Sign in to follow this  
Followers 0
Gaia Clary

What is "PE Weight" ?

36 posts in this topic

And what is [COUNT] ?

I found this in the newest viewer build "Second Life 2.7.4 (233413)":

peweight.png

1 person likes this

Share this post


Link to post
Share on other sites

I hope this is a work-in-progress. We need to see the physics ans streaming cost so we know which is limiting. Oh yes, and server cost if they don't do the sensible thing and abolish it.

Share this post


Link to post
Share on other sites

Some linked sets of mesh are showing with half the cost in "PE weight" that bthey show in Show Render Info streaming cost.. That would be nice, but it's probably a bug.

 

Share this post


Link to post
Share on other sites

This is about weight/mass in lindograms?

This is a neat info feature I imagine will be helpful for physics stuff, if it is physics related...otherwise I hate it and wish for this feature to be banished and shunned...yes, even called a doodie head and have a 'kick me' sign put on it's back..Ok, OK...just a joke. 

/me scowls at the new top sim info bar area and mumbles "good for nuthin', no vector info pile of no good junk."

Share this post


Link to post
Share on other sites

Hi, it is just buggy. When we see a string like [COUNT] in the interface, it means that there is a mismatch between the XML interface definitions and what the viewer C++ code fills in.

There is no patch on the build result and the JIRA for that change is private, but I would guess that maybe "PE" is "physics engine" and that could be the physics cost that used to appear after the slash.

Share this post


Link to post
Share on other sites

I can confirm that tyhe "PE weight" (PEwt) in viewer 233413 is the number subtracted from the parcel prim capacity. It is the higher of (a number that depends on physics shape type, presumably the) physics cost and the streaming cost in [show Render Info] (SRI) for a simple mesh.

For linked meshes I have looked at, it is much lower than the SRI streaming cost. For example, for my gallery which is 43 prims, the PEwt is 59.0, while the SRI streaming cost is 242.2. It is 59 that gets charged to the parcel prim allowance. However, I am sorry to say this discrepancy is a bug, because unlinking the gallery and leaving the pieces selected give PEwt of 256. and relinking them gives 251.

PEwt does not appear to have the "server cost" in it, as adding a script to a 12 triangle cube leaves the PEwt at 2.0.

Share this post


Link to post
Share on other sites

I have created 2 related jira's for enhancing the SL Object editor and give more detailed and less confusing information:

CTS-611: Primcost display could be made more intuitive ...
CTS-651: Add detail information about the meshes into the SL-Object editor)

That maybe a good location to add ideas or comment on the implementation ?

Share this post


Link to post
Share on other sites


Drongle McMahon wrote:

I can confirm that tyhe "PE weight" (PEwt) in viewer 233413 is the number subtracted from the parcel prim capacity. It is the higher of (a number that depends on physics shape type, presumably the) physics cost and the streaming cost in [show Render Info] (SRI) for a simple mesh.

So then PE is equivalent for "costs"... and the meaning is:

"1 Object" made out of  "1 Prim" will cost me "4 Prims" from my rezz-budget ...

Then why not name it as it is:

"1 object" made out of "1 mesh" will cost me "4 Prims"

But maybe "costs" is a word that has to be avoided these days...

Share this post


Link to post
Share on other sites

I agree entirely. The hiden double meaning of "prim" is badly confusing. It should say "i object" instead of "1 prim". I think  "1 mersh" would then be wrong for non-mesh objects. The meanings of "prim" are (a) standard or sculpted object, and (b) unit of cost charged against the parcel content allowance. Both are rather deeeply embedded in the UI and documentation, and it didn't matter as long as they were always the same. PE is only used if the (linked) object has at least one mesh; call it a "meshy" object. So I guess that is how they hope to avoid confusion.

Oh dear, it's worse than that....

Two separate objects, a mesh door and a standard box....

    door - "1 objects (1 prims) selected: PE weight 3.0."

    box - "1 objects (1 prims) selected."

    both - "2 objects (2 prims) selected: PE weifgt 4.0."

Now link with box root...

    linked - "1 object (2 prims) selected: PE weight 5.0."

    edit linked, door - "0 objects (1 prims) selected: PE weight 5.0."

    edit linked, box - "0 objects (1 prims) selected: PE weight 5.0."

So: (a) plural for one object/prim (b) still calling a mesh a prim (in the parentheses) whether it's alone or linked in a meshy object © zero objects (d) gives no information on cost of components of linkset (e) it still has the ridiculous effect of adding a prim (server cos) for the linking.

 

 

Share this post


Link to post
Share on other sites

We've actually had quite a few discussions on terms and vocabulary.

First of all, PE weight is indeed "Prim Equivalent" weight. It is the number of "prims" of land capacity used by the object and should be the highest of the three metrics we measure content on.

We are avoiding labeling this value as "cost" as that gets too closely conflated with the "fee" charged on upload.

The term "object" has generally referred to a linkset in our UI and documentation, which is distinct from a "prim" which is the smallest divisible unit in a linkset. The definition of "prim" is being extended to include mesh prims, just as sculpted prims are still referred to as prims.

Builders should be able to enable a more comprehensive view (showing physics weight and render weight) by enabling the debug option "ShowAdvancedBuilderOptions".

 

it is being displayed as zero objects because you do not have a full object/linkset selected. I can see how this could be confusing. Suggestions on what to show when a partial object is selected appreciated!

 

This is the first checkin of the new layout, I'm not surprised there is some confusion and probably some bugs. Definitely keep pointing out where the holes are so we can patch them up before release! Thanks!

