Nyx Linden has created a new tool that allows one to examine render complexity. When the tool is activated it colors objects in the area Blue, Green, or Red for low, moderate, and high complexity. The lower the complexity the faster things render.
This tool is based on the newest render complexity algorithm. According to Nyx the algorithm is not yet built into the Mesh Project Viewer. So, we may see differences in the values given in the mesh upload dialog and the Render Complexity Tool.
Nyx says there are a number of caveats one should consider and remember. So, we can ignore and forget those… er… wait… may be some of us could be smart this time around and remember this is all beta stuff and it will change, which is what Nyx is saying. So, don’t plan your tier savings based on the values you get from this tool.
Meshes, prims with animated textures, texture resolution, number of triangles, prims with media textures, prims with full bright faces, and some other things Nyx may have forgotten are likely to change value in future releases.
The Complexity Tool is alpha level in a beta version viewer. If you think it is going to work perfectly, please check to see if the words ‘silly rabbit’ are written on your forehead.
There are several setting you can control via Debug Settings.
RenderComplexityColorMax – color of the most expensive objects (default: red)
RenderComplexityColorMin – color of the least expensive objects (default:blue)
RenderComplexityColorMid – color of a mid-range complex object (default: green)
RenderComplexityStaticMax – set a static value for what render complexity equates to “max”. Any objects at or over this value will have a highlight of the color RenderComplexityColorMax. If this value is -1, the scale will be recalculated each frame to color the objects based on the current view. (default: -1)
RenderComplexityThreshold: objects less than this threshold will not be colored. This does not have any impact on the color scale’s minimum.
If you run into problems with the tool, contact Nyx Linden. Ask questions in the Mesh section of the forum.
Getting the Tool
To get a copy of the tool one must use the Test Viewer. The link is to the download page. The viewer is a version of SLV 2.6.2-225046 Mesh Project. Once things progress the feature will be part of the Mesh Project viewer.
The control to turn the tool on is: Top Menu->Develop->Render Metadata->Render Complexity see the image.
The settings are above.
The install places the viewer in its own folder: SecondLifeDeveloper. The viewer opens in Basic Mode… annoying that a developer’s viewer starts in BASIC…
The second restart put me in slow motion mode. The viewer even before login was taking 5 to 10 seconds to respond to the mouse. A reboot of the computer and I was ok and running.
I expected the viewer to take me into ADITI by default. It did not. So, one must select ADITI if that is where you are headed.
The image shows the tool in action. Most of the world is red… Notice the numbers. One can use the Debug Settings to tweak the color ranges to make the tool more useful.
The viewer itself is medium performance, 25 FPS in my little cottage.
The tool is incomplete and sort of a first draft. But, I think it is going to be helpful for content creators.
This tool will appear in viewers once the render cost algorithms for mesh are finalized. Documentation is in progress now.
WARNING: On the main grid I have not been able to use Group Chat. I can listen but, not chat. I get a connect fail.
Pingback: Popular SL News During March 2011 « Daniel Voyager's Blog
If I remember correctly, Nyx made this when mesh was in beta. Does this not just work off the Display Cost? With avatars, we have something similar with ARC or whatever it is called now. Every items has a display cost, and I was just saying on the forums that Display cost is probably the most relevant data point to control your environmental lag on a sim. The Display cost should have a spot on the marketplace to let consumers know approximately what it cost to display an item on their land.
I think this was an attempt by Nyx to provide insight to the math being used by the system. It made the math apparent without having to dig through the code. Values may have changed since it was written.
Display cost is client side and the effect on one’s FPS is highly dependent on on an individuals computer and settings. Attempting to put any value of render cost in the MP would just create confusion and drama as most users would see a different value on their system.
Display cost is a calculation of the object and its various properties. It does not change per computer. I run 3 different computers, of 3 different qualities, to test things on. None of them will ever show a different Display cost. Yes, each computer is affected differently by that display cost, but the number will not change, as it is a calculation of the objects properties.
If some1 were thinking of designing a whole region, the display cost would be a very good benchmark to make sure to limit the amount of data that every1’s computer would be downloading. This is the whole point of the display cost. The reason people don’t look at it, is that they either don’t know about it, or they make their items so inefficiently that they choose to ignore it, and then wonder why their SL lags.
IMHO, the display cost should be as important as Land Impact, and actually more important, as LI is not a measure of total data and has nothing at all to do with textures, shinies, bumpmapping, or glow. These areas can be massively abused if you are just looking at Li.
Yeah, you are correct. The viewer uses a fixed calculation. I’m wrong about people not seeing the same number in their viewer. They will. So, it could be used.
Oh, sorry Nalates. For some reason, I’m getting your old blog posts from over a year ago sent to my email. I have no idea why.
Some feeds will pick up any change or revision to an older post. I just revised a two+ year old post because of a broken link. So, you may get an email on that change.
I have not looked to see how that problem may be handled. But, I know other bloggers that have commented on the problem, but not provided a solution.