Hollow prims or single?

Hi!
This is probably a very silly question, but which building method is more efficient (in terms of sim load, rather than prim usage)?

(a) A structure made up of 3 prims - each 10m high by wide and linked to form a square which is open on one side.
(b) A structure made of 1 prim (a cube) which has been hollowed and cut so it acts as a structure with 3 solid walls with an opening on one side.

My feeling would be that the latter would be the best, but it just struck me that maybe the sim has to do more work to render a hollow object than 3 linked, flat prims!

Thanks

I believe that 3 separate prims are MUCH easier for the sim to deal with than a hollow prim. Concave objects are a hard problem for the physics engine.

Doesn't linking them cause them to become one object as far as the physics engine's concerned?

I assume these are for walls, so they normally aren't set as physics.

The server doesn't help you render anything on your screen. It sends you the information of what and where the prims are along with all its attributes and your own PC renders this. So, in this case, 1 prim, regardless of its attributes is less of a load than 3 prims, assuming they are all the same texture.

Your PC will need to calculate more as well with 3 prims. In terms of geometry, the hollowed out and cut prim has 20 polygons while the 3 prims has 36 polygons together. This is for rendering only, not physics collisions. This is assuming all the faces are triangulated once.

If they are set to physics, the mesh complexity (number of polygons) has a bigger impact on server load than the number of prims, I think. A very curly twisted torus for example has more polygons to calculate collisions than a few flat boxes linked together. The server calculates all the collisions and then updates you the information of which direction things are moving.

I defer to those with more knowledge of how the prims are rendered, as to server load. Another key thing to keep in mind in your decision between the two methods, is that a hollowed surface will be seen, texture-wise, as one continuous surface. This can present problems (as I learned when taking this approach) if you want, for instance, two of the surfaces to be one texture and the other surface to be a different texture. Also, just in terms of the texture itself, some textures will look odd and very stretched-out when applied to the hollowed-out object.

Hope this helps.