Real-time embedded applications are normally viewed as continuous processes and are often specified using Partial Differential Equations (PDEs) with certain boundary conditions. In this paper we present techniques for fast mass-parallel numerical solving of PDEs. We compose specialized lattices based on the integer number approximation specified with Sleptsov nets to be implemented as dedicated hardware, which we prototype on an FPGA. For mass-parallel solving of PDEs, we employ ad hoc finite-difference schemes and iteration methods that allow us to recalculate the lattice values in a single time cycle with appropriate accuracy suitable for control of hypersonic objects and thermonuclear reactions.