BFG's GTX 295 H2OC: Water-Cooled Graphics
Normally when we gather GPU temperature data, we log the results from an open-air testbed and leave it at that. There are far too many cases and ambient temperature conditions out there in the real world to possibly model. In the case of the BFG H2OC, however, there was good reason to put it into an enclosure. Without actually testing the card in a restricted environment against an air-cooled model, there was no way to compare the real advantage of water over air.
We've got a roomy Cosmos case that we figured would be just the thing for a watercooled card like the H2OC, and did not expect any problems until we saw BFG's image of the card in a case. See if you can spot the problem:
![]() |
|
Here's our starting point; a base comparison of GPU temperatures using the air-cooled GTX 295, the water-cooled BFG H2OC at low fan speeds, and the same card at its medium fan speed. While we could have modeled the card at maximum RPM, there's no reason to buy an $800+ card if you want the sound of an air turbine—you can literally buy that at half the price. One thing to remember here is that the air-cooled card was perfectly stable at all temperatures, despite the sky-high readings. Water is an easy winner, here, by a sizable margin.