accuracy of IcePak for heat conduction problem

I remembered there was a post arguing how the simulation could match the experimental data.

Today I just run a simulation for the below structure:
60mm (length)x 60mm (width) x 6mm (thickness) Al block, top surface there is  a 10mmx10mm size heat source with 20W input, bottom side has convective heat transfer coefficient of 100 W/m^2-K.

When the mesh size is 0.5mm(all in x,y,z) the steady-state temperature rise is 53K, when the mesh size is 0.3mm, the steady-state temperature rise is around 60.5K. While the exact analytical solution result is 60.53K.

The 0.3 mm mesh already took my dual-core laptop almost 30 minutes to solve the problem, most time spent on mesh.

So my message is that even for simple problem, an already fine mesh ~0.5mm with reasonable solving time will already yield noticeable error.