In the example you show, the difference is negligible to non-existent. Creating a solid source is so trivial, there's no real benefit in copying vs creating it again.
Should you be using a clip, using the layer router would be faster. Otherwise, the pixels would need to be read twice from disk.
Note that the layer router is only faster when it's routing a layer underneath it in the stack. Should you have your layer router on layer 1, and route in layer 3, you'll get 1 frame delay. This is because when layer 1 needs to render, layer 3 has not rendered yet. In these cases, it will access the previous frame from layer 3.
If your video is running at 30 fps, and Resolume at 60 fps, you'll probably won't notice this, because the frames will be the same anyway for that render pass. But it could become a problem when you have less overhead in your fps count.
Layer routers are a bit weird when it comes to this stuff, so I hope that made some sense