System setup and benchmarks
We've looked at it in detail. All the looking in the world won't tell us just how it performs, so on to the performance. The test setup was as follows.
·
Asetek
WaterChill watercooling kit
·
Intel
Pentium 4 3.0GHz 'C' 800FSB CPU
·
Asus
P4C800 Deluxe Canterwood Motherboard
·
2 x
256MB Corsair XMS3500C2 memory
·
ATi
Radeon 9800 Pro
·
300w
FSP PSU
·
61.4GB
IBM Deskstar 120GXP hard drive
·
ThermalRight
AX478 with Delta 80mm 68CFM fan (EHE)
·
Cooler
Master Fujiyama Heatpipe cooler (IHC-L71)
To fully push the system, I needed to find a suitably high wattage output from the Pentium 4 CPU. Raising both the overall speed and voltage to 3543.9MHz and 1.632v load voltage seemed like a stringent enough test.
Ambient temperature ranged from 21.3 - 21.7c. Prime95 was used to establish a load temperature after 20 minutes of running. Here's what we found.
I've also included figures for the WaterChill system with the Sunon 120mm fan switched off. Amazingly, it was stable, albeit at a precariously high temperature. The Cooler Master represents the same kind of noise levels as the WaterChill at 7v. In a word, quiet. The WaterChill managed to knock 13c off the CM's load temperature with similar sound levels. What's even more impressive from my personal point of view is that the WaterChill outperformed the incredibly loud ThermalRight AX478 and Delta 68 CFM combination at the quiet 7v setting. Turning up the radiator's fan to 12v shaved another 4c from the overall load temperature. The 12v WaterChill was, subjectively speaking, still far 'quieter' than the aforementioned air-cooling combination.
I also dipped a thermocouple into the reservoir. At 12v, it displayed a water temperature of 28.3c. At 7v, this rose to 31.9c, and 50.9c when run passively. This seemed to somewhat tie-in with the Asus Probe's load readings. Please bear in mind that testing was undertaken in an open-air environment.
All the runners were stable at 3.54GHz. The next step was to see just how high each cooling system would take the test 3.0GHz P4. Basic stability comprised of running Prime95 for 10 minutes without failure. Running with the same maximum load voltage of 1.63v, this is what we found.
The speeds were rounded up/down for easy reading. A 100MHz gain over a quality heatsink tells you that even lowering temperatures by ~ 15c is highly beneficial if you're really looking to push your system.
Let's wrap it up now.
