction between system complexity and calculation speed.
Solution Approach 2:
The patent implements dynamic task allocation between CPU and GPU based on real-time computing load and task characteristics. The system can dynamically adjust which processor handles which calculation task, optimizing resource utilization and achieving maximum dose calculation speed while adapting to varying computational requirements.
2Productivity
If CPU and GPU are used for joint dose calculation, then computing resources are fully utilized and calculation speed is maximized, but the system becomes more complex to control
Solution Approach 1:
The patent implements an automatic task distribution mechanism where the system autonomously decides which tasks should be executed by CPU and which by GPU based on predefined criteria and real-time performance metrics. This self-service approach eliminates the need for manual intervention in task allocation, simplifying operation while maximizing computational efficiency.
Solution Approach 2:
The system incorporates feedback mechanisms that monitor the performance and load of both CPU and GPU, automatically adjusting task distribution based on real-time conditions. This feedback loop ensures optimal utilization of computing resources while maintaining ease of operation, as the system self-regulates without requiring complex manual control.
3Measurement precision
If Monte Carlo method is used for dose calculation, then calculation accuracy is improved, but computation time increases significantly
Solution Approach 1:
The patent segments the Monte Carlo calculation process into multiple independent parallel tasks that can be simultaneously executed on GPU cores. By dividing the particle simulation into discrete, parallelizable units, the system maintains the accuracy of the Monte Carlo method while dramatically reducing computation time through parallel processing.
Solution Approach 2:
The patent replaces the traditional sequential CPU-based Monte Carlo calculation with a parallel GPU-based implementation. This substitution leverages the massively parallel architecture of GPUs to perform numerous Monte Carlo simulations simultaneously, maintaining calculation accuracy while reducing computation time by orders of magnitude.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The present application relates to a control method and apparatus for a dose calculation system, and a dose calculation system. The method includes: building a dose calculation geometric model; determining a CPU and GPU joint dose calculation mode from a computing speed of a CPU and a computing speed of a GPU; and performing, by the CPU and the GPU, particle simulation according to the CPU and GPU joint dose calculation mode and the dose calculation geometric model to obtain a dose calculation result. By using this method, computing resources of the dose calculation system can be fully used, thereby maximizing the improvement of a dose calculation speed.