Distributed Simulation and Layer Cache

In Miarmy 3.5, we can distribute simulate our agents based on the placement node.

This means we’re able to simulate a part of agents in scene and make them driven by cache. And then simulate a second part of agents. The second part of agent can interactive the cached first part of agents as well.

In production, the crowd artist can simulate some agents and get approve. Then continue simulate the other parts

Workflow

Suppose there are 3 parts of agents in your scene and they’re placed from 3 placement nodes, let’s cache them one by one group.

3 placement node and 3 groups

Please make sure turn off agent cache, and make sure fill the "cache folder" and "cache name".

Cache Setup

In the button of Miarmy tool, it will display the list of all placement nodes in scene.

The placement node will be listed in the Miarmy Cache Tool tab

We need to place all the agents from the place node which we want to cache. We can place them from selection or all the scene.

Place all or one group of agents

Place from group

If we're not sure the place node name, we can check the placement node name when you turn on shape node name display.

show place node

And then let's cache it from Miarmy Tool Layer Cache window. We need to select the place node name and them click "Make Cache" 

select the cache place name

cache confirmation

Then all of the agents placed from this placement node will be cached. And the cache file will be write into the sub folder of cache folder:

Agent place cache

After caching, we can enable/disable it in Layer Cache list.

When we hit the "Enable", there going to be a question, it will have you choose mode

Here we just introduce the "Full Drive" mode and which can drive your agents fully, and your agents will be controlled by cache solely.

Once the layer cache enabled, the place node name will start with an asterisk (" == ")

the walking agents in above picture shows the cache driven agents

And of course we can cache other parts of agents with the same method. 

Cache the agents from McdPlace0Shape

Enable cache after the second layer cache making, please notice the below picture:

The left and right side agents are driven by cache, middle agents are simulation agents. And the simulation agents can continue interactive with the cached agents.