Graphics Programming weekly - Issue 377 - February 2nd, 2025


Raymarching The Gunk

  • the detailed article provides an overview of the implementation of the Goo-like in “The Gunk”
  • explains the underlying voxel-based implementation, how it’s represented and interacts with gameplay
  • describes how it’s authored and rendered integrated into UE
  • additionally presents how the effect is rendered using a ray-marching-based approach


Sample for cluster-based continuous level of detail rasterization or ray tracing

  • the sample is part of Nvidia RTX Mega Geometry and presents how to use a continuous Level of detail through mesh clustering
  • presents how to rasterize using mesh shaders and use the new VK_NV_cluster_acceleration_structure for raytracing
  • provided documentation provides a detailed walkthrough of the sample code and the underlying implementation


DirectX 12 Agility SDK 1.716.0-preview Explained

  • the blog post provides an overview of the latest released features in the D3D12 Agility SDK update
  • the SDK introduces Shader hash bypass, reduced alignment requirements for buffers, as well as the recreation of resources at specific virtual addresses
  • as well as shared links to updates to the d3d12 video encoding functionality


[video] Calculating Specular Highlights - Custom Lighting Models - Episode 3

  • continuation of custom lighting model using Unreal and Unity
  • this part covers Phong, Blinn formula, as well as  Half Lambert, diffuse
  • implementation is shown using visual shader language


Agility SDK 1.716.0-preview: New D3D12 Video Encode Features

  • the blog post presents new features for D3D12 video encoding
  • allows developers to provide additional hints (such as motion vectors or dirty rectangles) to speed up the encoder processing
  • also now allows frames to be split into sub-regions with individual progress tracking and data access control


Now Available: Matrix-vector operations using tensor cores in Slang

  • first, experimental support for Cooperative vectors has been added to Slang
  • the blog post explains the features and presents two examples of how to use it
  • additionally provides a link to the SPIR-V extension


nv_cluster_lod_builder

  • Nvidia open-sourced a mesh processing library that allows meshes to be split into meshlets and calculate LOD cluster that allows for continues LODs
  • documentation explains the API, how the implementation approaches the decimation and split steps
  • additionally, explain the limitations (there are a number of them at this time)


Thanks to Angel Ortiz for support of this series.


Would you like to see your name here too? Become a Patreon of this series.