GPU Acceleration of Near-Minimal Logic Minimization