User:Ziqiu1995/sandbox

In physically-based simulations, solving the dynamic equations in each time step usually leads to solving a linear system $$Ax=b$$. The preconditioning technique can be used to reduce the condition number of the system by defining an approximation of the system matrix $$A$$, which is easier to invert. Based on this technique, the multi-threaded asynchronous preconditioning removes the factorization of precise preconditioners from the main simulation loop to a dedicated thread. This strategy allows obtaining a good preconditioner while releasing the computing expense in the main simulation loop.