Test Session

Our methodology of testing cooling systems implies measuring the temperatures of the heater element at different thermal loads created by a special testbed. This approach helps to accurately evaluate the performance of air coolers or water-cooling systems for which the temperature achieved is equivalent of efficiency. However, this traditional method does not produce understandable results with vapor phase-change cooling systems and other super-efficient cooling solutions for the PC. The fact that under a load of 200W the heater element temperature is -10°C does not tell anything comprehensible to the user. It does not make clear the value of extreme cooling and does not explain how the CPU core behaves under such conditions. So, instead of a brief “theoretical” test of the temperature modes, we carried out an extensive exploration of the operation of the ECT Prometeia Mach II GT on overclocked processors.

We tested the Mach II GT on as many as thirteen different processors and five different platforms and we were never disappointed. The temperature reduction down to 0°C and lower most positively affects the overclockability of any CPU. The diagram below shows you the overclocking gain the Prometeia provides over the maximum frequency a processor can achieve with air-based cooling. In each case the extreme cooling helps to reach a frequency which is absolutely unattainable for the given CPU core with an air cooler or a water-cooling system.

Here is a list of tested systems (the CPU model, core, and the mainboard):

• Athlon 64 3000+ (Winchester), ASUS A8V Deluxe
• Athlon 64 3400+ (Clawhammer), ASUS A8V Deluxe
• Athlon 64 3800+ (Newcastle), ASUS A8V Deluxe
• Athlon 64 4000+ (Sledgehammer), ASUS A8N-SLI Premium
• Athlon 64 FX-53 Socket 940, (Sledgehammer), ASUS SK8V Deluxe
• Athlon 64 FX-53 Socket 939, (Sledgehammer), ASUS A8V Deluxe
• Athlon 64 FX-55 (Sledgehammer), ASUS A8V Deluxe
• Athlon 64 FX-57 (San Diego), ASUS A8N-SLI Premium
• Pentium M 1.7 GHz (Dothan), AOpen i855GMEm
• Pentium 4 XE 3.2 GHz (S478, Gallatin), ASUS P4C800-E Deluxe
• Pentium 4 3.4 GHz (Northwood), ASUS P4C800-E Deluxe
• Pentium 4 660 (3.6 GHz, Prescott-2M), ASUS P5WD2 Premium
• Pentium 4 XE 3.73 GHz (Prescott-2M), ASUS P5WD2 Premium

The low result of the Pentium 4 Extreme Edition 3.73GHz is explained by our reaching the maximum FSB frequency the mainboard could support (350MHz is too high a FSB frequency even for an ASUS P5WD2 Premium). The same thing happened with the Pentium M: our i855GME-based mainboard could not yield a FSB frequency above 160MHz.

The AMD64 platform on average showed a 400MHz gain over the maximum frequency of the core with air cooling. This gain was achieved on nearly all of the tested AMD processors irrespective of the stepping.

It’s harder to see a solid trend on the Intel platform, but owners of modern Prescott/Prescott-2M CPUs, i.e. steppings E0 and N0, respectively, are going to enjoy a whopping 800-1000MHz frequency gain above the maximum frequency achieved with air-based cooling.