So I stuck a tulamp preheat ballast in the westy fixture on my ceiling to make it half preheat and half rapid start, but the funny thing is after a few hours of continuous running, the tulamp preheat ballast is noticeably hotter than the rapid start one, even though you'd think it would be vice versa. The tulamp ballast is running F40T12 lamps. Rest assured I am not running 34 watt in the preheat side, although the rapid start side is running the 34's, and that only raises further questions. When I touch the side of the housing where the ballast is, the tulamp ballast side is actually too hot to keep my hand there for very long but the rapid start, though hot, is at a low enough temperature that I can keep my hand there longer. Is this normal? Is this something I should be concerned about? It's not getting hot enough to trip the thermal protector, and the ballast is a Magnetek 205-TC-P, and Magneteks are known for having sensitive thermal protectors. If nothing else, the tulamp preheat ballast drives F40T12 lamps brighter than they would be on rapid start. There is a noticeable difference in luminosity too and that's a plus of the conversion. Thanks.

HPF RS are using as ballasting component the capacitor, what are nearly loss-free components. The only source of losses is then the transformer, what might be tuned in the way it has to handle nearly half VA then NPF used on most preheat (it start with less damage on both lamps and starters).
Other reason might be, the RS run the lamp at lower current then the Preheat one (as you mentioned the RS is dimmer) - due to manufacturing tolerances, corrosion eating-out some core material (from the magnetic shunt in the preheat), so the leakage inductance get's lower, so current higher, obviously causing higher losses...

The tulamp preheat ballast is a high power factor ballast and from what I've heard, they usually have fewer ballast losses than rapid start ballasts and as result I would think that they would run cooler so this only raises further questions.

If the comparison is HPF vs HPF, then indeed the RS has usually higher losses.
But the size and weight might play significant role, isn't the RS significantly heavier then the preheat?

The main dissipation source is the winding resistivity. If you for some inductive component choose larger core, as it has larger cross-section, you need less turns for the same section, so used wire would be shorter. And as less turns have to fit, the wire might be thicker. And the winding window on larger core is usually larger, so even thicker wire might be used. So you decreased the resistivity, so it's associated losses three fold.

The second dominant loss source are core losses, what are proportional to square of peak magnetic flux. So when you increase the core size and keep number of turns, the field might be lower, lowering associated core losses.

But both contributors have to be balanced: Too much turns mean too much resistive losses, while too few too much core losses (if not saturation problems), so trade-off optimization is needed.

So when you might use larger, heavier, so more costly design, the resulted losses would be lower (assume the optimization was done well), yielding lower operating temperature, so higher reliability and longer life.