Articles: Mainboards
 

Bookmark and Share

(18) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 ]

Testbed Configuration

We performed all our tests on a testbed built with the following components:

  • Gigabyte GA-Z77X-UP7 rev. 1.0 mainboard (LGA1155, Intel Z77 Express, BIOS version F5f);
  • Intel Core i5-3570K CPU (3.6-3.8 GHz, 4 cores, Ivy Bridge rev.E1, 22nm, 77 W, 1.05 V, LGA 1155);
  • 2 x 4 GB DDR3 SDRAM Corsair Vengeance CMZ16GX3M4X1866C9R (1866 MHz, 9-10-9-27 timings, 1.5 V voltage);
  • Gigabyte GV-T797OC-3GD (AMD Radeon HD 7970, Tahiti, 28 nm, 1000/5500 MHz, 384-bit GDDR5 3072 MB);
  • Crucial m4 SSD (CT256M4SSD2, 256 GB, SATA 6 Gbps);
  • Scythe Mugen 3 Revision B (SCMG-3100) CPU cooler;
  • ARCTIC MX-2 thermal interface;
  • Enermax NAXN ENM850EWT PSU;
  • 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 9.3.0.1020, AMD Catalyst graphics card driver version 12.4.

Operational and Overclocking Specifics

The system assembly on Gigabyte GA-Z77X-UP7 went smoothly. Unlike the mainboards from some other makers, such as ASRock or Asus, for example, which try to “improve” the processor operational settings and switch to non-standard operational modes by default, Gigabyte GA-Z77X-UP7 ensured nominal settings for all system components: processor, memory, graphics card. Upon system start-up the board displays a start-up image with all the hot-key reminders.

This is also an advantage compared with some other mainboards (for example, from Asus), but as we could see there was no hint to use the “Tab” key, which would normally turn off the start-up image completely and allow you to see the information displayed as we go through the boot-up process. Unfortunately, Gigabyte mainboards display no information at all during the start-up process. The F9 key seems to be a doubtful and unfair replacement, because it displays the same exact info as we can see in the mainboard BIOS. You have to press this key to get access to this info, but nevertheless, there is absolutely no mention of the memory frequency, which is a pity.

We noticed right away that the heatsinks on the processor voltage regulator components were suspiciously warm. Nothing critical: the infra-red Fluke 62 Mini thermometer read 42-43°C on the heatsinks, and maximum 49°C on the voltage regulator itself right next to the heatsinks. Unlike LGA 2011 mainboards, most of which require an additional cooling fan for the components around the processor socket, the heatsinks on LGA 1155 mainboards usually remain cold and warm up a little bit only when the CPU utilization is really high. In our case, the system was working in its nominal mode, but it started feeling warm even when idle. It was very unusual for LGA 1155 mainboards. Therefore, we decided to look into this matter immediately.

At first we confirmed that all Intel’s processor power-saving technologies were working fine. Most of the power-saving parameters in the BIOS were set to Auto, but even when we manually set them to “Enabled”, the system power consumption didn’t change. In other words, unlike the mainboards from some other makers (for example, Micro-Star), all power-saving technologies worked perfectly fine on Gigabyte GA-Z77X-UP7 by default lowering the processor clock frequency multiplier and core voltage, turning off idle processor units and switching it to energy-efficient state in idle mode. This is all great, but it means that we should blame the powerful voltage regulator circuitry with 32 phases for the excessive heating of the components. Large number of processor voltage regulator phases can only do good under heavy CPU load. If it is idle, or isn’t fully utilized, then there is no real need for a multi-phase voltage regulator and it could even become a problem for wasting precious power. To counterbalance this phenomenon many mainboards these days use special technologies that dynamically adjust the number of active voltage regulator phases depending on the current load level. However, we have already seen multiple times that this feature doesn’t really work on contemporary Gigabyte mainboards.

It was back in early 2008 that Gigabyte mainboards came out with rows of LED lights indicating visually how many voltage regulator phases were active at any given moment in time. At first “Dynamic Energy Saver” technology only worked once you launched the namesake utility in the operating system. Later the company implemented this technology in hardware and put it on a variety of mainboard based on different chipsets and designed for different processors. Everything changed in the beginning of this year, when they introduced the new 3D Power technology together with the seventh series chipsets. The transition to new digital voltage regulator circuitry was a great improvement, but it also brought with it at least one drawback: from now on Gigabyte mainboards no longer had Phase LEDs and we couldn’t see any practical effect from the dynamic changing of the number of active phases in the processor voltage regulator circuitry. This time we once again measured the system power consumption with the “PWM Phase Control” parameter set at its maximum “Extreme Performance” setting as well as with the minimal “Light Power” setting, but didn’t see any difference. It could be why Gigabyte GA-Z77X-UP7 consumes more power than regular mainboards, which you will see later in our today’s review. The energy-inefficiency of this particular mainboard model can be only partially explained by the additional PLX PEX 8747 hub, because it consumes more power than mainboards with the same hub onboard, and even more power than any of the Intel Z77 Express based mainboards we have tested so far.

However, we should admit that multi-phase processor voltage regulator is a perfect fit for an overclocker mainboard. It is a real pity that the dynamic adjustment of the number of active phases doesn’t work. This is the reason why the heatsinks are getting warm even in idle mode, which indicates clear waste of power. And do you know how the voltage regulator responded to CPU overclocking and heavy operational load? Well, it hardly responded at all. Remarkably, our overclocked processor loaded with LinX utility, which is a much heavier application than some regular task, has barely affected the readings off the processor voltage regulator. The heatsinks temperature rose from 42-43°C to 44-45°C, and the maximum component temperature next to the heatsinks increased from 49°C to 51°C. In this case the CPU clock speed changed from 1.6 GHz to 4.5 GHz, the total system power consumption almost doubled, but thanks to the exceptionally powerful voltage regulator circuitry on the mainboard it didn’t get affected at all by any of these. I am sure we would have missed this miniscule increase,  if we didn’t have an infra-red thermometer on hand.

