Asif

Not really sure where the issue is but could you try the latest code in the dev branch and see if that fixes anything? I've refactored some of the code.

@tigrazone : GLTF material support is not complete yet. Only metallic roughness workflow is used and transmissionFactor from the extensions: https://github.com/knightcrawler25/GLSL-PathTracer/blob/ddbba9a1f2bf1c9dec0f1c2634a1af3949e5f25a/src/loaders/GLTFLoader.cpp#L245-L279 For the last screenshot, you'll have to increase number...

@LiangliangNan : Could you also try the debug branch? It has just enough code to render cornell_box.scene and would be easier to check.

That's good news lol. I added a couple of commits each progressively removing a feature until its down to just bvh traversal and triangle intersection. Let me know if any...

(Edit: Ignore this, its just an effect of clamping negative values) Normals don't look right. Was the screenshot from 9a09313dc3c8deba67cf94fe6e0a7a540b726f07? This is what I see from that commit: 

Do you see the same image with the last two commits where mesh transformations have been removed?

I haven't tested this on Linux but maybe you might need to try with a fresh installation of DevIL as merlinND mentioned in this PR #7

DevIL is required as it is used to load textures. The library provided with the repo might have an issue on Linux. I changed the CMake file to not pick...

While building MoonRay (Step 4: https://docs.openmoonray.org/getting-started/installation/building-moonray/building-moonray-centos-7/#step-4-build-moonray), you might want to check the output of "cmake --preset container-release" to make sure it was able to find CUDA as well as OptiX...

The simplest way (that worked for me) is to first set the path variable to where nvcc is available on your system. This will allow cmake to find CUDA: `export...