Ready for Review

[cis565-fall-2016]

Ready for Review

[pjcozzi]

Looks great so far.

material has a lot of shader related parameters, which we cannot make most of.

It would also be fine to just do a geometry + texture only loader to keep things simple.

I think some knowledge of glTF is needed to write the vertex processing and primitive assembl

To keep this project focused on writing the CUDA kernals for rasterization, can we load all the geometry data into a pre-transformed single buffer or provide some starter code?

Maybe make glTF keyframe animation extra credit.

The pathTracer, I think will benefit even less from using glTF. I will vote for not using it in path tracer. But for WebGL project, I do think it is very beneficial.

Same comments as above - why not do a geometry-only load and pre-transform all geometry and store in a single buffer?

[pjcozzi]

@shrekshao also, is this code in a state yet where it would @likangning93 taking the time to review it?

[shrekshao]

To keep this project focused on writing the CUDA kernals for rasterization, can we load all the geometry data into a pre-transformed single buffer or provide some starter code?

The current code can confirm that all primitives will be in one single buffer. But before each vertex is assembled, they are still in separate arrays. One primitive each. So I left part of the vertex processing and primitive assembly code link (iterate through each primitive to call the kernel function) there which might be a little confusing at first, but it's still part of the assignment. I hope this won't be a problem then.

pathTracer

Previously constructing dev memory of the scene (a list of triangles) is part of the assignment. Do you think it's okay to build this dev array for the students in the starter code?

Is this code in a state yet where it would @likangning93 taking the time to review it

I think a glTF loader should handle the multi hierarchy case rather than leave it to the app. In light of this, I will try to contribute to the glTF loader instead of changing the rasterizer. So yes.

[pjcozzi]

The current code can confirm that all primitives will be in one single buffer...

Looks fine to me.

Do you think it's okay to build this dev array for the students in the starter code?

Yes, our real focus is on the kernels and parallelism. Plus this will be one less thing for them to do so they can focus on, for example, shared memory.