CPU GPU DCVS Co-optimization for Frame Processing Power

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for managing power consumption in multiprocessor systems, such as those with CPUs and GPUs, face challenges in effectively scaling frequency and voltage to optimize performance and efficiency, particularly during graphics frame processing, leading to suboptimal power management.

Innovation Solution

A system and method that co-optimizes dynamic clock and voltage scaling (DCVS) levels for both CPUs and GPUs based on activity data to minimize combined power consumption during graphics frame processing, using a CPU/GPU DCVS co-optimization module to select optimal DCVS levels that balance performance and power efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If separate DCVS algorithms are used for CPU and GPU to optimize performance within frame processing deadlines, then performance is improved, but power consumption is not optimized

Engineering Contradiction:
Improveframe processing performanceVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent merges separate CPU DCVS and GPU DCVS algorithms into a unified co-optimization framework that jointly determines DCVS levels for both processors. The system receives activity data from both CPU and GPU, evaluates combined power consumption across multiple DCVS level combinations, and selects the optimal combination that minimizes total power consumption while meeting frame processing deadlines, rather than optimizing each processor independently

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system dynamically changes DCVS parameters (frequency and voltage levels) for both CPU and GPU based on real-time activity data. By evaluating multiple DCVS level combinations and selecting the optimal pair that minimizes combined power consumption while satisfying performance constraints, the system adapts processor operating parameters to achieve both performance and power efficiency goals

Inventive Principle:
Principle #35Parameter changes

2Productivity

If DCVS levels are adjusted independently for each processor, then individual processor performance is optimized, but combined power consumption is suboptimal

Engineering Contradiction:
Improveprocessor performanceVSAvoidcombined power consumption
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The system implements feedback mechanisms by receiving activity data from both CPU and GPU, evaluating the impact of different DCVS level combinations on combined power consumption, and adjusting DCVS levels accordingly. The co-optimization module continuously monitors processor activity and power consumption, using this feedback to select optimal DCVS level pairs that minimize total energy loss while maintaining required performance levels

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3137965B1CPU/GPU DCVS co-optimization for reducing power consumption in graphics frame processing
Publication Date: 2019.02.20 QUALCOMM INC
  • EP3137965B1 patent drawingFigure 1
  • EP3137965B1 patent drawingFigure 2
  • EP3137965B1 patent drawingFigure 3

AI summary

Systems, methods, and computer programs are disclosed for minimizing power consumption in graphics frame processing. One such method comprises: initiating graphics frame processing to be cooperatively performed by a central processing unit (CPU) and a graphics processing unit (GPU); receiving CPU activity data and GPU activity data; determining a set of available dynamic clock and voltage/frequency scaling (DCVS) levels for the GPU and the CPU; and selecting from the set of available DCVS levels an optimal combination of a GPU DCVS level and a CPU DCVS level, based on the CPU and GPU activity data, which minimizes a combined power consumption of the CPU and the GPU during the graphics frame processing.