Non-Removal Semiconductor Topography Simulation with Hardware Acceleration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional EDA tools and algorithms for semiconductor topography simulations are computationally inefficient, particularly in simulating large-scale topography-changing processes, leading to prolonged simulation times and potential inaccuracies that can delay product development and increase experimental costs.
Innovation Solution
A voxel mesh model combined with ray-tracing algorithms and a computing system comprising a general-purpose CPU and hardware accelerator is used to simulate semiconductor surface movement, enabling parallel processing and significantly reducing simulation times without sacrificing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional step-by-step particle flight simulation is used, then simulation accuracy is maintained, but simulation time becomes extremely long for large-scale problems
Solution Approach 1:
The simulation domain is segmented into multiple spatial regions with different resolution requirements. A multi-resolution approach divides the semiconductor structure into coarse and fine grid regions, applying detailed particle flight simulation only where necessary while using simplified models in less critical areas, thereby reducing overall computation time while maintaining accuracy where it matters most
Solution Approach 2:
The method performs preliminary calculations to determine optimal simulation parameters and pre-computes certain geometric relationships and material interaction probabilities. By preparing lookup tables for material reactions and pre-calculating shadowing effects, the actual simulation runs faster without sacrificing accuracy in the critical regions
2Productivity
If conventional EDA tools are used for topography simulation, then basic functionality is provided, but computational efficiency is insufficient for modern nanometer-scale processes
Solution Approach 1:
The system dynamically adjusts simulation parameters such as spatial resolution, time step size, and particle emission rates based on the specific process being simulated and the required accuracy. By adapting parameters like grid density in different regions and adjusting the number of particles simulated, the tool achieves high computational efficiency for nanometer-scale processes without requiring excessive complexity in all areas
Solution Approach 2:
The patent replaces conventional sequential mechanical simulation steps with parallelized computational algorithms. By using GPU acceleration and distributed computing, the system substitutes traditional CPU-based step-by-step processing with highly optimized parallel algorithms that can process multiple particle flights and spatial regions simultaneously, dramatically improving computational efficiency
Data Source
AI summary
The present disclosure provides a method for topography simulation of a physical structure under a topography-changing process. The method includes initializing a voxel mesh as a three-dimensional (3D) representation of a physical structure by a general-purpose processor, generating a plurality of particles, simulating a flight path of at least one of the particles by a hardware-accelerated processor different from the general-purpose processor, identifying a voxel unit in the voxel mesh that intersects the flight path by the hardware-accelerated processor, passing information describing a collision between the one of the particles and the voxel unit from the hardware-accelerated processor to the general-purpose processor, determining a reaction between the one of the particles and the voxel unit by the general-purpose processor, and adding an extra voxel unit adjacent to the voxel unit based on the determining of the reaction.