Memory Performance State Control for Power and Thermal Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer systems generate excessive heat and consume excessive power, which can lead to damage and increased costs, as existing technologies do not effectively manage thermal output and power consumption based on activity levels.
Innovation Solution
The system and method establish multiple memory performance states in a computing system, allowing the operating system power management component to control core frequency and restrict available states based on power consumption and thermal output, enabling reduced power consumption and thermal output by transitioning to lower performance states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the processor operates at high performance, then computing speed is improved, but power consumption and thermal output increase
Solution Approach 1:
The patent implements dynamic power management by allowing the system to transition between multiple performance states (P-states) based on actual workload demands. The processor can dynamically adjust its operating frequency and voltage, switching between high-performance states when computing speed is needed and low-power states when workload is light, thereby resolving the contradiction between productivity and energy consumption.
Solution Approach 2:
The system changes operational parameters (frequency, voltage, performance state) of the processor based on workload conditions. By adjusting these parameters dynamically, the system can optimize the balance between computing speed and power consumption, selecting appropriate parameter sets that match current performance requirements without excessive energy use.
2Productivity
If the processor operates at high performance, then computing speed is improved, but thermal output increases
Solution Approach 1:
The dynamic performance state management allows the processor to adjust its thermal output dynamically based on workload. When computing speed is not critical, the system transitions to lower performance states that generate less heat, thereby managing thermal output while maintaining adequate productivity when needed.
Solution Approach 2:
The system proactively manages thermal output by transitioning to lower performance states before thermal conditions become problematic. The power management infrastructure anticipates thermal issues and adjusts performance states preemptively, preventing excessive heat generation before it becomes a problem.
3Use of energy by moving object
If power consumption is reduced, then energy efficiency is improved, but computing performance decreases
Solution Approach 1:
The system applies partial performance - using only the computing performance necessary for current workload demands. Instead of always operating at full capacity or always reducing to minimum, the system dynamically adjusts to provide exactly the right amount of performance needed, avoiding both excessive energy consumption and unnecessary performance limitations.
Solution Approach 2:
The power management system uses feedback from workload monitoring to dynamically adjust performance states. By continuously monitoring actual computing demands and adjusting performance accordingly, the system ensures that power consumption is optimized without sacrificing required computing performance, as adjustments are based on real-time feedback about actual needs.
Data Source
AI summary
A system and method is disclosed for providing memory performance states in a computing system. The operating system power management component of the computing system establishes a set of performance states, with each performance state being defined by a number of factors, including the core frequency of memory. The operating system power management component also defines the number of memory performance states that are supported by the computing system and the number of supported memory performance states that are available for use by the computing system. Whether a supported memory performance state is available is dependent upon a measure of the power being consumed by the computing system, the thermal output of the computing system, or both measures.