DDR Memory Training with Asynchronous Temperature Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern DDR DRAMs require extensive training of electrical parameters during system startup, which is time-consuming and disruptive, especially when power states change due to heating, leading to potential stalls and user experience issues.
Innovation Solution
Implementing a sensor circuit in the memory that allows asynchronous temperature measurement during low power mode, enabling adjustment of training parameters based on actual operating conditions without full retraining, using a self-timed sequencer to output temperature measurements asynchronously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full retraining of electrical parameters is performed when power states change, then memory operation reliability is improved, but system latency increases and productivity decreases
Solution Approach 1:
The patent applies preliminary action by performing initial comprehensive training of electrical parameters during system startup or low-power states, then storing these trained values for rapid retrieval and adjustment during power state transitions. This avoids the need for complete retraining during every power state change, thereby reducing latency while maintaining reliability.
Solution Approach 2:
The patent implements dynamics by enabling the memory system to adaptively adjust trained parameter values based on current operating conditions and power states. Rather than static parameters, the system dynamically selects and modifies pre-trained values to optimize performance for each specific power state, balancing reliability and response time.
2Stability of the object's composition
If comprehensive parameter training is performed during system startup, then memory stability is improved, but system boot time increases and productivity decreases
Solution Approach 1:
The patent applies preliminary action by completing comprehensive parameter training during system startup or initialization phases, then caching the trained values for rapid access during subsequent operations. This ensures memory stability is established upfront, while subsequent power state changes can proceed quickly using the pre-trained reference values.
Solution Approach 2:
The patent implements partial action by performing complete parameter training only when necessary (during initial system startup or when significant parameter drift is detected), rather than continuously. For routine power state transitions, only selective parameter adjustments are made based on temperature sensors and predefined thresholds, reducing overall training time while maintaining sufficient stability.
3Temperature
If frequent power state changes are implemented to manage thermal conditions, then temperature control is improved, but system latency increases due to retraining requirements
Solution Approach 1:
The patent applies preliminary action by pre-training parameter sets for multiple anticipated power states during system initialization. When power state changes are needed for thermal management, the system can rapidly switch between pre-trained parameter sets without performing complete retraining, thus maintaining fast transition speeds while enabling frequent power state changes for temperature control.
Solution Approach 2:
The patent implements dynamics by creating a dynamic mapping between power states, temperature conditions, and optimized parameter sets. The system continuously monitors temperature sensors and dynamically selects appropriate pre-trained parameter values, enabling rapid adaptation to thermal conditions without the latency penalty of frequent complete retraining cycles.
Data Source
AI summary
A data processing system includes a data processor having a memory controller, and a memory. The memory is coupled to the memory controller and is for reading and writing data synchronously with respect to a clock signal. The memory includes a sensor circuit that is responsive to a control signal to output a measured value without using the clock signal.