It seems like interpolation could be largely parallelized. I think it’s important because some other approaches are not as flexible, especially if you want to tweak a gradient value for only one point. This approach allows unitary modifications without affecting the rest of the curve
carolyn-wang
I agree with @myxamediyar. It seems like Catmull-Rom can be nicely parallelized because the computational process involves calculating independent curve segments and partitioning the data into subsets. GPUs could be well-suited to this kind of computation, or otherwise multithreading on CPUs might also work
It seems like interpolation could be largely parallelized. I think it’s important because some other approaches are not as flexible, especially if you want to tweak a gradient value for only one point. This approach allows unitary modifications without affecting the rest of the curve
I agree with @myxamediyar. It seems like Catmull-Rom can be nicely parallelized because the computational process involves calculating independent curve segments and partitioning the data into subsets. GPUs could be well-suited to this kind of computation, or otherwise multithreading on CPUs might also work