Tutorial Node: Attributes With CUDA Data¶
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.
Installation¶
To use this Node, you must enable omni.graph.tutorials
in the Extension Manager.
Inputs¶
Name 
Type 
Description 
Default 

a 

First value to be added in algorithm 1 
0.0 
b 

Second value to be added in algorithm 1 
0.0 
points 

Points to be moved by algorithm 2 
[] 
multiplier 

Amplitude of the expansion for the input points in algorithm 2 
[1.0, 1.0, 1.0] 
half 

Input of type half for algorithm 3 
1.0 
color 

Input with three doubles as a color for algorithm 3 
[1.0, 0.5, 1.0] 
matrix 

Input with 16 doubles interpreted as a doubleprecision 4d matrix 
[[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]] 
Outputs¶
Name 
Type 
Description 
Default 

sum 

Sum of the two inputs from algorithm 1 

points 

Final positions of points from algorithm 2 

half 

Output of type half for algorithm 3 

color 

Output with three doubles as a color for algorithm 3 

matrix 

Output with 16 doubles interpreted as a doubleprecision 4d matrix 