GPU Rendering Mode Switching for Embedded Power Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current graphics processing systems face challenges in efficiently rendering 3D graphics on embedded devices with limited computational power and memory, as they require significant CPU and memory overhead and inaccurate GPU state information, leading to suboptimal rendering modes selection.

Innovation Solution

A GPU is enabled to switch between direct rendering and binning modes based on command packets and real-time GPU state information, allowing it to determine the most efficient rendering mode for a frame, reducing CPU overhead and improving rendering performance and power management.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If direct rendering mode is used, then rendering performance is improved, but CPU and memory overhead increases

Engineering Contradiction:
Improverendering performanceVSAvoidCPU and memory overhead
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system dynamically switches between direct rendering mode and binning rendering mode based on real-time GPU state information. The GPU determines whether to use direct rendering or binning rendering on a per-frame or per-portion basis, allowing the rendering approach to adapt to current system conditions such as GPU load, power consumption levels, and memory availability, thereby optimizing the balance between rendering performance and resource overhead

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the parameter of rendering mode selection from a static CPU-driven decision to a dynamic GPU-driven decision based on multiple parameters including GPU load, power consumption, memory capacity, and computational power. This allows the system to select the most appropriate rendering mode (direct or binning) based on current operational parameters, resolving the contradiction between performance and overhead

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If binning rendering mode is used, then CPU overhead is reduced, but rendering performance deteriorates

Engineering Contradiction:
ImproveCPU overheadVSAvoidrendering performance
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The system employs dynamic mode switching where the GPU evaluates current state and dynamically selects between binning rendering (for lower CPU overhead) and direct rendering (for higher performance) on demand. This eliminates the need for static mode selection and allows the system to use binning rendering only when appropriate, thus reducing CPU overhead without permanently sacrificing rendering performance

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The GPU autonomously determines the rendering mode based on its own state information and the requirements of the rendering task, without requiring complex CPU intervention or inaccurate external state information. This self-service capability allows the GPU to optimize its own performance characteristics by selecting binning or direct rendering based on real-time conditions

Inventive Principle:
Principle #25Self-service

3Measurement precision

If GPU state information is inaccurate, then rendering mode selection is suboptimal, but system complexity increases

Engineering Contradiction:
ImproveGPU state information accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The command packet serves as an intermediary that carries accurate rendering task information from the CPU to the GPU. Instead of relying on complex state tracking or communication mechanisms, the command packet directly conveys the necessary information about the rendering task, allowing the GPU to make accurate mode selection decisions based on this reliable information rather than on potentially inaccurate state estimates

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback by having the GPU use real-time state information about its own operational conditions (power consumption, load) in conjunction with command packet information to determine the optimal rendering mode. This feedback loop ensures that mode selection is based on accurate, current information about both the task requirements and system state, improving decision accuracy without excessive complexity

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3134864B1Flex rendering based on a render target in graphics processing
Publication Date: 2021.01.06 QUALCOMM INC
  • EP3134864B1 patent drawingFigure 1
  • EP3134864B1 patent drawingFigure 2
  • EP3134864B1 patent drawingFigure 3

AI summary

A device comprising a graphics processing unit (GPU) includes a memory and at least one processor. The at least one processor may be configured to: receive a GPU command packet that indicates the GPU may select between a direct rendering mode or a binning rendering mode for a portion of a frame to be rendered by the GPU, determine whether to use the direct rendering mode or the binning rendering mode for the portion of the frame to be rendered by the GPU based on at least one of: information in the received command packet or a state of the GPU, and render the portion of the frame using the determined direct rendering mode or the binning rendering mode.