Contents Menu Expand Light mode Dark mode Auto light/dark mode
Logo. Logo.
PowerVR Docs
Developer Forum Email Developer Support
PowerVR Tools Downloads PowerVR SDK on GitHub PowerVR Tools EULA PowerVR SDK EULA
Imagination Technologies Careers
Introduction to PowerVR for Developers
Introduction to PowerVR for Developers
  • Introducing PowerVR
  • History of PowerVR
  • Modern GPUs
    • How Does a GPU Differ From a CPU?
    • Parallelism
    • Vector and Scalar Processing
  • PowerVR Architecture Overview
    • The PowerVR Advantage
    • Tile-Based Deferred Rendering (TBDR)
      • Vertex Processing (Tiler)
      • Per-Tile Rasterization (Renderer)
      • Further TBDR Details
    • Hidden Surface Removal Efficiency
    • Unified and non-unified shader architectures
  • Optimising for PowerVR
    • Do Understand the Target Device
    • Do Profile the Application
    • Do Not Use Alpha Blend Unnecessarily
    • Do Perform Clear
    • Do Not Update Data Buffers Mid-Frame
    • Do Use Texture Compression
    • Do Use Mipmapping
    • Do Not Use Discard
    • Do Not Force Unnecessary Synchronisation
    • Do Move Calculations ‘Up the Chain’
    • Other Considerations When Optimising for PowerVR
      • Do Group Per Material
      • Do Not Use Depth Pre-pass
      • Do Prefer Explicit APIs
      • Do Prefer Lower Data Precision
      • Do Use All CPU Cores
      • Do Use Indexed Lists
      • Do Use Level of Detail (LoD)
      • Do Use On-chip Memory Efficiently for Deferred Rendering
  • Glossary
  • Further Information
Back to top

PowerVR Architecture Overview#

  • The PowerVR Advantage
  • Tile-Based Deferred Rendering (TBDR)
    • Vertex Processing (Tiler)
    • Per-Tile Rasterization (Renderer)
    • Further TBDR Details
  • Hidden Surface Removal Efficiency
  • Unified and non-unified shader architectures
Next
The PowerVR Advantage
Previous
Vector and Scalar Processing
© Imagination Technologies Limited. All rights reserved.
Privacy Policy | Terms & Conditions
Made with Sphinx and Furo.