The 3D Solver takes any closed static mesh as an input, and outputs an animated triangular mesh.
Optionally, it can also take in a collider, a set of guide curves, and a vector field. (more on those later)

For a primer on how the solvers work, check out: General Concepts
If you haven’t already, take a look at that page first, as this guide will assume you’re already familiar with those concepts.
By default, if you put down the solver and start the simulation, it will only inflate your mesh, trying (not too hard) to avoid self-intersections. This is not particularly interesting.
We’ll now go over each of the tabs exposed on the solver one by one. These are generally meant to be used from left to right, as that’s mostly the order they’re executed in inside the solver.

This interface exposes all of the main parameters you need to create a basic (and even not so basic) simulation, however you can access more customization options by double clicking on the node to get access to the inner network:

Here you can add the rest of the GGro SOPs, or any other Houdini node you may need. Keep in mind that these are also executed in order, so anything you add to the left will affect nodes to the right, but not viceversa. These nodes are applied after all of the options present in their respective tabs. So for example, any node you insert in the direction step, will be applied after every other force exposed on the actual solver.
This page will cover the basic info you need in order to get started, while the other pages in this 3D Solver section will go more in depth on the exposed parameters, as well as the dedicated SOPs you can use inside the dive target network.
The first thing you’re going to want to do before you start simulating anything is set your Displacement Global Scale and your Remesh Target Resolution. The displacement scale is the maximum distance any point can travel in one step. The remesh target resolution is the target max edge length of the output mesh.
By default, these are linked at a value of 0.02, which works well for small scale simulations. You’ll need to set these appropriately for your specific scene before starting.