Dynamic Shader Engine Allocation for APU Power-Performance Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Larger APUs face inefficiencies in power consumption and performance due to idle graphics pipeline leakage and wasted power distribution when executing workloads with low concurrent active contexts, leading to suboptimal performance-to-power ratios and reduced battery life.
Innovation Solution
Dynamic and software-controlled allocation of shader engines based on application profiles and power configurations, enabling selective activation and deactivation of shader engines to optimize performance-to-power ratios by adjusting the number of active engines according to specific application needs and power sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all shader engines are activated in larger APUs, then maximum graphics processing performance is achieved, but power consumption increases due to idle graphics pipeline leakage and wasted power distribution
Solution Approach 1:
The system dynamically activates or deactivates shader engines based on runtime workload requirements. The driver or run list controller determines the appropriate number of shader engines to activate for each application, allowing the system to adapt power consumption to actual processing needs rather than maintaining a static configuration.
Solution Approach 2:
The system changes the operational parameter of shader engine activation state (active/inactive) based on application profiles and workload characteristics. By modifying this binary parameter dynamically, the system optimizes the balance between performance and power consumption for different application scenarios.
2Ease of operation
If shader engines are statically enabled at system initialization, then hardware resources are readily available, but the system cannot adapt to varying application requirements and maintains suboptimal performance-to-power ratios
Solution Approach 1:
The system transitions from static hardware configuration to dynamic runtime adaptation. Shader engine activation is determined at runtime based on application profiles and workload characteristics, enabling the system to optimize resource allocation for each specific application while maintaining ease of operation through automated driver-level control.
Solution Approach 2:
The system performs self-configuration of shader engine activation based on application profiles stored in the driver. The run list controller automatically determines the appropriate number of shader engines to activate without requiring manual user intervention, enabling the system to serve itself optimally for different application scenarios.
3Device complexity
If the system operates with a fixed number of shader engines, then hardware configuration is simple, but power efficiency cannot be optimized for different application types and workload intensities
Solution Approach 1:
The system implements dynamic shader engine allocation that adapts to different application types and workload intensities. By controlling the activation state of individual shader engines based on runtime requirements, the system optimizes power efficiency without complicating the underlying hardware configuration, as the complexity is managed at the software/driver level.
Data Source
AI summary
Techniques are described for implementing selective activation and deactivation of a dynamically allocated subset of shader engines, such as based on application-based profile information and/or on an active system power configuration. Instructions for execution are received from an application associated with a first application profile. Based on the application profile, a quantity of activated shader engines in a plurality of shader engines is modified. The quantity of activated shader engines is further modified responsive to receiving additional instructions from a second application, and/or to receiving one or more indications of an altered active system power configuration.