Articles: Mainboards
 

Bookmark and Share

(2) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 ]

Testbed Configuration

All performance tests were run on the following test platform:

  • Asus P8P67 Deluxe rev. 1.3 mainboard (LGA1155, Intel P67 Express, BIOS version 1702);
  • Intel Core i5-2500K CPU (3.3 GHz, Sandy Bridge, LGA1155);
  • 2 x 2048 MB DDR3 SDRAM Patriot Extreme Performance Viper II Sector 5 Series PC3-16000, PVV34G2000LLKB (2000 MHz, 8-8-8-24 timings, 1.65 V voltage);
  • MSI N570GTX-M2D12D5/OC graphics card (Nvidia GeForce GTX 570, GF110, 40 nm, 786/4200 MHz, 320-bit GDDR5 1280 MB);
  • Kingston SSD Now V+ Series (SNVP325-S2, 128 GB);
  • Cooling system: Scythe Mugen 2 Revision B (SCMG-2100) CPU cooler and an additional 80x80 mm fan for cooling of the area around the CPU socket during overclocking experiments;
  • ARCTIC MX-2 thermal interface;
  • CoolerMaster RealPower M850 PSU (RS-850-ESBA);
  • Open testbed built using Antec Skeleton system case.

We used Microsoft Windows 7 Ultimate SP1 64 bit (Microsoft Windows, Version 6.1, Build 7601: Service Pack 1) operating system, Intel Chipset Software Installation Utility version 9.2.0.1025, Nvidia GeForce/ION Driver 266.58 graphics card driver.

Operational and Overclocking Specifics

In an earlier review we talked about the difficulties we meet when assembling our test configurations, particularly when installing our Scythe Mugen 2 cooler on a CPU. Its wide back-plate often presses against the capacitor feet that stick out of the reverse side of the mainboard near the CPU socket, so the cooler has to be installed askew. However, this is just a small difficulty. The main problem with the P8P67 Deluxe is its central heatsink. One of its fastening screws prevented me from installing the CPU cooler altogether. Recalling that this heatsink didn't really cool anything and didn't need such secure fastening, I simply removed one of its screws and installed the cooler properly.

The mainboard worked flawlessly in its default mode, but provoked some difficulties when I tried to overclock the CPU and memory. I began my tests using BIOS 1703 beta which was supposed to ensure better compatibility with memory modules. However, my memory refused to work with that BIOS at 1866 MHz, so I installed BIOS version 1702 and used it for all the tests.

As opposed to many other mainboards which would issue errors or refuse to boot the OS up at a CPU frequency of 4.8 GHz, the P8P67 Deluxe started up normally but was not stable. It could work for an hour without a problem one time, but the next time it would produce a BSOD 15 minutes into a test. I was steadily increasing the voltage to achieve stability and the resulting power consumption grew up very high under load, so I decided to disable the option for counteracting the CPU voltage drop under load. This helped at first and the system began to pass the tests successfully, but the voltage lowered considerably when I additionally increased the memory frequency to 1866 MHz. So I had to lift the CPU voltage up more, losing in energy efficiency. When the stability and performance tests were passed and the screenshots captured, I only had to measure the power consumption of the system in idle mode and under load (while running the LinX utility). Unfortunately, the overclocked P8P67 Deluxe could only run LinX for a few seconds, hanging up quickly. We had used this utility to check out overclocked CPUs for stability and it is not supposed to be a test of memory modules, but this time around the problem was in the memory modules indeed: LinX would still hang the system up at a lower CPU frequency but ceased to do that after I had reduced the memory clock rate.

So once again I have to say that LGA1155 processors are easy to overclock but hard to test for stability. And without stability, there is no point in overclocking. We used to stress-test older CPUs with the LinX utility but it is not good for LGA1155 CPUs because it can run normally where Prime95 produces errors. Prime95 itself doesn’t guarantee anything, either. When we tested the Gigabyte GA-P67A-UD4 and GA-P67A-UD4-B3, they passed Prime95 successfully, but couldn’t then pass an Adobe Photoshop test. The ECS P67H2-A2 also passed every test except for the simple single-threaded SuperPi utility. And this time around, LinX reveals that the memory modules, rather than the CPU, are overclocked too much. I really wish there were a single program that would be able to quickly and reliably test an LGA1155 platform for stability.

As a result, I had to lower the memory frequency to 1600 MHz, which was partially made up for by lowered timings. The system was absolutely stable then at a CPU clock rate of 4.8 GHz.

We, at X-bit labs, do not overclock to set new records and capture pretty screenshots. Our goal is that the overclocked system were stable and could be used continuously without any problems. Therefore we do not turn off any components or technologies and do not fix the voltages, so the mainboard retains all of its features. All of the power-saving technologies keep on working, lowering the CPU multiplier and voltage in idle mode.

Overclocking the CPU and memory on the ASUS P8P67 Deluxe was not easy because the numerous interrelated parameters require a lot of testing and also some experience on the overclocker's part. However, the mainboard offers automatic overclocking tools which make the whole process much easier for beginner users, even though you can't expect to get maximum or optimal overclocking results with them. In order to enable automatic overclocking, you only have to change the position of the TPU switch on the mainboard. If you don't want to open up your system case, you can simply choose the OC Tuner option in the BIOS. In either case I got the same results after rebooting:

The system not only increases the CPU frequency multiplier to x42 but also lifts the base frequency up to 103 MHz, effectively overclocking the CPU to 4.33 GHz. The power-saving technologies keep on working at that:

Of course, the mainboard does not know the capabilities of the particular CPU sample, so the results are lower than what I got by overclocking manually. My CPU could work at 4.5 GHz without increasing the voltage, but manual overclocking takes a lot of time and testing whereas the automatic overclocking feature accelerates your system in just a few seconds, which is going to be perfect for inexperienced users.

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 ]

Discussion

Comments currently: 2
Discussion started: 07/07/11 08:22:04 AM
Latest comment: 07/16/11 01:10:04 AM

View comments

Add your Comment