Quote:People certainly weren't indifferent to our article, as websites from across the globe reported its findings. Furthermore, we received feedback from a number of Dutch webshops, who stated that they were not aware of this practice, and that they weren't exactly thrilled about them. Gigabyte did promise us an official statement, but despite the fact that two months have passed, we have yet to receive one. Gigabyte's Dutch office informed us that they're dependent on Gigabyte Taiwan for an official statement, and we have obviously yet to receive anything official from Gigabyte Taiwan.

What we did receive was a pair of emails from Gigabyte's global PR manager during Christmas. The main criticism that he provided was that the kind of throttling we observed when using revision 2.0 of the B85M-HD3 motherboard could also occur when using motherboards made by other manufacturers, and that we were thus providing a biased report because we failed to consider other manufacturers' products.

Based on this response, and on other statements that Gigabyte Taiwan sent to other websites, it appears our primary argument has not been understood. In our article, we do not fault Gigabyte for producing a motherboard that throttles due to overheating. However, we do fault them for the fact that they're selling significantly altered products that offer less functionality and reduced quality, while having the exact same product codes / bar codes. We merely included the "throttle test" to demonstrate that the GA-B85M-HD3 revision 2.0 indeed substantially differs from revision 1.0, seeing as how the two boards clearly exhibit different behaviour in our test.

We have however taken steps to verify Gigabyte's implicit accusation of other motherboard manufacturers also making budget boards that throttle when having to provide power to a heavily loaded Core i7 processor: we are currently in the process of completing a test in which we compare budget H81 motherboards made by the four big brands with regard to this phenomenon. That said, we would again like to stress that this was not the focus of the original article, and that this does not diminish the severity of substantially changing the attributes of a product during its lifespan.
...
Since Gigabyte remains unwilling to comment on the situation, even after many reminders from our side, it is difficult to consider this case as definitely closed. The fact that a number of significantly changed products have been given a new name and new codes is certainly a positive thing. At the same time, however, the act of producing new revisions continues, and it remains unclear whether Gigabyte will in future once again significantly change existing boards in a negative fashion without adjusting their product codes.

With Gigabyte being unable or unwilling to clarify the situation, there's only one piece of advice we can give to consumers. If you're looking to purchase a Gigabyte motherboard – because of a review on Hardware.Info or another site, for example – you should before ordering always visit Gigabyte's website and verify whether new revisions have been released in the meantime. If so, you should use the photos to try and determine whether any significant changes were made. If you find that components have been removed or adjusted, but are unclear on what these changes exactly entail, you should contact Gigabyte and ask them to explicitly state what has been changed. Moreover, if you're going to order a product, you should first contact the shop where you're planning on purchasing the product, and ask them to check what revision they have in their warehouse and are going to send to you. While this is a fair amount of work, it is also for the time being the only way to ensure that you'll receive a motherboard that is identical to the motherboard on which you've read a review.

We will of course keep an eye on things. Should you, or somebody you know, happen to encounter a new revision that has identical codes but features substantial alterations, be they positive or negative, we would love to hear about it. We secretly hope that Gigabyte has, despite the lack of a formal response, learned from the resulting commotion, and will henceforth only sell minor revision and improvements under the same product code.

Quote:The reason for this test is our original article about Gigabyte's revision policy, in which we found out that while testing the 2.0 revision of the GA-B85M-HD3 the mosfets of the CPU power-supply heated up too much, causing the CPU to clock down. Gigabyte defended itself by claiming similar scenarios also occured with other manufacturers, ignoring the fact that the main point of our article was not about the throttling issue, but instead the re-use of productcodes. But enough about that; Gigabyte's claim that this could also happen with other manufacturers was more than enough reason to conduct a new test of budget motherboards.
...
The results of our VRM-test are shown in the table below. As we wrote on page 3, it's the maximum temperature measured during 10 minutes of H.264 video-encoding workload. We performed the test with an Intel boxed cooler as well as a Cooler Master Hyper 212 Evo towercooler. Temperatures above 90 °C are in the critical zone in our opinion and are shown in red. Values below 80 °C on the other hand are excellent and are shown in green.

The good news is that none of the motherboards are throttled, something we previously saw with the B85M-HD3. At the same time we noticed that the three motherboards that use three phases and two mosfets per phase (the two ASUS boards and the Gigabyte H81M-S2V) are in the critical zone when paired with a tower cooler. It wouldn't come as much of a surprise to us that the CPU would throttle when subjected to a high CPU-load for more than 10 minutes, due to the VRMs heating up too much.

We can also see that ASRock's decision to put a four-phase powersupply on the H81 does have effect. The temperatures of the mosfets remain low in all cases. There's also some appreciation from our side for the Gigabyte H81M-HD3. This board is the only one that has an enhanced turbo and is therefore under more load during this test yet its temperatures remain low. This is definitely a case of the use of better components and/or better use of the phases.