For any of you posters, please consider that liquid viscosity decreases exponentially with temperature. In practical terms, it means every so many degrees the temperature goes up, the viscosity halves. Viscosity directly affects power requirements for the oil pumps, inversely affects efficiency of closely spaced fins because of laminar (higher viscosity) versus turbulent (lower viscosity, like with air), with turbulent being far more efficient in dissipating heat.
If I wanted to do this, I would use a low-viscosity mineral oil (without additives, mind you) and aim the oil pump output to all heat sinks, after passing it through a temperature-controlled cooler. A submerged system has the advantage that, when temperature climbs, viscosity lowers and more fluid passes through, so more cooling where needed. Pure mineral oil should cause no corrosion issues on the board.
All in all, a submerged system has inherent self-regulation because of the viscosity factor.
Good idea.
P.S.:
I was thinking on replacing cooling fluid in my car radiator with some sort of high-temperature oil, but then I have to design for a more efficient radiator, higher flow speeds for both oil and air, and so on. The advantage would be the end of corrosion for engine parts and the self-controlling temperature due to viscosity changes.
I heard the old VW motors were cooled by air, but in fact the air cooled the oil, and then the oil cooled down the engine, with varying degrees of efficiency (or not ;-)))