Palladio

[biaozeng]

I hope that the plug-in can support compress block to improve the generation efficiency企业微信截图_20230206003428.png

[ChristianKnieling]

Dear biaozeng,

Thank you for reporting your issue!
Palladio is already implemented in such a way that it should maximize CPU utilization even without the use of compiled blocks. Would mind describing in more detail what your use case of compiled blocks is? If this is about pure performance, would you be willing to share the rule file you are working with so we can profile Palladio with your use case?

Best, Chris

[biaozeng]

I hope to use cga in building generation in big cities. I connected 1200 buildings in the test process, and the result was wrong. It was correct to connect only dozens of buildings, so I had to try to use foreach. There are about 400000 buildings in the scene I want to generate, so I need its calculation process to be as fast as possible

[ThomasFuchs]

Thank you for your reply. To give further advice, more information is needed.

• Can you please share more insights into what exactly went wrong?
• Did the model generation work?
• Does the Palladio generate node have maximum CPU utilization on your system?

Concerning your use case:

• Are the buildings procedurally generated or pre-modeled assets?
• Is it possible to distribute the computations to multiple systems?

[ThomasFuchs]

Thanks @biaozeng, for providing the reproduction data. 

The issue is caused by not separating the footprint geometries. Therefore the rule package (rpk) is applied on a single big geometry that fails to fully generate.

To fix this, add an additional attribute "primCls" with the value "$PR" (Primitive Number) before assigning the rpk:
ThomasFuchs_0-1678286478408.png

This will lead to the correct result:
ThomasFuchs_1-1678286610367.png

[biaozeng]

The scope showed different resultsQQ截图20230423094718.jpgQQ截图20230423094347.jpg企业微信截图_20230423094328.png

[chris_knieling]

Thank you for reporting this issue.

Could you please clarify what output you expect and how it differs from the actual result?

[biaozeng]

I use the size of the scope to determine the front and sides of the building. If scope. sx>scope. sz, then x is the front, and if scope. sx<scope. sz, then x is the side. Then, different elevation rules are applied to the front and sides, and this operation is correct in cityegnite. Package this rule into Palladio encountered an error when running in Houdini. Please refer to the comparison screenshot I uploaded earlier