However, despite all this, we failed to overclock our test processor to its maximum of 4.6 GHz and had to stop at 4.5 GHz. I don’t know about you, but I found it very strange. Until today we had a very simple strategy: we tried hitting 4.6 GHz frequency, and if we failed we lowered our overclocking goal to 4.5 GHz, which was easy to achieve almost on all mainboards we have reviewed so far. However, Gigabyte GA-Z77X-UP7 mainboard seemed to be so powerful and made such a bold statement with its looks and feature set that its inability to overclock our processor to its maximum seemed absolutely unnatural. Therefore, this time we decided to make an extra effort and run additional tests.

We always overclock mainboards in such a way that they could be used permanently in this mode. Therefore we do not try to make our life easier by disabling any of the mainboard’s features, e.g. onboard controllers, and try to keep the CPU’s power-saving technologies up and running. However, traditional processor overclocking recommends disabling all unutilized controllers and functions, including the power-saving modes. After that you should lock the processor frequency. The processor Vcore should also be set at a permanent fixed value. And only then you can proceed to stability tests. If you can’t achieve stability, then you should raise the Vcore. If the temperature increases too much, then you should lower the clock frequency. After all of that tweaking, you end up with a stable overclocked system.

I really wanted to say that this approach produced great outcome momentarily, but I have to admit that it didn’t work out that well at first. The thing is that I haven’t used this primitive overclocking algorithm for years and therefore was setting the core voltage too low at first. However, once I found that balance between the processor core voltage and Vdroop, things came together rapidly. I didn’t time myself, but it felt like it had taken me less than an hour to get a stable and reliable system working at 4.6 GHz. Just for your information, I spent an entire day trying to find optimal parameters for energy-efficient overclocking, but had to stop at 4.5 GHz. So, what was different this time? And the only difference is the system power consumption in idle mode, which equaled a little over 80 W with power-saving technologies enabled, and increased to 110 W with power-saving technologies disabled. This is exactly why we believe it is unacceptable to overclock while disabling all power-saving technologies.

I am sure you understand the difference between reasonable use and waste of power. When you have five lights on in your room, but that much light is necessary to read, draw, stitch or do anything else that requires good lighting, it is OK. But it is wrong to have the light on in an empty room, because someone simply forgot to turn it off. The same is true for computers. If the system consumes 100, 200 or even a 1000 W, it is OK as long as this power is used efficiently, and it doesn’t really matter if it is used for work or for entertainment. However, it is unacceptable to overclock your system without retaining the power-saving technologies. Yes, it is much easier to succeed when overclocking in this manner, but it will most likely satisfy only young teenagers, whose maximalism requires getting results at any cost. They sincerely believe that only movie super-heroes can save the world and do not yet realize that they could also contribute to maintaining ecological balance by simply turning off the light or overclocking while keeping power-saving technologies up and running. This is why for years we have been encouraging users to approach overclocking in a reasonable, thought-through and balanced manner and keep power-saving technologies up and running. In our opinion, this overclocking is truly worthy of a responsible adult. Yes, it is more complicated, because instead of one fixed frequency at one fixed voltage setting, you have to ensure stability in a broader frequency interval with constantly changing voltage. It is indeed harder to succeed with so many variables and the obtained results may often be lower than during overclocking “at any cost”, but this is the right way.

We cannot explain why some boards overclock better and some worse. So far, only ASRock mainboards have been continuously overclocking our test processor to its maximum frequency of 4.6 GHz. Not all Asus boards, and not all Gigabyte boards did the same. Micro-Star mainboards seem to be unfit for energy-efficient overclocking. They cannot increase the processor core voltage in Offset mode, i.e. by simply adding a desired value to the nominal setting. MSI boards can only set the core voltage at a fixed value. Yes, it is easier to overclock this way, but all power-saving technologies will be turned off. Therefore, optimal overclocking on Micro-Star mainboards is only possible without changing the processor Vcore. In fact, it is a pretty good option, too, and I personally always do it this way on my home computer systems, because in this case you boost the performance almost without any increase in the system power consumption. However, when we review products, we have to fully uncover the potential of the tested boards that is why we also increase the voltages, but do whatever it takes to keep all power-saving technologies up and running.

There were reports about Gigabyte GA-Z77X-UP7 setting an overclocking record for an Intel Core i7-3770K processor, which managed to hit 7.102 GHz frequency. Of course, this frequency was reached with extreme cooling methods involving liquid nitrogen and with all extras disabled for good, including the power-saving modes. Yes, this record-breaking frequency does show high overclocking potential of this product, but our specific Intel Core i5-3570K processor could only be overclocked to 4.5 GHz.

At the same time we increased the memory frequency to 1867 MHz and corrected its timings accordingly.

And of course, we overclocked our system in such a way that it could be used permanently in this mode. We did not disable any of the mainboard’s additional controllers and kept CPU’s power-saving technologies up and running. As a result, they lowered the processor clock frequency multiplier and core voltage, turned off idle processor units and switched it to energy-efficient state in idle mode.

 

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 ]

Discussion

Comments currently: 18
Discussion started: 11/26/12 09:48:37 PM
Latest comment: 01/10/14 02:39:52 AM

View comments

Add your Comment