 |
|
|
|
OpenExec supports efficient evaluation of computational behaviors encoded in USD scenes. In brief, this works as follows: Schemas publish computations, atoms of computational behavior that take inputs from other computations, and from values authored in the scene, to produce output values. OpenExec ingests a UsdStage, along with registered computations to produce an internal representation of the instantiated computations in the scene. Clients can then request sets of computation outputs to be computed. In addition to evaluation that results in the requested values, OpenExec supports caching of computed values and invalidation of caches in response to scene changes.
This document describes the concepts on which OpenExec is built, with links to detailed documentation for the APIs that implement these concepts, and provides tutorials that demonstrate the details of how OpenExec features are used in practice.
Computations are the basic unit of computational work in OpenExec. A computation definition is a template for producing data flow nodes that take inputs from other computations and produce an output value.
Each computation is associated with a scene object that hosts the computation, referred to as the computation provider. The provider object, either a UsdPrim or a UsdAttribute, anchors the computation in a USD scene, providing a reference point for the computation to use for drawing input values. For example, an attribute computation may take input from the authored value of the attribute that provides the computation.
Computations that are associated with schemas are called plugin computations; OpenExec will load the plugin that defines computations for a given schema the first time any computation is requested for a prim that uses that schema. Plugin computations are registered using an embedded computation definition language, which allows schema authors to register the callback that implements the computation, along with a specification of the input values required by the computation.
OpenExec provides a fixed set of builtin computations, computations that are always provided by a given type of provider object. For example, attributes provide a computeValue builtin computation that yields the value of the attribute at the current evaluation time.
See the builtin computation documentation for a desciption of the builtin computations that are available.
Coming Soon:
1.9.6