Open
Description
Deformable simulation currently takes advantages of thread-level parallelism to boost its realtime rate. This can be configured with DeformableModel::SetParallelism(). This function is currently marked as an internal only function because it can easily be misused. In particular, we want to prevent a user from calling SetParallelism(Parallelism::Max()) on a model that has already been parallelized at a higher level. The quadratic thread spawning will hurt performance instead of helping it. We need to restrict the entry point to parallelism configuration so it's less likely to be misused.