Dynamic Memory Controller Frequency Adjustment for GPU Power Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory controllers in SoC systems face challenges in optimizing power and performance, particularly in managing the operating frequency of memory access by multiple processing units like GPUs and CPUs, which affects overall system efficiency and power consumption.
Innovation Solution
Implementing a dynamic clock voltage scaling (DCVS) mechanism that monitors the active time of processing units, compares it with an active threshold, and adjusts the memory controller's operating frequency accordingly, starting at a higher frequency for moderate to high GPU utilization to reduce active time and relax timing requirements for CPUs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the memory controller operates at a higher frequency to serve GPU memory requests, then GPU performance is improved, but system power consumption increases
Solution Approach 1:
The memory controller frequency is made dynamic rather than static. The frequency controller continuously monitors GPU activity and adjusts the memory controller operating frequency in real-time based on actual demand, allowing the system to operate at high frequency only when necessary and reduce frequency during low-demand periods
Solution Approach 2:
The operating frequency parameter of the memory controller is changed dynamically based on GPU activity thresholds. When GPU activity exceeds a threshold, the frequency is increased to improve performance; when activity is low, the frequency is reduced to save power
2Use of energy by moving object
If the memory controller operates at a lower frequency to reduce power consumption, then system power is reduced, but GPU memory access speed decreases
Solution Approach 1:
The system transitions from a static low-frequency operation to a dynamic frequency adjustment mechanism that responds to real-time GPU activity, ensuring frequency is increased only when performance is actually needed
Solution Approach 2:
A feedback loop is implemented where the frequency controller monitors GPU activity and uses this information to adjust memory controller frequency accordingly. This closed-loop control ensures the system maintains optimal performance while minimizing power consumption
3Loss of time
If the memory controller frequency is increased to reduce GPU active time, then GPU processing speed is improved, but CPU timing requirements become more stringent
Solution Approach 1:
The frequency adjustment occurs periodically based on monitored GPU activity patterns. The system monitors activity over time and adjusts frequency in response to sustained activity patterns rather than instantaneous spikes, creating a more predictable timing pattern for the CPU
Data Source
AI summary
In some aspects, a method for adjusting an operating frequency of a memory controller is provided, wherein a graphics processing unit (GPU) accesses a memory via the memory controller. The method includes monitoring activity of the GPU to determine an active time of the GPU, comparing the determined active time with an active threshold, and, if the determined active time is greater than the active threshold, increasing the operating frequency of the memory controller.