Gigabyte Gives Lifetime Warranty to X79 Board Corrective BIOS Users, Isolates Problem

[btarunr]

On Tuesday, a major problem associated with Gigabyte's X79-UD3, X79-UD5, and G1.Assassin 2 motherboards came to light after a Gigabyte press-release, where enthusiasts subjecting their boards to voltage-assisted overclocking with stress-testing, ended up with burnt CPU VRM. Till that press-release, the scale of the problem was not known. Gigabyte announced remedies to existing owners, which included either updating their motherboards' BIOS to the latest "F7" version posted on the company website, or sending their boards dead or alive for free replacements.

We're getting to know now that to all those who opt to keep their boards and update their BIOS, Gigabyte is offering a lifetime product warranty, an extension of the limited warranties their products come with. Gigabyte's own version of what went wrong with these motherboards is that it shipped several of its motherboards with bad BIOS firmware that did not have "overclocking limits", which motherboards by other manufacturers did. This claim means that "japan0827", the overclocker from XFastest community who ended up with a burned X79-UD3 that he posted on YouTube, would have been running his setup way off spec, electrically.



View at TechPowerUp Main Site

 

[qwerty_lesh]

I highly doubt this is an Isolated incident because one person was running their board out of spec.

Seen this happen at optimised defaults... read replies from other forum members with the same experience too...

brand new boards with failing VRM on defaults, that is not an 'out of spec' problem that is shoddy SMDs, simple.

 

[HTC]

Gigabyte's own version of what went wrong with these motherboards is that it shipped several of its motherboards with bad BIOS firmware that did not have "overclocking limits", which motherboards by other manufacturers did. This claim means that "japan0827", the overclocker from XFastest community who ended up with a burned X79-UD3 that he posted on YouTube, would have been running his setup way off spec, electrically.

So they're putting the blame on the overclocker despite admitting they're @ fault: figures ...

I thought there were cases where the user who got his board fried wasn't OCing: am i mistaken in this?

 

[NC37]

Lifetime...with a limited pool of enthusiasts buying them, prolly wouldn't hurt them much. Plus they'll likely get good PR from announcing they'll back their products for life.

Could hope this kicked off a lifetime warranty spree among other makers, but to keep the parts on hand to fulfill that wouldn't work. Gotta be a clause in it where they'd just give you a new board. Then there is the trouble of lifetime warranties getting popular. It sounds like a good idea, but I remember over the years seeing products with lifetime warranties being cheaper made than those with limited.

 

[R_1]

Highly doubt that explanation, because +12 phase mobos have VRM&PWM rated at 450Watt at +35 Watt per vCPU phase. No way i7-E is getting near that power consumption in Prime95, regardless overvolting.

 

[qwerty_lesh]

I thought there were cases where the user who got his board fried wasn't OCing: am i mistaken in this?

multiple instances of this happening yep, and on the UD7's.. the model which wasnt mentioned in the CN press release at all even tho every GB x79 got F7 bioses to "fix" this problem..

 

[ensabrenoir]

Yeah i'll take two

Free board replacement......how about my $1000 cpu?

 

[MetalRacer]

From Gigabytes press center.
http://www.gigabyte.us/press-center/news-page.aspx?nid=-869

 

[FreedomEclipse]

Plus they'll likely get good PR from announcing they'll back their products for life.

+1

Usually Gigabyte like to sit back and deny everything, If it was just the one or small handful of related incidents then thats what they probably would have done and ultimately blamed the user. the fact that it is an design flaw with their boards (possibly rectifiable with a bios update???) and not just user error means they have to come out of the closet do a pirouette around the room a few times before backing off slowly with its head bowed.

Its a bigger problem then Gigabyte would like to admit to itself

 

[qwerty_lesh]

+1

Usually Gigabyte like to sit back and deny everything, If it was just the one or small handful of related incidents then thats what they probably would have done and ultimately blamed the user. the fact that it is an design flaw with their boards (possibly rectifiable with a bios update???) and not just user error means they have to come out of the closet do a pirouette around the room a few times before backing off slowly with its head bowed.

Its a bigger problem then Gigabyte would like to admit to itself

I don't agree with you on your claim that they sweep problems under the rug. years prior to the Intel Cougar Point issues, Gigabyte had problems with their line of GTX275 graphics cards, and gave excellent support to their customers to resolve the issues (free physical repairs and video card bios updates, case by case basis) at least in my country, they also admitted to the issues with the 460/465 2D/3D clock problems that Fermi had at the time, and provided support to resolve the problem.

My qualm is specifically because of their rushed to market patsburg boards being buggy and problematic due to poor uefi bios and bad QA for SMD parts.

the upside from all this though is that we'll see a prompt change to Rev 1.1/2.0 boards which (should) have better surface mount devices on them. and well if not, at least the other big brand patsburg boards have proved so far to be very reliable (all other then GB and intel from what i've dealt with so far), so either I'll just be using those in the mean time or from hereon out.

