Download

Some tests I did with poly-count

Hey guys, I did some tests regarding the polycount threshold of the engine. Basically I created 3 cubes, each with different subdivisions and inserted on the engine for tests.

Here’s the first cube, along with the stats from the engine.

The second cube, much higher poly count. I found strange that the stats haven’t changed.
49bbd7c217c240ca2680b94e74d9aff09cf2654d.jpeg

The third cube is almost 2 million polys and still the stats are the same.

I did some further testing by making 16 copies of the cubes:


ad6088506723d4a964b0a8520c869f1d0de9afc4.jpeg
efe140c109f91db057ecbc6ca05222f02675deea.jpeg

Here the stats changed. Maybe because the engine’s poly threshold is very high.
Feel free to contribute with further knowledge on this subject :slight_smile:

From what I’ve seen other developers on the forums, AnswerHub, and in general say about poly count and the engine is that this isn’t going to be as significant a hit as one would think.

Now when you start using dynamic lights and add materials and a whole lot of these you’ll start to see some issues.

Especially with any dynamic lights on those meshes casting dynamic shadows! Once dynamic shadows are applied it’s then taking into account all of those meshes and their poly counts for calculation.

It’s still good practice to use as low a poly count that gets the detail you need. Using a good LOD setup will work wonders as well.

While the poly count may not affect your scene here, when you start adding in particle effects, sounds, lights, and all your materials you’ll be wanting to save FPS where ever you can and having a more optimized mesh will help with this!

Keep your tests going! I’m curious to see what you find! :slight_smile:

Tim