I wish I could find the reference. I read a report years ago that compared column temperature control designs for temperature uniformity. To my surprise, the forced air design had better spatial and temporal uniformity than the block design. (Also better than convection, but that was no surprise.)
There are a lot of subtleties to a good block design, and a poor design will actually have worse heat transfer than forced air. Sorry to say, I inflicted one such thing on the world.
A related issue is preheat of the mobile phase. I have been testing a method at 50 °C, and for a 4.6mm column at 2 mL/min it nearly doubled the plate count. For a 2.1mm column at 0.5 mL/min, it nearly halved the count.
Here at Dionex, our better lines use forced air, and have a preheat option. The low-end lines use a block, also with a preheat option.