Multi-Core GPU Frame Scheduling Using Future Frame Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current graphics processing techniques are associated with sequential frame processing, limiting the utilization of information from later frames to optimize earlier frames, leading to increased overhead, power consumption, and potential visual stuttering.
Innovation Solution
A method that involves obtaining frame data from multiple graphics processor cores and altering the workload of one frame based on previously executed workloads and shared resources from another frame, optimizing the processing by discarding unnecessary resources and reducing on-chip memory transfers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sequential frame processing is used, then processing simplicity is maintained, but frame rate and resource utilization are reduced
Solution Approach 1:
The system performs preliminary actions by executing future frame workloads in advance on available graphics processor cores while previous frames are still being processed. This allows the GPU to prepare rendering operations for frames that will be displayed soon, utilizing otherwise idle processing time and improving overall frame rate without significantly increasing complexity through automated workload scheduling.
2Productivity
If future frame information is utilized, then processing efficiency is improved, but information availability constraints are overcome
Solution Approach 1:
The system overcomes information availability constraints by performing preliminary rendering of future frames in advance. The graphics processor executes workload commands for frames that have not yet been fully processed sequentially, using the time between frame submissions to perform optimization operations that would normally require waiting for future frame data to become available.
3Loss of energy
If workload optimization is applied, then power consumption is reduced, but computational overhead increases
Solution Approach 1:
The system reduces power consumption by performing workload optimization in advance on future frames before they need to be rendered. By analyzing and preparing optimization commands for frames that are upcoming, the system can efficiently manage GPU resource allocation and avoid unnecessary rendering operations, thereby reducing overall power consumption without proportionally increasing computational overhead through automated optimization scheduling.
Data Source
AI summary
This disclosure provides systems, devices, apparatus, and methods, including computer programs encoded on storage media, for optimization of graphics workloads through future frames. The method includes obtaining first frame data at a first graphics processor core of a graphics processor and second frame data at a second graphics processor core of the graphics processor, where the first frame data is associated with a first time instance and the second frame data is associated with a second time instance. The method includes altering a first workload associated with the first frame data at the first graphics processor core based on a portion of a second workload for the second frame data that has been previously executed by the second graphics processor core and a shared resource between the first graphics processor core and the second graphics processor core and outputting an indication of the altered first workload.