I'm going through the tutorials, When I open an operator it has the @P.x instead of $TX for instance. Sometimes the @P.x works, sometimes it causes problems I can resolve by changing it back to $TX.

Why @P.x now? And why does it cause problems?

Thanks :?

Hi,
The move is on right now to the multi-threaded vex language $TX , $ is not multi-threaded. Vex is the way to go

Rob

The @P.x in place of $TX was done in H14 for consistency with wrangling which has become an often used method to wank attributes about efficiently.

@P.x is the same whether you are in a wrangler parameter or a parameter field for a SOP geometry node's parameter field.

$TX is called a Local Variable and the issue is that many SOP Geometry Nodes have their own unique local variables you would have to reference from the Help Docs for that SOP.

By also allowing @P.x as a local variable to fetch the attribute by name and index, it makes for a more consistent path to learn attribute wrangling in Houdini moving forward.


If you can post the specific cases where @P.x fails, that would be great.


Rob is also correct in that moving to more modern methods using wrangling and VOPs is preferred for performance reasons.

Absolutely learn with the tried and true SOP Geometry nodes from older tutorials as those methods are still working. When performance is in order, and you get more comfortable with Houdini, you can replace the slower parts with VOPs and wranglers.


There are so many ways to create the same things in Houdini and each has it's pros and cons. Just plow away and if you get the results you like, you win. Then and only then do you start performance tuning your network using the Performance Monitor to pick out the slower operations in your network, if it's even necessary.

Cheers guy's that helpful to know.