Voor de echte die-hards heeft Overclockers Australia een review van de Senfu radiator, waarmee water in een water-cooling systeem gekoelt kan worden. Onze Australische kameraden hielpen een Celeron 366 dankzij water koeling op 616MHz, waar deze normaal gesproken niet boven 583MHz wilde gaan:
I fired up the system to see what I could get from it. I'm not using a peltier at the moment as I don't want to seal the jacket/chip/slocket assembly up just yet and I don't want to run a peltier in these hot and humid Sydney summer days without decent sealing. I used the Soyo SY6-VBA-133 motherboard, of course, because it can monitor the internal temperature diode of the CPU. Using the same C366 as in my water-cooling and peltier experiments, I was able to run stably at 616MHz @2.2v for an entire day, looping the Unreal flyby with SETI@HOME in the background. This CPU is only capable of email@example.com normally with a GlobalWin FDP-32 (big heatsink/fan) on it. The temperature in my room hit a high of 27C during the 8 or 9 hours it was running and the reported on-die temperature was consistently about 9 degrees above ambient. The radiator and the water in the reservoir never felt noticeably warm. The jacket felt slightly warm. I tried it the next day with a C400. It was stable at 630MHz at 2.3v, the ambient temperature was a little warmer, hitting 28.5 degrees in my room for about 2 hours in the middle of the day. Again, the on-die temperature never rose more than 10 above ambient. With the FDP-32, and at 600MHz, this chip usually sits at about 43C.