Reimplement camera in taichi?

[Vesnica]

The current implementation of the camera does not leverage Taichi's acceleration advantages, which significantly impacts simulation efficiency during camera rendering operations. Is it possible to reimplement the camera computational logic(perhaps including other sensors as well) using Taichi in the furture? Doing so would greatly enhance the efficiency of acquiring simulation data.

[zhouxian]

Yes we are trying to rebuild the whole rendering pipeline - faster, while being more photorealistic

however, using taichi for rendering is not necessarily a bonus: rendering is already gpu accelerated, and most of our existing rendering pipeline is jitted via numba already

stay tuned for a more advanced renderer :)