FPS difference between binary sample and compiled from source code


I noticed that there is a big FPS difference between the bundled binary in the sample folder and that which I compile. For example. for SimpleDepthViewer-SFML, the bundled binary gives a FPS range of 25-47 fps, while the one I compile from source will give me about 7.6-8.1 fps.

Is there any optimzation flags needed to get better FPS? I am running on Ubuntu.


The CMakeList used in the Astra SDK doesn’t specify a target build. There’s a PR that fixes this: Default CMakeLists.txt build type to release by SirDifferential · Pull Request #30 · orbbec/astra · GitHub</tit

I don’t fully understand what kind of build the CMakeList generates without this PR, but it at least has asserts in it.

I tired modifying the CMakeList, seems like it has little effect. The FPS is still very low. :frowning:

*Edit: my bad, there is a typo in my CMakeList. After fixing that, the fps is much better. Getting about 17-19 fps now.