This adds access to the velocity and heat fields from the smoke simulation to Cycles and the viewport.
As the velocity fields do not exist in the high resolution part of the simulator, the velocities for the high res grids are sampled from the low res grid ones, I stay confident that we shouldn't get too blocky of a result.
For the heat field, judging from the fact that heat is based on a linear interpolation of the ignition temperature and maximum temperature based on the value of the flame field at the current voxel, such field was added to the high res solver (as we already have the flame one and it is computed the same way as the low res one).