 

[FreedomEclipse]

I don't agree with you on your claim that they sweep problems under the rug.

theyve been doing it for a while. but more with isolated incidents then big ones like this....

One of the major ones i remember is when they were shipping some of their 5850s without 2D clocks so the card would be running at full speed all the time. when they were contacted by a few users about the issue (I think there were at least 2-3members here in this forum that had the same problem) and asked about a bios update instead of editing the bios themselves, they strictly denied that it was an issue at all and that it was the way they designed the card.

There are numerous issues TPU members have had with Gigabyte support. Maybe they were pretty decent back in the GTX275 days but quite a few people on these forums would say their level of custom support has dropped quite substantially.

you are welcome to disagree all you like. Maybe you were just lucky you got someone who actually gave a crap when you contacted their support, many havent been so lucky and its got to a point where its starting to effect and weigh heavily on peoples decisions when it comes to deciding what hardware to buy.

Im sure Gigabyte can make some good hardware. Their EP45-UD3R 775 mobo was a seriously good overclocker that took many high end C2Qs to 4Ghz & above. but if support is lacking then that becomes a major fly in the ointment

 

[cadaveca]

Highly doubt that explanation, because +12 phase mobos have VRM&PWM rated at 450Watt at +35 Watt per vCPU phase. No way i7-E is getting near that power consumption in Prime95, regardless overvolting.

You greatly underestimate SB-E's power consumption.



Just FYI, that 130W TDP rating...isn't for the whoel CPU package...it's JUST for the cores.

Not my pic, but:

At a paltry 4.4 GHz, I'm pulling well north of 275W. At 5GHz and above.. it's totally possible that the power consumed is too much for the VRM.


However, it IS hard to beleive that on a standard all-in-one cooler like the user @ XFastest used, was able to deal with that heatload.


What's really interesting to see is that the GA-X79-UD7 is now on the list as well.

 

[Steven B]

Nice pic, I am glad you read it

Yea and it is also interested to see that this really isn't a big deal to them, becuase in all actuality it really isn't. If you read the GB blog you would see that Tim only recommended flashing if you not running at stock. Which to me means at stock the boards are fine. They didn't offer lifetime for USA. I think they only did that in TW (in TW warranty isn't 5 years out of the box tho), because those users at xfastestest cried about GB TW warranty service. Surprising is that GB even said they will cover warranty if you OC and break your board, b/c the truth is OCing anything voids its warranty, mobo included.

one thing tho, i do want to say to GB users, if you flash BIOS F7 you cannot back flash currently. Meaning it will flash your main and backup BIOS to F7, and you canot flash back from F7 to previous BIOS. This really doesn't affect OCing at all, my max OC under air is the same, and shoudl be under LN2. GB has removed 0% LLC, as this is what they said was a real issue. i guess raising voltage 60mv under load is a bad thing lol. The new F7 BIOS doesn't lower OTP or OCP, it does however lower maximum SF from 400khz down to 320khz. Which really doesn't affect the UD7 at all I can get teh same OC at 180khz or 500khz. Funny thing is that they didn't lower the stock SF from 220khz, they left it raised. When the boards first released it was at 200, and then after a few updates it went to 220, I was expecting them to lower it, but they didn't. Does SF have any affect on your UD5?

Bottom line: If you want to unnecessarily hurt a company and the entire platform lineup, make a video of the board blowing up and post it on every forum you can find, but don't post it in the motherboard or Intel section, please this next part is key, find where the site has its most hits, usually news section, and post it there! Then give it a name, like "R4E self-ignites and burns down house in a blaze"(For this one you will need to go to your neighbor's house and video tape through a window). Or how about "BigBang self-ignites and makes bang sound, one of its bullets hit my child in the eye!" (You can probably pay a 12 year old to be the target) Or "Assassin 2 self-ignites and its gun shot me in the ass!" (Make sure you get a tripod) You might even get an upgrade, nicer board, nicer CPU, make sure you ask for one of those $500 64GB memory kits!. GB has already been taken care of, who is next? ASUS? ASRock? Or MSI? Seriously i can make any board do what that UD3 did, even at stock.

 

[cadaveca]

I'll be checking out the new BIOS tomorrow. I did my initial testing with F4, as it was the most recent at the time, but given the situation, it seems like I should be using F7 for the review.

 

[LiveOrDie]

Who's watched the video and i missing some think a overclocker with not even one temp motioning program open stress testing his CPU under high load

 

[Delta6326]

I'll be checking out the new BIOS tomorrow. I did my initial testing with F4, as it was the most recent at the time, but given the situation, it seems like I should be using F7 for the review.

Review is the only thing I heard!

