Jump to content

Tier reduced 40% -- How are you celebrating?


Qie Niangao
 Share

Recommended Posts

With the new physics types for Mesh, applied (skillfully) to prims, on average 40% less land is needed to support the same number of prims as before.

(If we expected an even greater tier reduction, well, Mesh itself permits builds with even more efficient "Land Impact" -- although there are very, very few Mesh builds fit to look at yet, let alone own, so it will be a while before that effect kicks in.)

Or are you planning to spend your tier decrease on more detailed builds?

Link to comment
Share on other sites

I thought I might use the money to buy - eventually - a better graphics card which supports the more resource hungry mesh-equipped viewers, so that those avatars wearing mesh clothes didn't look ridiculous, with rubber swimring aids around their hips.

Although I might miss all the ladies being topless.

Link to comment
Share on other sites


Perrie Juran wrote:


Qie Niangao wrote:

With the new physics types for Mesh, applied (skillfully) to prims, on average 40% less land is needed to support the same number of prims as before.

 

But are you going to live in 40% less space?

I'm getting claustrophobia here.

Penny and I have blogs on scale and proportion to solve that. :D

 

In seriousness, this can reduce the cost of some prim builds, and raise the cost of others - and it takes skill and trial and error to find which.

 

Link to comment
Share on other sites


Dora Gustafson wrote:


Perrie Juran wrote:

But are you going to live in 40% less space?

I'm getting claustrophobia here.

In that case you are building too big and too close to the property lines and out of avatar proportion:smileyhappy:

That's my experience

Well I don't have a whole SIM but I have  2048 sqm.

On that I have a small house, a yard with a pool, a hot tub and a gazebo.

I have around 400 prims on the property.  My critters are the biggest 'prim hogs.'  I have a a dog and a number of birds that live in my trees.  Not clearly visible in this picture are all my (one prim) flowers.

So I live fairly comfortable.

 

Home Lay Out

Link to comment
Share on other sites

Until I can learn how to make Mesh stuff that doesn't actually have a higher PE than what I can make with prims, it will have very little effect on me at all. For that matter, most of what I have tried to make so far with mesh hasn't worked right at all.

So far, my efforts to make even the most simple mesh items have utterly failed. And I'm not going to trash all my self-built prim stuff and spend tons of money on other people's creations, that most likely won't be what I want anyway.

Link to comment
Share on other sites

Quite the opposite for me on the first point. I'm a newbie to ownling land and only have one 1024 sq. meter plot on the mainland. I actually want to expand into a neighboring plot if the owners decide to close up shop.

I purchased my land not long before Mesh Beta started gathering steam, so I've always had it in the back of my mind when building on the plot. At the moment I'm working to replace my main structure with a mesh version (hopefully for significatly less land impacts with more detail than the prim version).

When I get enough experience, I hope to deck out my plot with as much detail as I can while still leaving enough prims free for my projects.

Link to comment
Share on other sites


Pamela Galli wrote:

Qui, can you give a link to that data on avg. mesh LI compared to sculpts/prims, and any other similar data?  Thanks!

Well, I can't do much better than to link to the various Land Impact (aka "Prim Equivalence") calculations in the wiki, which are still marked as draft but haven't changed since before the Mesh beta hit the main grid. (Here, here, and here, mostly.)

I should be a little more precise about my 40% "average", which wasn't about Mesh, per se, but the savings in land impact from adjusting prims to use the new physics types that were introduced with Mesh.

I had a whole article for the Bay City Post on this, a couple months ago, and I don't want to bore folks with all the details here (as I probably did there).  I'd just start with the caution that the method is best performed in a sandbox, on objects you can readily replace.  It doesn't work with prims that must be linked to Mesh, sculpties, or anything scripted, and only really applies to boxes and cylinders--otherwise you won't get any savings with physics types of "Convex Hull" or "None".  And that's why I estimated only 40% savings: you get two-for-one with almost all unscripted boxes and cylinders that can be linked together, and most parcels contain a vast, vast majority (like 80%) simple prims of that sort.

Amazingly enough, you can link two gigantic megaprim boxes and have it count as only one prim equivalent, but the tiniest sculpty or a single inactive script anywhere in the assembly invalidates the savings for the entire linkset.

Anyway, presumably obviously, I was being a bit tongue-in-cheek with this thread, but not without reason: I want folks to think for a moment about land tier in broader terms than simply square meters.  Land is a place to put content, so when there's a sea change in content, there's also a change in what land is, and in what tier is buying us.

Link to comment
Share on other sites


Pamela Galli wrote:

Thanks Qui -- I would like a link to your article.  The links you provided I will have to study a bit but they contain stuff like this:

 

Image:PhysicsResourceCostEqn.png


One of the great puzzlements to me is this whole impact thing.

My basic undersanding was that Mesh would allow two things:

1.  Nicer builds In World and nicer clothes and other Avatar accessories.

2.  Lower IMPACT.

But if the impact is LOWER why is everyone I have talked to getting a LOWER frame rate with V3 Viewers.  (And I know your experience may vary). 

For many it is only a slight difference.  But the past two nights I have been dancing at two different clubs and the subject of Mesh has come up.  Of the 15 different individuals, only one was using a Mesh viewer.  And the universal comment of the others as to why they weren't, it was simply this, IT WAS SLOWING THEIR COMPUTERS DOWN TOO MUCH.  So much for lower impact.

