Test Results - Liquid3D
The benefits of heat-pipe cooling are two-fold. Not only is the processor cooled with a liquid which undergoes phase-change, it's still reaping the benefits of the thin-fin copper cooler. It's actually an amalgamation of water-cooling, phase-change, and air-cooling. So what are the end results? To-date I've tried only two of three mounting positions
. I began by situating the SP-94 in the upward-loop position running my P4 2.8C@3.22GHz (230FSB / 1.550Vcore) the SP-94 maintained a mean temp of 29.5 °C.
With Drempels (3D Screensaver/Desktop) I experienced highs of 34 °C, lows at 28.5 °C. The TIM used is Artic Silver 5
(soon to be released), and the motherboard is the Abit IS7-E. Monitoring software is Winbond's DrHardware. Most importantly, I used the Vantec Tornado 92mm/119CFM fan. Room temps varied from 16°C to 22°C, I ran the cooler in each position for 48-hours.
Liquid3D's Test Setup
|CPU ||Pentium 4 "C" 2800Mhz @ 3.22Ghz (1.55v vcore) |
|Mainboard ||Abit IS7-E|
|Cooling ||Thermalright SLK-947U / Thermalright SP-94|
Vantec Tornado 92mm/119CFM
|Memory ||2* 512Mb PC4200 OCZ|
|Video ||Asus FX5900 |
I then oriented the heatsink such that the heatpipe-loops faced down and was amazed at the difference. The cooler maintained a mean temp of 26 °C, with lows to 24 °C and high's reached a comfortable 28.5C. Again all system parameters were the same as above.
Comparatively the SLK-947-U ran an approximate mean temp of 31.50C, albeit a slightly different setup, it's not difficult to make a comparison. For example, hardware used were the following: P4 @3.0GHz (250FSB / 1.550Vcore) Vantec Tornado 92mm/119CFM Abit IS7-E, using Nanotherm XTC. Idle temps which in this case were considered non-overclocked, 2.4C at default (200FSB / 1525Vcore) speeds, and voltages averaged 28C. Below is a screen shot of the SLK-947-U's performance under the condition I've just described:
Until I finish a back to back SLK-947-U / SP-94 in a controlled environment, meaning the only variables being the coolers themselves, I cannot be specific as to the extent at which the SP-94 exceeds the performance of its predecessor. This is how the current situation looks:
The SP-94 outperforms the SLK-947U despite the fact that the CPU is producing more watt/heat!
I will offer you the following theory, extrapolated from my entire body of data. The SP-94 does out-perform the SLK-947-U and that margin seems to exponentially increase with stress (> watt's/heat). Check JMke’s test results for further prove of this. It seems the heat-pipe technology does require higher temps, to fully exploit the vaporization ~ condensation phase-changes. I believe the SP-94 to be the ideal Overclockers weapon of choice in the thermal wars. One surprising (and profitable) fortuity of the SP-94 may be it's ability to perform better in poorly ventilated cases, environments wear the ambient temps are high, and perhaps even LAN boxes, providing there's enough room. I wonder how the cooler would work with a 1U copper cooler? The average desk-top end-user's case is often cluttered, poorly ventilated, and even if a large 92mm fan were installed with a SLK-947-U, the result may simply be forcing hot air, onto an even hotter CPU.
Defeatist? Perhaps not, but certainly a waste of resources. I believe Thermalright's marketing strategy towards the Sally and Joe-desktop user should be; "If they can be convinced, they will come." Once you communicate the propensity for thermal damage with today's often inadequately laid out cases, and the plethora of anomalies which can occur under such conditions, people are usually willing to make some changes.