Task Queue Checkpoints for Dynamic Execution State Monitoring

The patent segments the task execution process into multiple checkpoint instructions distributed throughout the code. Each checkpoint creates a monitoring point that allows the system to dynamically check task state without disrupting overall execution. This segmentation enables both simple implementation (by inserting small checkpoint instructions) and effective monitoring (by creating multiple observation points).

The patent applies preliminary action by pre-defining checkpoint instructions at key locations in the code before task execution begins. These checkpoints are predetermined positions where monitoring will occur, allowing the system to prepare monitoring capabilities in advance while maintaining simple task execution flow. The checkpoints are embedded during code compilation or setup phase.

The patent applies local quality by differentiating data importance levels and applying redundancy processing selectively rather than uniformly. Critical data fields receive full redundancy protection (e.g., replica redundancy or ECC), while less critical fields receive reduced or no redundancy. This localized approach ensures data reliability for important information while minimizing storage overhead for less important data.

The patent changes the redundancy parameter based on data importance classification. For high-importance data, it uses higher redundancy levels (such as multiple replicas or strong error correction), while for low-importance data, it uses lower redundancy levels or none at all. This parameter adjustment optimizes the balance between reliability and storage efficiency.

The patent reduces memory access power consumption by applying redundancy processing only to critical data fields that require protection. By identifying and marking important bits or fields, the system performs redundancy operations locally only where necessary, rather than processing entire data structures uniformly. This selective approach significantly reduces the number of memory accesses required for redundancy operations.

The patent applies partial action by performing redundancy processing on only a subset of data (the important/critical portions) rather than all data. This partial processing approach provides sufficient reliability for critical information while avoiding the excessive power consumption that would result from processing entire data sets, achieving an optimal balance between protection and energy efficiency.