Articles: Cooling

Bookmark and Share

Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 ]

14 Super Coolers in Action

Now comes the most interesting part of our test session, that part that took us over two weeks of tests and over 4 days of editorial work. When we ran Linpack tests our quad-core 45nm CPU overclocked to 3.76GHz (+41%) on the weakest cooling system of all participating. The processor core voltage was increase only by 0.075V to 1.275V (+6.3%) in the mainboard BIOS.

We activated “Load-Line Calibration” that lowers the voltage drop on the segment between the voltage regulator and the CPU itself. The memory voltage was set at a fixed value of 1.525V, its frequency equaled 1500MHz with 8-8-8-18 timings:

All other parameters in the mainboard BIOS connected with the CPU and memory overclocking remained unchanged (left at “Auto”).

In this part of our session all coolers were tested with their default fans. If the cooler was equipped with a rotation speed controller, it was tested in two modes: at min and max fan rotation speed. Cooler equipped with PWM-controlled fans were tested in Auto mode (“Turbo” profile in the mainboard BIOS). Since Thermalright TRUE Copper ships without a fan, we tested it with a fan from its fellow cooler – Thermalright Ultra-120 eXtreme-1366 RT – fan model TR-FDB-1600. The same is true for the Thermalright IFX-14 cooler, which also ships without a fan. The tower-heatsinks – Prolimatech Megahalems and ThermoLab BARAM – were topped with a fan from Scythe Mugen 2 with the rotation speed varying between 250 and 1300RPM. The last cooler that has no fan with it, Thermalright AXP-140, was tested with the most suitable fan – Scythe Kaze Maru at 1150 and 1920RPM rotation speeds. I would like to stress once again that we used the same thermal grease for all coolers: it was Arctic Silver 5. The case was also the same: Antec Twelve Hundred.

The red carpet and drum roll are on. Let’s check out the results now:

The coolers on the diagram are lined up in order of efficiency starting with the most efficient one at the very top. Let’s try analyzing the obtained results. The leader is Thermalright IFX-14 with a 4°C advantage over the competitors, which is quite a lot. Overall, when we tested Thermalright IFX-14 we got the impression that this cooler was preeminent among the others. Further tests will confirm this observation.

The leader is followed by a dense group of 11 coolers from Scythe Mugen 2 to Thermalright AXP-140 (at 1920RPM). The difference between the best and the worst in this group is 4°C. The best ones here are Scythe Mugen 2 and Prolimatech Megahalems; only one degree behind them are ThemoLab BARAM and Thermalright Ultra-120 eXtreme that cool the CPU equally well (in this test mode). Noctua NH-U12P SE1366 with two fans at maximum and medium rotation speeds loses another degree. The next one is again Noctua NH-U12P SE1366 at the slowest fan rotation speed, Thermaltake BigTyp 14Pro at 1050RPM and Cooler Master V8. The 11-cooler group is closed by Thermalright AXP-140 with a fan at 1920RPM that yielded to the cooler of the same type – Thermaltake BigTyp 14Pro.

The remaining coolers fall into the outsider group, however, it is important to remember that they are the least efficient coolers of the absolute best out there. Here we see Thermaltake BigTyp 14Pro and Thermalright AXP-140 at the minimal rotation speed of their fans, Cooler Master V10 that has suddenly yielded to the V8 model on Intel Core i7 platform. The very last one in this race is Zalman CNPS9900 LED, which has limited the maximum CPU overclocking for this test. And since we know the maximum CPU overclocking result achieved on the weakest cooler of the today’s testing participants, it is time to check out what the overclocking maximums will be for the rest of the pack:

The difference between maximum overclocking results for each of the cooling solutions is not dramatic, and there is simply no difference between the coolers within their efficiency group. Nevertheless, we can notice a difference in the CPU temperatures at the same frequency and voltage settings. The indisputable leader here is again Thermalright IFX-14 winning 5°C from its closest rival, Scythe Mugen 2. The latter is closely followed by Noctua NH-U12P SE1366 with two fans and two Thermalright Ultra-120 eXtreme coolers. The all-copper TRUE model has finally proved worthy having outperformed its aluminum brother by 2°C. By the way, to ensure CPU stability when tested with Ultra-120 eXtreme with aluminum heatsink, we had to increase the processor core voltage to the next level. At the same time voltage increase didn’t help improve the Core i7 stability at 3958MHz with ThermoLab BARAM cooler, so we had to set the base frequency 1 step lower. Please, do not forget that we are testing these coolers with their default fans, so Thermalright coolers have a slight advantage thanks to a slightly faster fan (although we could argue about it, because the 9-blade Scythe Slip Stream at 1300RPM will hardly yield to the 7-blade Thermalright fan at 1560RPM in terms of the created airflow strength).

After that we see gradual lowering of the stable CPU frequency and increase in its core temperature. The least successful ones here are Cooler Master V10 and Zalman CNPS9900 LED. The new Prolimatech Megahalems rests modestly somewhere in the middle losing in efficiency to the ultimate cooling masters.

Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 ]


Comments currently: 13
Discussion started: 03/25/09 05:15:06 AM
Latest comment: 12/03/09 02:32:14 AM

View comments

Add your Comment