Lots of people are saying the new Firestorm Viewer 5.0.1 is faster than previous versions. My personal experience is it is. But, what I really want to know is it faster than the default Linden Viewer 5.0.0. Here I find the results a bit mixed.
The image shows my front porch where I did some of the testing. In the image starting on the left I have 3 Firestorm and then 3 Linden viewer numbers. The first FPS from ground level at my house is labeled FS 5.0.1 and shows an FPS of 78.3. The number was bouncing between 75 and 84 FPS.
I set the graphics settings so they are similar. Draw Distance is turned down so it is all with in the region. I am looking north while standing at the southern edge of the region. My house is on the south edge of the region. I waited for the readings to stabilize.
With the Linden viewer standing in the same place I get 60.3 FPS and bounce is 56 to 64 FPS. The settings are labeled SL 5.0.0.
I have a sky box at 1,500 meters that I use for green screen stuff. So, it is empty. It is made of 3 prims. With Firestorm, I get 161.8 FPS and Linden 60.6 FPS. Yeah, that is a 100 FPS advantage to FS.
Next I flew over to a landing hub with 26 avatars present. With Firestorm, I get 46.9 FPS and with Linden 50.7 FPS. There is some bounce. But, I spent some time to get a sense of what the average was.
In both cases I have ACI set at 350k. I wasn’t seeing any Jelly Dolls. So, I’ll say that isn’t a factor.
These are preliminary measurements. It will take more time for me to decide if Firestorm is enough faster for me to say it is ‘usefully’ faster. For me being faster in an empty area and slower in a crowd is no big advantage. The other way around would be better.
The ground level test standing on my porch shows Firestorm with 130% of the performance of the Linden viewer. That is useful for me. When I go shopping it should be a nice plus.
There is only a small disadvantage for Firestorm in a crowd. (8% less) If that holds up in numerous places with a variety of avatars, it isn’t going to be enough to turn me off. Both Firestorm and Linden viewers are going to be about the same at events. I probably wouldn’t relog to take advantage of the performance boost.
In a broad sense, it does appear to be faster.
Is it more stable? I don’t yet know. That is going to take some time to experience and decide.
But what I really would like to know: How does it compare to Singularity?
Like a figher jet compares to the Wright Brothers plane: Singularity is so much behind that it literally is the butt end of all viewer jokes. Wich is a shame. Singu is still my preferred viewer but it doesnt even ahve VMM. The latest version also crashes way more often.
It’s not mixed for me as a Mac user. The firestorm viewer is currently double the FPS of the official viewer. it is SO visually noticeable that having to log into the official viewer feels like switching to a decades old slower computer. But some of this might just be related to my account due to some bizarre corruption that only effects the official viewer.
60 FPS? I think the LL Viewer is capping due to Vsync. I’m not entirely sure anymore if its possible to disable (weither the Viewer accepts the Debug or not) but you should look into the Debug Settings and disable Vsync, then restart and do the test again. Comparing unlocked 100+ FPS against capped 60 FPS seems unfair.
I agree it is probably a setting that is different between the 2 viewers. maybe Frame Limiting is enabled to 60 FPS?
If you want accurate fps results don’t use the viewer stats but use something like FRAPS etc since having a menu open uses fps and also the fps depends on where you have your mouse cursor. For example if you move the cursor to the blue border on the top of the viewer the fps will go up I expect because it’s not updating anything.
Opening the Stats panel does slow the viewer. FRAPS itself slows the computer and viewer. But, I haven’t seen Viewer Stats and FRAPS disagree.