Quick way to find a point inside a manifold object

[pyenthu]

Is there a fast way to find whether a point is inside this manifold object based on the structure you have?
My objective is to perform some tasks with vertices of another mesh that intersects a manifold object without actually doing a Boolean operation.
What I am looking for is to iterate over its vertices to see if they lie in the manifold. It has to be fast and near real time.

[pca006132]

We do have a bvh-based collider, but it is not exposed to the public API.

[pca006132]

It may be possible to do that, but it is not our priority for now because warp is rarely the performance bottleneck, and performance with GPU is actually not that simple. This library had CUDA backend previously, but was removed because we can get faster performance by better CPU parallelization. We will enable CPU parallelization on js when tbb support with Emscripten is more mature.

[elalish]

Actually, WarpBatch() was added for exactly this reason, but it looks like it's only been hooked up to the Python bindings so far - would you like to make a PR to add it to the WASM bindings too? Our WASM doesn't touch the GPU, but your batch function could, I believe. I'm also curious if we might need to think through some async issues...

[pyenthu]

Yes absolutely I can use the gpu.js function to modify this in async. See in three mesh one can run the array of positions to the gpu and it is really fast. then you get that as a batch and just assign the vertices, update the normals and it works fine. If you have the capability to access the vertices and modify them in a batch it would be fantastic.
But the key is to be able to do this with an array of vertices in a parallel process on the gpu. That coupled with all the csg functionality you have would be a game changer.
The other option is that I build the object in three and use the "batch" process but then I lose the capability to do fast csg your library does. I want rhe functionality of the three morphing along with manifold csg. That is a very good combination.

[elalish]

SG - would you like to try exposing our WarpBatch() in our WASM? I think it shouldn't be hard - you can follow the existing Warp() API. You're ideal since you can test it and let us know if there are practical problems we haven't thought through.

[pyenthu]

Let me try. I am not that savvy with several technologies. If there is a guide to doing this let me know k will try.