Adaptive family guidelines for performance

[trevor.clarkM88RA]

I'm building some common families for my firm and some might benefit from the more powerful controls of an adaptive family. However, I'm concerned about performance and I hate to run down a dead-end when the families are put into a real-world Revit file of 600mb and 200 sheets. Are there any best practices, recommendations, guidelines, or test results on the performance impact of using adaptive families? For example, is the process required for added adaptive points linear or exponential? Should regular families be nested into the adaptive instead of modelled directly in the adaptive or does it not matter to performance? What is the impact of each instance placed in the model, if there are 100 is this negligible or severe?

[ToanDN]

What are they for?

[trevor.clarkM88RA]

A lot of roof applications, roof crickets, tie-off davits & window washing systems, walk pads, solar panel anchors, etc. Basically families that need to have a consistent application across irregular surfaces of a low slope roof. But it's come up on interior design, such as unusual shaped acoustic wall panel, etc. Situations where we need a consistent function across a few different geometries that are not rectilinear. However, if using adaptable families is going to sink performance of the model verse just using model in place elements or a larger number of standard families, it would be worth the extra model management to avoid them. I was hoping to find some references or user experiences on using adaptive families in big projects before getting into a situation near the end of CD's where the model runs so badly we need to spend more time replacing the adaptive families for something else when the project team is already at peak workload.

[barthbradley]

If you've got to ask about performance, I got to imagine that you are talking A LOT of Adaptive Component Families. Honestly, I think there is a more efficient way to place "rectilinear" components " across irregular surfaces" than using Adaptive Families - which would require picking a minimum of four points for EACH component placed. 

Have you considered Pattern-Based Panels on a Divided Mass Surface?  

[trevor.clarkM88RA]

I must be setting the adaptive families up differently, they typically only take one click per instance placed. I don't think a pattern-based panel on a mass surface would be very effective, few layouts are regular enough for that and the added layer of editing the layout via a hidden mass always creates confusions on the team. I guess, just imagine the layout of roof tie-offs and an associated life safety line - it is always vertical and a fixed height from the roof surface, but is set along the roof relative to key features, not at a regular interval. A roof may have 60 such tie-offs. Using an adaptive family, I can set the base of the tie-off 7" below the roof surface as it would be mounted below the insulation and it can be vertical rather than set perpendicular to the 3/8":12" roof slope. If needed, several can be in a single adaptive family to show a life line with a connecting graphic between. But is using that many adaptive families across many types of roof families (Crickets, roof drains, solar panel mounts, etc) be significantly more detrimental to the model performance than other methods? It doesn't seem like anyone has yet tested how much adaptive families impact Revit performance.

[ToanDN · Accepted solution]

---

@trevor.clarkM88RA wrote:

A lot of roof applications, roof crickets, tie-off davits & window washing systems, walk pads, solar panel anchors, etc. Basically families that need to have a consistent application across irregular surfaces of a low slope roof. But it's come up on interior design, such as unusual shaped acoustic wall panel, etc. Situations where we need a consistent function across a few different geometries that are not rectilinear. However, if using adaptable families is going to sink performance of the model verse just using model in place elements or a larger number of standard families, it would be worth the extra model management to avoid them. I was hoping to find some references or user experiences on using adaptive families in big projects before getting into a situation near the end of CD's where the model runs so badly we need to spend more time replacing the adaptive families for something else when the project team is already at peak workload.

---

I think the reception that adaptive families take long time to process is they are commonly used in repeaters in a large quantity.  It takes horsepower to flex a shear amount of geometry when the host changed.  

It also comes down to the complexity and quality of the families.  I have seen non-adaptive families with a nightmare of constraints and formulas that can fill the Vietnam War Memorial wall that takes minute to flex one instance.  Without knowing how your families are built it is hard to judge.

[barthbradley]

---

@trevor.clarkM88RA wrote:

... just imagine the layout of roof tie-offs and an associated life safety line...

---

Roof tie-offs and an associated life safety line?  Are we talking about modeling fall protection?