I didn't answer that question, in all the ranting about heat flow, so here goes:

The "Metal Glaciers" are possible when the whole planet is that hot, and the atmoshere has been stripped away to leave a few residual millibars of metal vapours.

On a "normal" planet, even if you develop a regional-scale magma ocean at ~ 2,000 K in the immediate post-impact scenario, it will act as a heat pump for the existing atmosphere, not as a source of voluminous metal glaciers. (there will be a local zone of superheating around the magma sea, with exhalative deposits of metals, then sulphides, then more normal fumarolic minerals, but mostly the existing atmosphere will be driven frantic.

There is the option to significantly change the state or thickness of the atmosphere by melting ice and vaporising volatiles. This is how Segura et al., and my good self, envisage ephemeral atmospheres on early Mars - impacts trigger volatile evolution, which makes thick atmospheres, which then collapse and rain out, forming the valley networks - i.e. NOT a steady-state wrm and wet early Mars, but a frozen snowball (White Mars) with rare warm and wet episodes.

On Earth, you'd melt the icecaps if any, and for a really huge impact boil the oceans and sterilise the planet. Norman sleep uses this model to explain why all of the most primitive branches of life are heat-resistant extremothermophiles. he suggests that they are the sole survivors of a near-extinction even from the last giant impactor that wasn't quite big enough to sterilise everything. He suggests that life could have originated multiple times on the very early Earth, and been snuffed out each time by impacts, prior to this "not quite big one"...