Jump to content

Question of the Day - Which is Better?


You are about to reply to a thread that has been inactive for 1873 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

Now I know there are some VERY smart folks here and while they each still have their "opinions", they also test a lot.   There has been an ongoing discussion on another forum (I really didn't START the discussion but it has my name on it -- you know how topics can morph) on this question but responses seem to be mostly based on supposition and hearsay and such. So while I really don't want to start another ongoing thread of opinion, by now I would really like to base MY opinion on something like -- well tests and facts. And I am counting on you guys to have that.

Thanks.

 

Question:  Which is easier on the server -- a Mesh Cube (in this case with six materials, mapped as a prim would be - one on each face and full LODs for all levels -- so a replacement for the default cube) or a prim cube? 

Link to post
Share on other sites

I seriously don't think you (or any others apparently) are asking a very relevant question.

The server load of either is minimal. All the server has to do as far as I know is pass on the data to the viewer and both types of boxes only cost a couple of kB.

But what really makes the question irrelevant is this: full LoDs on a box? A mesh box doesn't need any LoD models, you use 6x4=24 verts and 6x2=12 tris, that's a lot less than the 54 verts and 108 tris (off the top off my head) prim box has on the highest LoD.

So replacing a prim box with an identical mesh box would be wasting geometry, quite a lot as you can see.

But to answer the question....just build yourself a mesh box, upload it to SL and see what the server cost is. It's 0.5 I bet :) That without a doubt  means LL thinks the differences between the two are so small, the boxes are of equal load. Overhead will determine the weight, not the actual object.

If you still want a definitive answer, I'm sorry I can't provide that. Only someone who knows the ins and outs of the SL code and server hard- and software can answer this one.

Link to post
Share on other sites

Kwakkelde Kwak's response is accurate.

You're asking about a special case, the cube. That isn't representative. It doesn't much matter. But, apply a twist to the prim cube and the difference becomes significant in favor of a mesh.

Also, how one handles the LoD's makes a big difference in how efficient the mesh is. There is a lot of information on the web about how to make efficient LoD's. It is pretty much the same for all virtual worlds.

For basic prim shapes you can't improve a lot. For complex shapes you CAN produce a significantly lower Land Impact and render cost with mesh. It can also be a disaster with outrageous render cost. It depends on the designer and how good they are.

When the Quick Graphics viewer makes it to main viewer status we will see more pressure for good mesh design. As it is now the casual user of SL has no way to differentiate between poor and good design.

----

Server Side: The difference between basic prims and mesh prims is the amount of data to be downloaded. Prims have their mesh vertex list built into the viewer, so there is little to download, just some parameters. With mesh prims the 4 LoD lists of vertexes has to download. In general these lists are far smaller than a single texture.

Plus, we are now using CDN. So, most of the download is shifted off the Lab's servers.

If you watch the Viewer Stats, which show server stats and compare the server performance for mesh and non-mesh areas, you will find the servers run at near no load values in both cases. It is only when there are lots of scripts and avatars that server performance degrades.

With new HTTP changes to the server and viewer the load on servers is decreasing. Poor clothes design, because of no Land Impact cost, is out of control. So, in crowded areas we see lots of performance hits. Expect that to start to change when the Quick Graphics viewer promotes to main viewer.

Link to post
Share on other sites

I definitely agree about the clothes issue and scripts :D.  And good to know about the new viewer.

 

So from my out on the edge viewpoint (and I don't think the folks arguing will ever agree) it seems like the MAIN reason to use a mesh prim in specialized cases would be so that your building would be 100% mesh -- which for some people is important.

 

I'll keep watching for any other techincal info.

Link to post
Share on other sites
You are about to reply to a thread that has been inactive for 1873 days.

Please take a moment to consider if this thread is worth bumping.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...