Optimizing performance of large assembly loading times

marcus_wong438

Hi, I had a question about the best way to manage large assemblies such that loading times are bearable. The current way my team's projects are set up is by having a few sub-assemblies in folders of the document, while there is a master assembly of the complete product that uses these sub-assemblies. All of this is a part of a single document.
I am unsure if this rumor is true, but I heard from someone that an individual Onshape document is allocated a single worker thread from Onshape's servers. If this is the case, would it be beneficial to have sub-assemblies in separate documents while the main document derives these assemblies into a master assembly? Therefore, when a change is made, only the document of the subassembly needs to do a complete regeneration instead of regenerating the entire large assembly that contains hundreds of parts.
If this is not true however, what would be the most optimal way to organize large projects such that documents can be loaded in faster for weaker client computers?

nick_papageorge073

It makes a big difference in performance in my experience. 

Besides the performance, it’s better for working on a project with others. If all in the same document and making a part studio change, the asm is affected immediately. Another person working in another part studio within that document might want a stable asm for their area of responsibility. 

ilya_baran

The "rumor" is false.  The threading model in Onshape is way too involved to go into here, but at any given time, we almost always have many threads (across multiple AWS instances) working on a given open document, or even a given tab within a document.

That said, it is a best practice to break up a large design into multiple documents, for all sorts of other reasons (some reasons are more "obvious" like not having to update all referencing tabs on every change, the way we do if the tabs are in the same workspace, and some reasons are more "internal").