Selective Memory Margin Testing for Boot Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing systems face increased latency during boot processes due to signal training, which becomes necessary as the number of memory modules increases, and environmental changes can cause instability even with signal compensation, leading to potential system failures.
Innovation Solution
Implementing a memory tester that reduces latency by testing only a subset of memory blocks and signals during the boot process, using a Rank Margin Tool to identify margin quality and trigger retraining if necessary, allowing for a fast boot mode without full signal training in every boot process, and leveraging software-based solutions to avoid custom hardware designs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full signal training is performed on all memory modules during boot process, then memory stability is improved, but boot latency increases
Solution Approach 1:
The patent segments the memory testing process by dividing memory blocks into different groups and selectively applying margin tests only to certain segments rather than all memory blocks. This segmentation allows the system to maintain stability checks on critical memory regions while skipping less critical areas, thereby reducing overall boot latency without compromising essential memory stability.
Solution Approach 2:
The patent implements partial action by performing margin tests on only a subset of memory blocks rather than conducting exhaustive testing on all memory modules. The system determines which memory blocks require testing based on previous training results and environmental conditions, applying testing only where necessary to maintain stability while avoiding unnecessary delays from comprehensive testing.
2Quantity of substance
If the number of memory modules increases, then system capacity is improved, but signal training time increases
Solution Approach 1:
The patent applies segmentation by organizing memory modules into groups and selectively applying margin testing to specific segments rather than all modules. This allows systems with multiple memory modules to maintain stability verification on essential segments while reducing the overall testing burden, thereby scaling memory capacity without linearly increasing training time.
Solution Approach 2:
The patent implements preliminary action by using results from previous signal training and environmental monitoring to pre-identify which memory blocks require margin testing. By determining the testing scope in advance based on historical data and current conditions, the system avoids unnecessary testing on stable memory modules, thus reducing training time as system capacity increases.
3Reliability
If environmental changes are compensated with signal training, then memory reliability is improved, but system complexity increases
Solution Approach 1:
The patent applies partial action by implementing margin testing only on memory blocks that are susceptible to environmental changes rather than all memory blocks. The system identifies critical memory regions that require environmental compensation and applies testing selectively to those areas, maintaining reliability for environmental adaptation while avoiding the complexity of comprehensive system-wide testing mechanisms.
4Measurement precision
If margin testing is performed on all memory blocks, then testing accuracy is improved, but testing time increases
Solution Approach 1:
The patent segments the memory testing process by dividing memory blocks into different groups and applying margin tests selectively to specific segments rather than all memory blocks. This segmentation maintains testing accuracy on critical memory regions while reducing overall testing time by excluding non-critical segments from exhaustive testing.
Solution Approach 2:
The patent implements partial action by performing margin testing on only the necessary subset of memory blocks determined by previous training results and environmental conditions. This approach maintains sufficient testing accuracy for system reliability while avoiding the time penalty of exhaustive testing on all memory blocks.
Data Source
AI summary
Method, systems and apparatuses may provide for technology that executes a margin test of a first memory storage based on a subset of first signals associated with the first memory storage. The technology determines, based on the margin test, first margin data to indicate whether the first memory storage complies with one or more electrical constraints. The technology determines, based on the first margin data, whether to execute a signal training process.