Articles: Mainboards
 

Bookmark and Share

(6) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 ]

Power Consumption

We measured power consumption using Extech Power Analyzer 380803 device. This device was connected before the system PSU, i.e. it measured the power consumption of the entire system without the monitor, including the power losses that occur in the PSU itself. When we took the power readings in idle mode, the system was completely idle: there were even no requests sent to the hard drive at that time. We used LinX program to load the Intel Core i7-860 CPU. For more illustrative picture we created a graph showing the power consumption growth depending on the increase in CPU utilization as the number of active computational threads in LinX changed in nominal mode as well as during overclocking. Besides DFI LANPARTY DK P55-T3eH9 we also included the results for all previously tested LGA1156 mainboards.

Frankly speaking, if I hadn’t been the one running all these tests, I would have never believed the results I saw. It can’t be that the board showing the same power consumption in idle mode as all other solutions would demand more and more energy as the workload grew higher.  Under maximum load the power consumption difference reached 50 W compared with the least efficient mainboard, which until now has been a DFI solution, and 60 W compared with most of the other tested mainboards! It is totally unbelievable! And I didn’t believe it and started retesting and rechecking everything over and over again, making sure that I didn’t forget to bring down to nominal all the voltages in the mainboard BIOS. Nope, I didn’t, all optimal settings were in place, but the board kept consuming enormous amounts of power. So what are we going to get during CPU overclocking? Power consumption close to an inflammability limit? No, the results demonstrated by Intel DP55KG mainboard were hardly any different from everyone else’s.

Equipment doesn’t lie, and if it does then not so dramatically.  If the results are repeatable, then no matter how unrealistic they seem, there should be a logical explanation. Now we just have to find one. Why is the power consumption of Intel DP55KG mainboard almost the same as that of all other testing participants during overclocking and so unbelievably high in nominal mode?

One of the possible explanations is the way the board sets higher voltages on its own. In nominal mode nothing could prevent it from doing that, as we barely change the automatic settings. During overclocking, however, we kept in mind this peculiarity of the board and set the voltages lower than necessary, so that the resulting values could turn out exactly what we needed them to be. Yes, it could explain a 10 W difference, maybe a 20 W one, but not a 50-60 W difference!

The second hypothesis is the use of non-optimal algorithms for processor fan rotation speed adjustment. The board maintains very low processor fan rotation speed of less than 500 RPM. It is enough for a high-performance cooler to keep the CPU temperature in nominal mode within acceptable range even under heavy load, but the processor voltage regulator components heat up more than usual in this case because they lack airflow. The higher the temperature, the higher the power consumption. Pretty logical, but a 50-60 W difference is still way too big. Besides, this assumption doesn’t explain normal power consumption during overclocking. In this case the rotation speed control feature uses the same non-optimal algorithms, even the CPU temperature is higher, but the power consumption of Intel DP55KG mainboard is not any different from the rest of the testing participants.

So, looks like the third explanation makes most sense of all: processor voltage regulator circuitry is not functioning properly or is not designed correctly. Only under heavy processor load created during overclocking it proved efficient enough, but wastes a ton of energy when the CPU runs in its nominal mode.

Well, let’s check things out now. I can’t uncover the defects inside the processor voltage regulator circuitry on my own, but I can manually adjust the voltages in nominal mode and see how it affects the power consumption readings. I can set the fan at its maximum rotation speed and repeat the tests to see how the power consumption changes. Finally, I can do both at the same time. So, you can imagine how surprised I was when even before I made any changes to the testing conditions for Intel DP55KG mainboard, its power consumption in the nominal mode got close to normal.

So, let’s replay everything one more time. At first we overclocked our processor and ran the performance and power consumption tests. Then we loaded the optimal system settings and ran the tests one more time and this is when the power consumption of Intel DP55KG mainboard turned out considerably higher than the rest. After that I reflashed the old BIOS version back onto the board to take the screenshots for the review and then returned to the latest BIOS at the time of tests. Could reflashing the BIOS one more time have brought the board back to its normal operational mode? Let’s see. We overclock the processor one more time. By the way, I have already said that the early BIOS versions for Intel DP55KG mainboard do not have the utility for updating the BIOS and do not allow saving user settings profiles. However, despite the fact that we reflashed the BIOS twice already, the overclocking settings profile remained intact.

So, we overclock the processor one more time, then load the optimal system configuration settings and check the power consumption. Yes, the results are again incredibly high. Let’s try and set the CPU Vcore manually as close to the nominal as we can, and again the board’s power consumption comes back down to normal levels. So, here is what’s going on: when we load optimal system settings, the board doesn’t change the voltages back to their nominal values. Everything in the BIOS looks the way it is supposed to, but in reality only the frequencies get lowered, all other parameters go back to their nominal settings, but not the voltages: these remain where they were during CPU overclocking. You can’t notice it in idle mode, because the voltages are lowered by operational processor power-saving technologies. But under heavy load core voltage increases resulting in abnormally high power consumption. All our super-high power consumption readings taken off Intel DP55KG mainboard were correct and completely real and can be explained by a BIOS error.

Further tests showed that you don’t need to reflash the BIOS to bring Intel DP55KG mainboard back to normal. Just use the onboard jumper or a back panel button that will return all settings to their nominal values. Now it is the right time to run our power consumption tests, when Intel DP55KG really does work in its nominal mode.

As you see, things are far from the catastrophe we have anticipated after the first round of tests. The power consumption of Intel DP55KG mainboard in nominal mode turned out comparable with that of other mainboards. However, the users who only tried once to overclock their processor on Intel DP55KG and then went back to the nominal system settings, did not really know that they mainboard started to consume way more power. And the result of 50-60 W is definitely not the limit. The higher you set the voltages during your overclocking experiments, the more power will Intel DP55KG mainboard consume. Make sure you use the jumper or button to return the board back to normal.

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 ]

Discussion

Comments currently: 6
Discussion started: 12/03/09 04:01:06 PM
Latest comment: 05/17/13 04:19:24 PM

View comments

Add your Comment