Workflow Query - Many Railclones or Single Railclone for performance?

[OC1963]

Hey guys,

Quick one I hope - Just about to start testing some things to check some load times but in case you or someone already knew the answer as to whether it's faster to have many smaller railclones on individual splines, whether linked by a Master Style or one mega railclone attached to one spline. Also if that matters either with the railclone proxies at all or that now means either way would be the same? Had a few projects where we're building cities in the kilometres recently and making some obscene number of buildings and variations visible in aerials so every little helps.

One other workflow I was trying was using a group per floor plate to then randomise new towers procedurally through logic compositions to create more variations but found across such huge scales this was trying to account for all the objects in each group and startup times and even caching was procing troublesome unless I then bit the bullet and proxied each floorplate as one single object per floor. This may just be Max limitation of hating a large number of objects, was hoping through some dark magic/wishful thinking that the group would somehow be converted to one object in RC calculations.

Cheers

[Michal Karmazín]

Hi,

There is no general rule as every project may require a slightly different workflow. However, in most scenarios, a reasonable balance between these two approaches leads to the best performance. It is not necessary to attach all these splines and manage all calculations in one or a few RC objects. However, having a lot of RC objects may take time to process them.

Using the Proxy Cache feature would be highly recommendable for these purposes. It's very efficient - it'll load and render much faster because RailClone doesn't need to perform so many calculations to rebuild them, which is especially useful for complex styles. In theory, there shouldn't be any significant difference between the Embedded / External mode though the Embedded might be a bit faster.

I hope that helps.

Best regards,