Changelog#

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.5.0] - 2024-08-29#

Fixed#

  • Fixed dirty Usd stage after the creation of a post render graph.

[0.4.5] - 2024-05-22#

Fixed#

  • Ensure the RenderGraph is cached correctly for each Render Thread, in the pre and post render execution graph.

[0.4.4] - 2024-03-25#

Fixed#

  • Fixed crash when setting the graph state variables after a fabric snapshot is applied on the farbric stage.

[0.4.3] - 2024-03-07#

Fixed#

  • Fixed crash when building graphs where nodes have attributes with unknown types.

[0.4.2] - 2024-02-28#

Changed#

  • Move the initialization of the state variables to a pre process task.

[0.4.1] - 2024-02-26#

Changed#

  • Initialize the state variables during the build stage, to avoid spike at runtime and deadlock.

[0.4.0] - 2024-02-14#

Added#

  • The CudaInteropExecutor now executes CUDA interop nodes both on the render thread and as cuda commands.

  • The execution framework sets graph variables with the cudaStream_t and other execution state necessary for post render graph nodes

Changed#

  • Changed the ComputeParamsBuilder methods to return a reference, to allow multiple separate add* calls on the same instance.

[0.3.2] - 2024-01-02#

Changed#

  • Updated the description of the extension in the readme file.

[0.3.1] - 2023-12-13#

Fixed#

  • Skip running the pre / post render graph when the state is invalid.

[0.3.0] - 2023-11-22#

Added#

  • Added support for compound nodes in pre and post render graphs.

[0.2.5] - 2023-11-16#

Added#

  • Added private interfaces for execution nodes CudaInteropEntryNodeDef and RenderProductInteropEntryNodeDef

Fixed#

  • Fixed bad cast to check for CudaInteropEntryNodeDef nodes in the CudaInteropExecutor

[0.2.4] - 2023-10-18#

Changed#

  • Using the added ‘canEvaluate’ ABI from core

[0.2.3] - 2023-10-10#

Changed#

  • The RenderGraphState is now stored in thread local storage to avoid data races in MGPU setups

  • Assert when attempting to update the PreRenderGraphDef in parallel from separate threads.

[0.2.2] - 2023-08-31#

Fixed#

  • Attach execution definition to the pre-render and post-render stage update root nodes.

[0.2.1] - 2023-08-25#

Fixed#

  • Ensure the runtime data is initialized before executing post render graphs.

[0.2.0] - 2023-08-17#

Added#

  • Introduced the IPreRenderGraph and IPostRenderGraph interfaces.

  • Implemented the execution graph for pre and post render graphs.

[0.1.1] - 2023-07-14#

Changed#

  • Improved the documentation of ComputeParams and ComputeParamsBuilder types.

[0.1.0] - 2023-05-12#

Initial Version#

  • Initial version introduces the ComputeParamsBuilder and ComputeParams types.

On this page