Testbed Configuration and Testing Methodology
I have to say a few words about the upgrades of our testbed. First of all, it is now built around Intel Siler DX79SI mainboard with LGA 2011 processor socket based on Intel X79 Express chipset:
Even though the BIOS of this particular mainboard is pretty peculiar, it is fully-functional and allows not only to skillfully “juice” the CPU dry, but also ensure that it remains stable at the highest clock frequencies. We used BIOS version 0308 from November 28, 2011.
The second significant upgrade is, obviously, the CPU. Our Intel Core i7-3960X Extreme Edition processor manufactured using 32 nm process is an engineering sample made in
The typical TDP for this processor at its nominal clock is declared to be 130 W, and the maximum temperature upon exceeding which thermal throttling kicks in is set at 90°C (confirmed empirically). The contact surface of the CPU heat-spreader is 38x38 mm:
According to the following schematics, the processor die is 19.4x14 mm big and it is positioned along the processor socket (parallel to the DIMM slots, in other words):
Therefore, one can assume that for maximum efficiency the cooler should be installed with the heatpipes going across the die, i.e. perpendicular to the memory slots. By the way, the die dimensions on the illustration above taken from an official Intel document do not correspond to the official declared die size of 435 mm2: according to the dimension on that document the die size adds up to 271.6 mm2, not 435.
Nevertheless, we didn’t check out whether this assumption is correct in our today’s test session because in order to correctly estimate the dependence of cooling efficiency on the cooler position on the CPU we need to have perfect contact between the cooler base and the CPU heat-spreader in all positions. As you have already seen, none of the testing participants can actually boast that.
So, we performed all cooler tests inside a closed system case. Here is our testbed configuration:
- Mainboard: Intel Siler DX79SI (Intel X79 Express, LGA 2011, BIOS 0380 from 11/28/2011);
- CPU: Intel Core i7-3960X Extreme Edition, 3.3 GHz, 1.2 V, 6 x 256 KB L2, 15 MB L3 (Sandy Bridge-E, C1, 32 nm);
- Thermal interface: ARCTIC MX-4;
- Graphics card: Asus Radeon HD 6770 DirectCU Silent (EAH6770 DCSL/2DI/1GD5) GDDR5 128 bit, 850/4000 MHz (with a passive heatsink from the DeepCool V4000 VGA cooler);
- Memory: DDR3 3 x 2 GB OCZ Platinum Low-Voltage Triple Channel (Spec: 1600MHz / 7-7-7-24 / 1.65 V);
- System drive: Crucial m4 256 GB SSD (SATA-III,CT256M4SSD2, BIOS v0009);
- Drive for programs and games: Western Digital VelociRaptor (300GB, SATA-II, 10000 RPM, 16MB cache, NCQ) inside Scythe Quiet Drive 3.5” HDD silencer and cooler;
- Backup drive: Samsung Ecogreen F4 HD204UI (SATA-II, 2 TB, 5400 RPM, 32 MB, NCQ);
- System case: Antec Twelve Hundred (front panel: three Noiseblocker NB-Multiframe S-Series MF12-S2 fans at 1020 RPM; back panel: two Noiseblocker NB-BlackSilent PRO PL-1 fans at 1020 RPM; top panel: standard 200 mm fan at 400 RPM);
- Control and monitoring panel: Zalman ZM-MFC3;
- Power supply: Xigmatek “No Rules Power” NRP-HC1501 1500 W (with a default 140 mm fan).
For initial tests and summary diagrams we overclocked our six-core processor with the clock generator frequency set at 125 MHz, the multiplier set at 35x and “Load-Line Calibration” enabled to 4.375 GHz. The nominal processor Vcore was increased to 1.385 V in the mainboard BIOS. After that we tested all super-coolers at an even higher frequency and voltage settings. Turbo Boost was disabled during this test session, and Hyper-Threading technology was enabled. The memory voltage was at 1.65 V and its frequency was 1.666 GHz with 8-8-8-16_1T timings. All other parameters available in the mainboard BIOS and related to CPU or memory overclocking remained unchanged.
All tests were performed under Windows 7 Ultimate x64 SP1 operating system. We used the following software during our test session:
- LinX AVX Edition version 0.6.4 – to load the processor (memory - 4500 MB, Problem Size – 24234, two 11-minute cycles);
- Real Temp GT 3.66 – to monitor the processor core temperatures;
- Intel Extreme Tuning Utility version 3.0 – for monitoring and visual control of all system parameters during overclocking.
So, the complete screenshot during the test session looks as follows:
The CPU was loaded with two consecutive LinX test runs with the settings as indicated above. The stabilization period for the CPU temperature between the two test cycles was about 8-10 minutes. We took the maximum temperature of the hottest CPU core for the results charts. Moreover, we will also provide a table with the temperature readings for all cores including their average values. The ambient temperature was checked next to the system case with an electronic thermometer with 0.1 °C precision that allows hourly monitoring of the temperature changes over the past 6 hours. The room temperature during our test session varied between 24.0-24.5°C.
The noise level of each cooler was measured between 1:00 and 3:00 AM in a closed room about 20 m2 big using CENTER-321 electronic noise meter. The noise level for each cooler was tested outside the system case when the only noise sources in the lab were the cooler and its fan. The noise meter was installed on a tripod and was always at a 150 mm distance from the cooler fan rotor. The tested cooling systems were placed at the edge of the desk on a sheet of polyurethane foam. The lowest noise reading our noise meter device can register is 29.8 dBA and the subjectively comfortable noise level in these testing conditions was around 36 dBA (do not mix it up with low noise level). The fan(s) rotation speed was adjusted in the entire supported range using our in-house controller by changing the voltage with 0.5 V increment.
Besides their default fans, all coolers were also tested with two Scythe Slip Stream 140 fans at 800/1000/1300 and 1750 RPM speeds. The Intel RTS2011LC Water Cooler system was tested with two Thermalright TR-FDB-2000 fans. The rotation speed of all fans was controlled suing the same controller with ±10 RPM precision.
Well, it is time to check out the obtained results.