Jump to content

PE - Light at the end of the tunnel?


arton Rotaru
 Share

You are about to reply to a thread that has been inactive for 4113 days.

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

Recommended Posts

Sorry, arton, but I have to disappoint you. This was still happening with the upload dialog cost number, but was corrected in the Show Render Info figure some time ago. That is the expected quadratic function now. I will add a comment to make that clear. The Show Info number is agreeing with PE in the edit dialog now and is the figure they seem intent on.

Oh, wait. Runitai has added a comment about LOD array index, but I don't see how that can affect the power of the dependence on radius. Will have to look after the fix to see. The inworld numbers were behaving exactly as predicted by the algorithm in the wiki last time I tested them. The inflection points between the parabolic pieces are the clincher. Also the plateau above 34.66m .... which is amost exactly 40x40x40 for a cube if you want to try it.

Link to comment
Share on other sites

OMG. Now it is all really seriously screwed up. This change is wrong. If you upload a mesh with the same file in all LOD slots, the Show Render Info now goes down as you start to increase from tiny, then turns round and goes up again. (eg 0.1m->10.3, 2.5m->4.1, 10m->16.1). And the inflection points are all messed up. The plateaus are different for the same high LOD with different combinations of low LOD. That "fix" needs to be reverted asap. I will try to documant it properly. There vare some circumstances where it agrees with PE.

The streaming cost in the uploader has also gone even more wrong. It is still givging parts of the graph with cost proportional to radius^4 instead of radius^2, but now it has multiple segments where there's only one LOD step and there  should be only one. Asrrrgggghhhh!

I don't know what changes were made. I can't find them in the getStreamingCost code that came with 224066.

 

 

Link to comment
Share on other sites

I looked at the changes. Nothing dramatic, I thnk. An increase in what is treated as the total area objects will be seen in will stretch the curves on the radius axis, including the inflection points, by about 1.25. It will also reduced the max cost on the vertical axis, but.....

There is a big change in the rational for the vertical scaling. This is supposed to be related directly to the percentage of the region triangle budget (250,000) being equivalent to 15000 prims, rather than the download data size. This reflects the change from streaming to rendering as the major concern, that Runitai mentioned. However, the triangle count used for now is just the bytes of data divided by 16 per triangle. So really it's still the data size (with adjustments for header overhead etc). I guess Runitai said they were highly corellated though, even considering sharp vs smooth edged triangles. So may be that is a better measure of rendering load than an actual triangle count, which would treat sharp and smooth the same.  Anyway, that means a totally new vertical axis (cost/weight) multiplier, so it could be a big change either way!

Must sleep. That's kept me up much too late!

 

Link to comment
Share on other sites

I always thought that linking objects together into a linkset should be encouraged and not penalized. So we are now encouraged to only make linksets of small items ?  It starts getting too much for me:

 

  1. I take a bowling pin (2.23 meters high)... PE=2 (selection streaming cost tells me 1.9)
  2. I create 4 Bowling pins (same height) ... PE=8 (selection streaming cost of the unlinked set: 7.6)
  3. I link the 4 pins to a linkset ... PE=9
  4. Make the pins smaller, until the linkset PE < 8. Now the link set gets cheaper than the 4 isolated pins.

 

I need a break.

Link to comment
Share on other sites

Uhmmm thats strange

 

Yesterday evening i showed a pal of me the SL Meshes on the beta grid.

And also wanted to show him some downsides .. as you mentioned the resizing and also the other strings.

 

I have a small 5x5x5m cube tunnel still there at HQ1 and also my 2 fireplaces.

I grabbed the tunnel and rsized it to 10x10x5m to demonstrate the increasing prim cost and discovered that this, with this increase, was no longer the case Oo

 

Ok .. well .. Uhmm .. lets wait for monday then again and ask then. I dont know what to think now :(

 

Link to comment
Share on other sites


Gaia Clary wrote:

... It starts getting too much for me:...

I need a break.

Yup, same here. I tried to keep up with it since october last year. But now I feel more and more "just get back to sculpts and prims and everything is just fine".

Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 4113 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
 Share

×
×
  • Create New...