 

 -Nyx

 

 

Share this post


Link to post
Share on other sites

"Suggestions on what to show when a partial object is selected appreciated!"

"0 objects" -> "1 prim" would fit with your extended definition of "prim" (or "n prims" if n are selected).

Share this post


Link to post
Share on other sites

 


Nyx Linden wrote:

We've actually had quite a few discussions on terms and vocabulary.

...

We are avoiding labeling this value as "cost" as that gets too closely conflated with the "fee" charged on upload.

Then would it make sense to rename "Resource Costs" to "Prim Equivalent" in the importer ? Or is resource cost again something different from PE ?

And out of curiosity: Is the Resource cost which is now shown in the debug display anywhere near to what we will see later in the PE ?

Share this post


Link to post
Share on other sites

@Nyx: Now we see that there are three costs associated with mesh: streaming, physics and server. If the server cost is less than the streaming cost, will the final PE equal the streaming cost?

Share this post


Link to post
Share on other sites

If the streaming weight is the highest weight of the three, the prim equivelancy will be equal to the streaming weight. In that case the server cost will be irrelevant for the currently implemented system.

Share this post


Link to post
Share on other sites

Btw. Mesh Experimental Regions are updated. Now PE is the same what is shown in the Show Render Info display.

Share this post


Link to post
Share on other sites

Im still confused.

The advanced Option reveals the following with one of my meshes

 

Object1.JPG

 

Display weight = 35 PE is 5

Now i wanted to see what one of my other meshes say about this

 

Object2.JPG

 

Display weight 45 therefore 10 points more then the previous mesh.

But the PE is only 2.0 ?

If i use the Render Info from the advanced menue it says the first mesh has a streaming cost of 5.4

And the second has a streaming cost of 1.5

That fits the PE, but the difference in the display weight is somewhat confusing ?

It cant be the physics .. so maybe i get something totally wrong here ?

 

EDIT:

Forgot to mention im using

Second Life 2.7.5 (233657) Jun 22 2011 14:28:19 (Project Viewer - Mesh)
Release Notes

You are at 206,030.0, 178,884.0, 23.1 in MeshHQ 1 located at sim7007.aditi.lindenlab.com (216.82.40.77:12035)
Mesh Import 11.06.16.233117

Share this post


Link to post
Share on other sites

i was wondering all the time what "Display Weight" is and if it is not just the "rounded (Resource costs * 10)" ? So what do we have now ? And how is it related to each other:

 

  • Resource Cost (RC)
  • Physics Weight (PW)
  • Display Weight (DW)
  • Prim Equivalent Weight (PEW or PE?)

 

Share this post


Link to post
Share on other sites

Oh ...

Now that i think about it.. i was assuming that display weight was the new value for the geometry cost.

Maybe thats where i go wrong.

 

I did a quick experiment.

If i copy the second mesh with 45 display weight and link them together, the display weight change to 110 :matte-motes-sour:

Also the PE changes to 3 ( from 2 in an unlinked state ) and accordingly to the Streaming cost ( Enabled from the debug menue ) that says 3.1 it make sense, somehow.

If i unlink them and look at them individually the streaming costs are 1.5 for each mesh. So together 3.1 ( maybe a rounding issue ) is what i can understand if we set the PE cost = Streaming cost. But still i cant figure out how the display weight comes into play, and what it means. So .. if you want to get the best out of your PE for a given mesh, the streaming cost value is ( to my understanding ) still the most reliable source of information.

Beside that it seems we still cant go below 2 Prims

A testcube with 12 Triangles and 24 Verts as high LOD and autogenerated lower LODs and no physics still shows as streaming cost 0.3 and physics cost = 0.4. And still also shows a PE of 2 while Display weight is 20

 Thats for a cube with the dimensions 0.5 x 0.5 x 0.5

If i expand the cube to 10 x 10 x 10 the Streaming cost changes to 0.4 Pe is still 2 and Display weight changes to 50

So i get the impression that Display weight has something to do with the pure dimensions of the mesh.

 

Share this post


Link to post
Share on other sites

I think that minumum of 2 is the server weight. One for the prim and one for the linkset (although it's only one object). If that's higher than streaming or physics cost, then it becomes the PE. So an isolated mesh object can never be less than 2. As part of a larger linkset, it can cost less.

Share this post


Link to post
Share on other sites

Thanks Drongle

 

I still think thats a BIG mistake from LL side. but i guess we never going to change their mind, eh?

 

Share this post


Link to post
Share on other sites

"Display weight" is really just a relabeld "render cost". That's our new term for it and its currently displayed for informational purposes only. Before we can start to limit display weight, we will need to switch to the new algorithm, get the new algorithm on the simulator, and do a fair bit of tweaking so that the numbers are scaled appropriately. Currently "display weight" should just be the amount your ARC will increase if you attach the object. If this is not the case, its a bug, and let me know about it.

The minimum for a single isolated mesh is indeed the server weight. If you look at the documentation you can see we charge 1 prim for the object as a whole and 0.5 prims for each piece. So an isolated mesh object will cost 2 if its streaming cost is low, but the per-prim cost is only 0.5. The values are rounded to the nearest prim.

1 person likes this

Share this post


Link to post
Share on other sites

@Nyx: Thanks for clearing up the display weight confusion. But, I do have a question. I believe that we can see the streaming cost under "Develop -> Show Info", and we can see the physics cost in the edit window. But, where do we see the server cost?

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  
Followers 0