SimXACT Gate-Level Simulation X-Pessimism Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Gate-level simulation is hindered by X-pessimism, which generates false unknown values (Xs), leading to inaccurate results due to physical optimizations and low-power requirements, with existing solutions either being non-generic, time-consuming, or requiring significant memory and search space.
Innovation Solution
The SimXACT method analyzes sequential fan-in cones to identify and generate fixes for false Xs, using algorithms to trace and replace Xs with correct values, producing auxiliary code that eliminates combinational false Xs and ensures accurate simulation results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If gate-level simulation is performed with physical optimizations and low-power requirements, then design verification capability is improved, but false Xs increase due to X-pessimism
Solution Approach 1:
The patent segments the simulation process into two distinct phases: standard gate-level simulation and formal analysis. The formal analysis phase specifically targets and analyzes false Xs identified during simulation, separating the detection and resolution of X-pessimism problems from the main simulation flow. This segmentation allows each phase to optimize for its specific purpose without compromising the other.
Solution Approach 2:
The patent introduces an intermediary formal analysis component that acts as a mediator between the simulator and the design netlist. This formal analysis tool receives the design netlist and simulation waveforms, performs specialized analysis to identify false Xs, and provides corrections back to the simulation process. The intermediary handles the complexity of X-pessimism resolution without requiring changes to the core simulation engine.
2Measurement precision
If formal analysis is applied at simulation time to find false Xs, then simulation accuracy is improved, but analysis time increases
Solution Approach 1:
The patent performs preliminary formal analysis on the design netlist before full simulation begins. During this preliminary phase, the system pre-identifies potential sources of false Xs and prepares analysis data structures. This preliminary action reduces the computational burden during actual simulation by having much of the analysis work completed in advance.
Solution Approach 2:
The patent applies partial formal analysis only to specific regions of the design where false Xs are most likely to occur, rather than performing exhaustive analysis on the entire design netlist. The system identifies critical paths and registers prone to X-pessimism and focuses analysis resources on these areas, achieving sufficient accuracy without the time cost of complete analysis.
3Measurement precision
If all possible paths for X-propagation are analyzed, then false X identification is improved, but search space requirements increase significantly
Solution Approach 1:
The patent extracts and isolates specific signals and registers that are prone to false Xs from the larger design netlist. By identifying and extracting only the relevant portions of the design that contribute to X-pessimism, the system reduces the search space from the entire design to just the critical subsets that need analysis. This extraction approach maintains identification accuracy while dramatically reducing computational complexity.
Solution Approach 2:
The patent applies different analysis depths and methods to different regions of the design based on their local characteristics. Critical registers and combinational logic paths receive more thorough analysis, while less critical areas receive lighter analysis. This local quality approach ensures that resources are concentrated where they are most needed for false X identification, rather than uniformly analyzing the entire design.
4Reliability
If sub-circuit duplication is performed to find false Xs, then X-pessimism resolution is improved, but memory requirements and time consumption increase
Solution Approach 1:
The patent creates simplified copies or models of sub-circuits rather than duplicating the full detailed netlist. These copies contain only the essential logic needed to analyze X-propagation behavior, omitting unnecessary details. This copying approach provides sufficient information for false X identification while using far less memory than full sub-circuit duplication would require.
Data Source
AI summary
A computer executable tool analyzes unknowns (Xs) in gate-level simulation and traces their sources to determine if the Xs are generated due to X-pessimism. For Xs generated due to X-pessimism, fixes are generated to correct simulation results. Corrected simulation results match real hardware behavior and greatly reduce the analysis effort of engineers.