OmniGraph Node omni.graph.tutorials.CudaData

omni.graph.tutorials.CudaData Properties

Name	Value
Version	1
Extension	omni.graph.tutorials
Icon	ogn/icons/omni.graph.tutorials.CudaData.svg
Has State?	False
Implementation Language	C++
Default Memory Type	cuda
Generated Code Exclusions	None
uiName	Tutorial Node: Attributes With CUDA Data
__memoryType	cuda
__categories	tutorials
Generated Class Name	OgnTutorialCudaDataDatabase
Python Module	omni.graph.tutorials

omni.graph.tutorials.CudaData Description

This is a tutorial node. It performs different functions on the GPU to illustrate different types of data access. The first adds inputs ‘a’ and ‘b’ to yield output ‘sum’, all of which are on the GPU. The second is a sample expansion deformation that multiplies every point on a set of input points, stored on the GPU, by a constant value, stored on the CPU, to yield a set of output points, also on the GPU. The third is an assortment of different data types illustrating how different data is passed to the GPU. This particular node uses CUDA for its GPU computations, as indicated in the memory type value. Normal use case for GPU compute is large amounts of data. For testing purposes this node only handles a very small amount but the principle is the same.

omni.graph.tutorials.CudaData Inputs

Name	Type	Default	Required?	Descripton
inputs:a	float	0.0	Y	First value to be added in algorithm 1
	__default	0.0
inputs:b	float	0.0	Y	Second value to be added in algorithm 1
	__default	0.0
inputs:color	colord[3]	[1.0, 0.5, 1.0]	Y	Input with three doubles as a color for algorithm 3
	__default	[1.0, 0.5, 1.0]
inputs:half	half	1.0	Y	Input of type half for algorithm 3
	__default	1.0
inputs:matrix	matrixd[4]	[[1.0, 0.0, 0.0, 0.0], [0.0, 1.0, 0.0, 0.0], [0.0, 0.0, 1.0, 0.0], [0.0, 0.0, 0.0, 1.0]]	Y	Input with 16 doubles interpreted as a double-precision 4d matrix
	__default	[[1.0, 0.0, 0.0, 0.0], [0.0, 1.0, 0.0, 0.0], [0.0, 0.0, 1.0, 0.0], [0.0, 0.0, 0.0, 1.0]]
inputs:multiplier	float[3]	[1.0, 1.0, 1.0]	Y	Amplitude of the expansion for the input points in algorithm 2
	__memoryType	cpu
	__default	[1.0, 1.0, 1.0]
inputs:points	float[3][]	[]	Y	Points to be moved by algorithm 2
	__default	[]

omni.graph.tutorials.CudaData Outputs

Name	Type	Default	Required?	Descripton
outputs:color	colord[3]	None	Y	Output with three doubles as a color for algorithm 3
outputs:half	half	None	Y	Output of type half for algorithm 3
outputs:matrix	matrixd[4]	None	Y	Output with 16 doubles interpreted as a double-precision 4d matrix
outputs:points	float[3][]	None	Y	Final positions of points from algorithm 2
outputs:sum	float	None	Y	Sum of the two inputs from algorithm 1