Liquid3D's Test System
|CPU ||Intel P4 550 Retail-3.4GHz LGA 775 |
|Mainboard ||Abit AA8|
|Memory ||Kingston HyperX PC2-5400 1024MB kit|
|Video ||ATI Radeon X800 XT|
|Power Supply ||OCZ Technology PowerStream 420W|
|Operating System ||Windows XP SP2|
In addition to the components listed above, the cooling system includes a double heater core, 2x120mm/92CFM Sunon fans (pull), Danger Den's clear bay reservoir, and 1/2" tubing except at the pump where adapters run the Delphi's native 3/8" fitting's to the 1/2" system.
Although the DDC-12V pump is absolutely perfect for those wanting their water cooling neatly tucked away in their case, for ease of access and simplicity I've located the water cooling components externally. Besides the reasons to follow, the Hydor was in the same location and changing the location of the pump would in fact change the pressure. For this reason I've propped the DDC-12V on its box, so that the impellers were at the same height. My unconventional component placement simplifies measurement of ambient/case temperatures, and makes for ease of access to the devices. I removed the side panels from my TTGI USA TT-561T2-BK
to eliminate any temp differences between ambient room, and internal case temps.
This simplifies the formulae, eliminating additional variables. It's also much easier to judge the pump's decibel level. On this point, I must say this is definitely the quietest pump I've ever tested. Its floating rotor/impeller
is self-aligning and suspended in the water, this greatly reduces vibration, ergo noise. Since the pump's fittings are native 3/8", I used adapters to insert the pump into a 1/2" ID system required for the TDX. As I stated earlier, the pump arrived when I was completing a review for the TDX on Socket 775, using an overclocked P4 550 on the Abit AA8. Therefore the test system only require that I switch out the Hydor L-40 pump replacing it with the DDC-12V pump seen below.
Before providing charts which will exemplify performance with respect to CPU temperature under the cooling prowess of the TDX, I'll include a few screenshots reflecting methodology. The pump includes a variable voltage connector, which can be plugged into the motherboard and flow rate reduced through the BIOS. For these tests I ran the DDC-12V flat out, which simply involves leaving the variable voltage connector unplugged. This is the ideal scenario for any comparative test as the number of variables in our evaluation is reduced to the pumps themselves.
One couldn't ask for a better test environment, with the exception there be absolutely no variation in the ambient room temp between tests. To generate the maximum amount of heat from the P4 550, I employed Prime95
running in Torture Test, Small FFT's. In my opinion looping this utility does generate a large amount of heat, however; I do prefer SETI
. The problem with the latter is its variation. To view the thermal clock throttling as it occurs, I've used the recently released Throttle Watch
from Panopsys. This utility tracks thermal clock throttling TM1, and TM2 on P4, Xeon, and Pentium-M processors.
Our first screenshot was taken running the Hydor L-40 pump. Our P4 550 is running at default speed 3.4GHz (200FSB) at 1.38Vcore, CPU-Z window, Abit EQ utility, and Throttle Watch display.
The next screenshot captures the system under the power of the DDC-12V pump, under the same conditions. P4 550 running at default speed 3.4GHz (200FSB), and 1.38Vcore.
From the tests above there's virtually no temperature difference running either pump in the system at this point. This is because the Hydor L-40 was running with a slightly lower ambient temp (-2° C). Up till now we've only seen idle performance running the system at default speeds/voltage. Our next screenshot will show the system powered by the L-40 once again this time overclocked to 4.0GHz (240FSB), and at Prime95 LOAD, Torture Test / Small FFT's node (with the P4 Hyper Threading activated this translates into a 50% CPU load)
We see the Hydor L-40 does its job here, keeping H20 flowing and temps remain at 33°C. Can the DDC-12V keep up with the Hydor L-40 flowage beast? At 2800l/h you could supply Los Angles basin with it's drinking water, but could it pump that water over the hills to its overclocked homunculi-headed robots?
Let us see...