Maybe I should just start my own thread about this.

Link to comment
Share on other sites

Not really much change for ordinary prims. The physics shape can now be set individually per prim, to either "prim" (which is the default), "convex hull" (which closes all holes and makes it a solid object) and "none" (essentially turning the prim phantom). Objects can have a mix of all types.

Beware though, changing the physics type to anything other than prim "opts in" the whole linkset to mesh accounting, i.e. "Land Impact". It can lessen the prim count, and it can easily skyrocked the prim count.

As an example, rez a box and set it to physics shape "convex hull". You'll find that it still counts as one prim but if you link two such boxes together, it'll STILL be one prim because individually they're 0.5 land impact, which gets rounded up to one.

Now do the same with a torus... preferably in a sandbox or someplace with a bunch of available prims. :)

More or less incomplete:

  • Scripted objects will retain their land impactt or get a higher land impact
  • Sculpts and torii will get drastically higher land impact (especially left at prim physics shape)
  • Changing objects' physics shape impacts sim performance, specifically the physics engine
  • The ability to turn individual prims in a linkset "phantom" by setting their physics shape to "none" makes funky scripting un-neccessary in most cases
  • In a complex linkset it does make a lot of sense to individually fine tune the physics shape. It can often cut the prim count (land impact) in half, even if there's sculpts or torii in there
  • Fine tuning all objects can make a drastic difference in sim performance. A mall I'm involved in went from zero spare time (plus lag) to 2ms spare time (and no lag) after mucking with the physics shape of a few thousand prims, one by one. I now only get called Prim Nazi... figures.
Link to comment
Share on other sites


Qie Niangao wrote:

With the new physics types for Mesh, applied (skillfully) to prims, on average 40% less land is needed to support the same number of prims as before.

(If we expected an even greater tier reduction, well, Mesh itself permits builds with even more efficient "Land Impact" -- although there are very, very few Mesh builds fit to look at yet, let alone own, so it will be a while before
that
effect kicks in.)

Or are you planning to spend your tier decrease on more detailed builds?

I do not see any reduction in tier regardless of prim counts.  Show me reduction in USD and I will show you a happy resident;-)

 

Link to comment
Share on other sites

Simple, just dump 40% of your land and reap the savings in real US$s.  The new rules let you keep more stuff on less land.  That's the point.
Of course, that doesn't work at all for breedables, among other things.  To be fair, though, it's probably fitting that the Lindens now penalize breedables, after all the free promotion they've given them for years.
Link to comment
Share on other sites

Yeah... it can have a very dramatic impact on land impact, but I'm surprised you're seeing any change in spare time at all, just by adjusting physics type.  If that's happening, I'd have to guess that the sim performance was heavily "collision-bound" before the changes.

To take advantage of this, I strongly recommend separating linksets into those prims that benefit into one linkset, and those that don't benefit into another.

Now, one performance-affecting thing this practice does achieve is to identify scripts that just aren't needed.  Static particle displays, for example, and texture animation scripts, are often left behind after their effects have been established.  (There are lots of other prim-property-setting script functions, but these are the ones I see most often.)  Individually, they have very little impact on the sim -- they don't actually do anything, but they introduce a tiny bit of scheduling overhead, so they do eventually add up.

On the other hand... I almost hesitate to mention this, but... because the presence of scripts in a linkset is penalized, there's incentive to use fairly expensive script functions to dynamically link and unlink prims so that scripts can operate on them without having to incur the land impact of keeping them linked all the time, and dynamically adding and removing scripts to make it all work.  I expect this practice will spread and ultimately hurt performance more than any savings that might have been anticipated by penalizing scripts in the first place.

Link to comment
Share on other sites

Wow, thanks for the heads up.  Don't know how this got by me, but I had no idea.  I just gained 36 "land impact points" converting a set of stairs and porches to convex hull in a matter of minutes.  It does not seem to make a difference to legacy viewers either, other than they still see the original prim count, though land impact still shows the savings (obviously since one is client, the other server).  On my way to land impact heaven. Woot!

Link to comment
Share on other sites

Well... savings depend a lot on the actual sim. In this case, most of the vendors there are not scripted, i.e. just boxes set for sale. The savings come from the fact that physics in any shape are a problem for the physics engine. The more solid things there are, the more any physics engine has to calculate. When there's avis around anyway. Running into walls etc.pp.

The important thing is that convex hull is cheaper (for the physics engine) than prim physics. Setting the prim phys shape to "none" takes it entirely out of the equation. Just ask yourself how many prims have a physics shape when they don't need any? In the case of the mall I think I disabled physics on about half of all prims, effectively turning them phantom. I'll grant that I went nuts on it. Roof, windows, door frames... everything you would not normally collide with I turned into physics shape "none". Decorative sculpts same thing, though there i just opted for turning them phantom and not opting them into the LI accounting.

And you're absolutely right about prim property scripts... they utterly suck. I tend to toss them too. Texture animations usually, particle scripts, sometimes prim phantom scripts (the old way of making one prim of a linkset phantom).

Equally a pet peeve are scripted windows... with a script in every freaking window prim. Or lights. Actually... don't even get me started on scripts :)

Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...