Memory Transaction-Level Modeling Selective Bank Status Updates

The patent divides the bank status tables into different groups based on their update requirements. Instead of updating all bank status tables uniformly, the system segments them into those that need updating and those that don't, based on the specific memory command being executed. This segmentation allows the simulation to maintain accuracy for relevant banks while skipping unnecessary updates to others, thereby resolving the contradiction between simulation accuracy and efficiency.

The patent applies partial action by updating only the necessary subset of bank status tables rather than all tables. The system determines which specific bank status tables require updates based on the current memory operation, and performs updates only on those tables. This partial updating approach maintains the required simulation accuracy while significantly reducing the computational overhead associated with updating all tables, thus improving simulation efficiency.

The patent employs preliminary action by pre-determining which bank status tables need to be updated before actually performing the updates. The system analyzes the incoming memory command and identifies the specific banks that will be affected, then prepares the update operations only for those banks. This preliminary identification prevents unnecessary computational work on banks that wouldn't change state, thereby maintaining result reliability while reducing computational overhead.

The patent applies local quality by making different parts of the bank status table structure have different update behaviors. Instead of treating all bank status tables uniformly, the system applies updates selectively to specific local regions (individual banks or groups of banks) based on the current operation's requirements. This localized updating approach ensures reliability for affected banks while minimizing computational waste on unaffected banks.