When analyzing a 3D model, the implied gate size of the model is of critical significance when predicting injection pressure for a part.
The gate contact diameter is used by the solver to set the appropriate viscosity of the tetrahedron nodes at/near the injection point. All nodes inside the gate diameter are treated as injection nodes and will have the same temperature, pressure, and shrinkage.
If you want to compare 3D and Dual Domain pressure results, a runner system is required.
In order to achieve accurate injection pressures for 3D Fill+Pack results from a modeled gate, model the gate as a tab connected to the cavity. Then you can define the whole face of the tab as the injection location, and regardless of the mesh size, you will have the same area for the injection gate. As an alternative, you can pick more of the surrounding nodes in the fine mesh, and so make a gate that has approximately the same size as the gate in the coarse model.
With the ability to use beam runner feed systems with a 3D cavity model, there is a second option available that allows you to compare pressures in 3D Fill+Pack against those in Dual Domain Fill+Pack. If you model a short beam runner system (even just the gate will do) for both 3D and Dual Domain, and then pick the end node as the injection node, then the gate size in the two analyses will match.