Fast Clear Shared Surface Graphics Driver

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing solutions for clearing shared surfaces across multiple applications in modern computing systems are inefficient due to CPU/GPU synchronization issues, which prevent the application of optimizations like compression and fast clears, leading to increased memory bandwidth demands and processing bottlenecks.

Innovation Solution

Implementing a method that allows for fast clear operations on shared surfaces by using a compressed control surface, where a graphics driver determines if a shared surface is shared with a desktop compositor or other applications, and utilizes a resolve flag to manage the need for a resolve operation, ensuring transparent optimization without introducing synchronization issues.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional clear operations are performed on shared surfaces, then surface correctness is maintained across multiple applications, but memory bandwidth demands increase and processing bottlenecks occur

Engineering Contradiction:
Improvesurface correctnessVSAvoidprocessing efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the clear operation into two parts: a fast clear on a compressed control surface (smaller data structure) and a selective resolve operation on the actual shared surface. This segmentation allows the majority of clear operations to be performed efficiently on the compressed representation, while maintaining correctness by resolving only when necessary.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates a compressed control surface that is a simplified representation or copy of the actual shared surface. This compressed copy contains only the essential information needed to determine what needs to be cleared, allowing fast clear operations without processing the entire high-resolution surface data.

Inventive Principle:
Principle #26Copying

2Productivity

If compression and fast clears are applied to shared surfaces, then memory bandwidth demands are reduced, but CPU/GPU synchronization issues arise across multiple applications

Engineering Contradiction:
Improveclear operation speedVSAvoidsynchronization complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces a graphics driver intermediary that manages the compressed control surface and coordinates clear operations across multiple applications. The driver acts as a mediator that handles the complexity of synchronization, allowing applications to use fast clears on the compressed surface without directly managing the synchronization themselves.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements a resolve flag mechanism that provides feedback about whether a resolve operation is needed. This feedback system allows the system to dynamically determine when synchronization is necessary, enabling fast clears when safe and triggering resolve operations only when required to maintain correctness.

Inventive Principle:
Principle #23Feedback

3Loss of energy

If fast clear operations are performed on shared surfaces, then processing costs are reduced, but applications lose visibility into when resolve operations are needed

Engineering Contradiction:
Improveprocessing costVSAvoidresolve operation visibility
Core Design Contradiction:
Loss of energyVSLoss of information

Solution Approach 1:

The patent makes the resolve flag visible and accessible to applications, allowing them to self-determine when resolve operations are needed. Instead of the system hiding this information, applications can check the resolve flag themselves and perform resolve operations only when necessary, giving them control over their own optimization decisions.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS9563930B2Techniques for clearing a shared surface
Publication Date: 2017.02.07 INTEL CORP
  • US9563930B2 patent drawing
  • US9563930B2 patent drawing
  • US9563930B2 patent drawing

AI summary

Various embodiments are generally directed to an apparatus, method and other techniques to determine that a shared surface is shared between a first application and a second application, determine that a fast clear operation has been performed on the shared surface, the fast clear operation comprising clearing one or more locations of one or more buffers. Further, various embodiments may include writing pixel value information to the one or more locations of the one or more buffers and performing a resolve operation on the shared surface.