 

[HTC]

I'll be checking out the new BIOS tomorrow. I did my initial testing with F4, as it was the most recent at the time, but given the situation, it seems like I should be using F7 for the review.

Suggest you review it first with the BIOS you have now and again with the new BIOS. Regardless if you find any differences in performance with the new BIOS, i believe you should add them to the review.

This is a special case because, ordinarily, reviewers don't remake a review because of new BIOSes but, then again, BIOSes aren't to blame for fried motherboards.

This way, whether or not the new BIOS has any sort of significant impact on performance, readers of this review will know what to expect.

 

[cadaveca]

From the way the problem has been reported by Gigabyte, I do not expect there to be any performance differences.


Intel spec sheets say VRM is supposed to send signal to CPU to initiate CPU throttle if VRM exceeds safe temperatures. It makes sense to me that all that they have adjusted is when that point is triggered, as well as possibly the tweaks listed by StevenB above. None of these should have any impact on performance during normal operation, even when overclocked. I will check anyway. Really, to me, all of this is a non-issue.

 

[qwerty_lesh]

theyve been doing it for a while. but more with isolated incidents then big ones like this....

One of the major ones i remember is when they were shipping some of their 5850s without 2D clocks so the card would be running at full speed all the time. when they were contacted by a few users about the issue (I think there were at least 2-3members here in this forum that had the same problem) and asked about a bios update instead of editing the bios themselves, they strictly denied that it was an issue at all and that it was the way they designed the card.

There are numerous issues TPU members have had with Gigabyte support. Maybe they were pretty decent back in the GTX275 days but quite a few people on these forums would say their level of custom support has dropped quite substantially.

you are welcome to disagree all you like. Maybe you were just lucky you got someone who actually gave a crap when you contacted their support, many havent been so lucky and its got to a point where its starting to effect and weigh heavily on peoples decisions when it comes to deciding what hardware to buy.

Im sure Gigabyte can make some good hardware. Their EP45-UD3R 775 mobo was a seriously good overclocker that took many high end C2Qs to 4Ghz & above. but if support is lacking then that becomes a major fly in the ointment

Fairy nuff, support quality may also vary vastly in different countries. which could account to mostly why I disagreed, since I never caught them performing such shady practices in Aus in the past. :\

 

[HTC]

From the way the problem has been reported by Gigabyte, I do not expect there to be any performance differences.


Intel spec sheets say VRM is supposed to send signal to CPU to initiate CPU throttle if VRM exceeds safe temperatures. It makes sense to me that all that they have adjusted is when that point is triggered, as well as possibly the tweaks listed by StevenB above. None of these should have any impact on performance during normal operation, even when overclocked. I will check anyway. Really, to me, all of this is a non-issue.

I don't really understand the technical aspect of this but neither do most people. Just because, in theory, A should not affect B doesn't make it so and a good way to find out is to review it.

I do realize this means more work for you but you should also realize that your efforts could help many people deciding whether or not any of the affected boards are a good buy.

 

[cadaveca]

Well, news came out today that F7 BIOS has been used to set quite a few benchmarking WR.

http://valid.canardpc.com/show_oc.php?id=2169786

 

[EarthDog]

Who's watched the video and i missing some think a overclocker with not even one temp motioning program open stress testing his CPU under high load

I didnt see the video but if he was subzero cooling.. there isnt a point. And wtih benchmarking there is no stress testing outside of the benchmark completing. WAY different methods there. Not to mention, most applications either cant monitor VRM temps or have no clue what temps are the VRMs...

@ cad - care to attach that CPuz screeny to TPU? I cant see that site at the office!

Well, news came out today that F7 BIOS has been used to set quite a few benchmarking WR.

http://valid.canardpc.com/show_oc.php?id=2169786

Thats good news. So just keep the VRM temps down is the moral of the story here. Much ado about nothing.

Testing both bios for performance differences is nothing but a waste of time.

 

[cadaveca]

@ cad - care to attach that CPuz screeny to TPU? I cant see that site at the office!

Sure.

 

[HTC]

Well, news came out today that F7 BIOS has been used to set quite a few benchmarking WR.

http://valid.canardpc.com/show_oc.php?id=2169786

Where does it say it's the F7 BIOS being used?

If true then it whatever the problem was, Gigabyte seems to have fixed it.

EDIT

Testing both bios for performance differences is nothing but a waste of time.

Ordinarily yes but this isn't an ordinary thing.

If this new BIOS does indeed fix the problem while keeping the performance of the old (if not increasing it) then great but if the new BIOS fixes the problem @ a performance cost, then it's a whole new story.

 

[EarthDog]

I would re-read cad and SteveB's posts... They explain what the issue is. Which should not affect performance at all.



@ CPUz record you linked, that was done on a RIVE, not a Giga board.