Gigabyte Issues Statement Regarding SOC Voltage on AMD's AM5 Motherboads Under 1.3V

[TheLostSwede]

We would like to address the recent media reports regarding the SOC Voltage exceeding 1.3 V on GIGABYTE's AMD AM5 motherboards especially when EXPO is enabled in the latest beta BIOS.

GIGABYTE respects and appreciates media's support and favor to GIGABYTE motherboards for long time. In terms of the SOC Voltage measurement, the authentic measurement point and method is critical since the SOC Voltage will differ by that. The CPU internal SOC Voltage (SVI3 interface) is the most crucial indicator to motherboards, and, in general, the PWM Output Voltage will be higher than the CPU internal SOC Voltage (SVI3 interface) due to various physical factors.




Users can use HWiNFO software to accurately monitor the CPU internal SOC Voltage (SVI3 interface), which is indicated as "CPU VDDCR_SOC Voltage (SVI3 TFN)" in HWiNFO. This value is reported by the internal sensor of the CPU to show the true voltage. GIGABYTE uses HWiNFO to monitor the SOC Voltage on AM5 motherboards and it is under 1.3 V, which is officially confirmed by AMD. The full lineup of GIGABYTE AM5 motherboards follows AMD's latest SOC Voltage 1.3 V Guideline when EXPO is enabled. Please refer to the video below for more details.

GIGABYTE consistently works closely with AMD and follows AMD's guidelines to fully comply with AMD's official specifications and requirements. We appreciate all the attention and support from media and customers. It has always been our motivation to deliver more innovative and breakthrough products to our customers. As a leading motherboard manufacturer, we will continue to provide top-quality products for users to enjoy the best PC experience.

View at TechPowerUp Main Site | Source

 

[Bomby569]

am i reading this right, basically saying some Youtubers are doing it wrong

 

[TheLostSwede]

am i reading this right, basically saying some Youtubers are doing it wrong

AMD is measuring the Voltage at the SVI3 interface, which is apparently unfamiliar to people, hence why there seem to be some misunderstandings.

 

[user556]

So, yes, they're saying the youtubers are doing it wrong.

 

[R0H1T]

This is getting ridonculous, while the 1.3v would be a "guideline" & just that (alone) won't kill your CPU. AMD would give themselves and their partners more leeway without getting sued! Also no one can say for sure what voltage is absolutely safe for your CPU, every chip is different & that limit may vary a lot in some cases ~ so don't take it like anything above 1.3v will absolutely kill your chip

It's just that the chip will degrade faster around those limits, but again your mileage will vary. And yes running pime95, OCCT or any power virus around those limits will degrade your CPU faster!

 

[Zubasa]

am i reading this right, basically saying some Youtubers are doing it wrong

Yes this one apperently.
ASUS is NOT the ONLY ONE: Gigabyte - EXPO and SoC Voltages Before & After the BIOS Update - Hardware Busters (hwbusters.com)

 

[TheLostSwede]

A more technical explanation of the SVI3 interface.

AMD’s Scalable Voltage Interface 3 (SVI3) for Dummies and how to check if your AM5 Mainboard is SAFE | igor´sLAB

Terrible SoC voltages everywhere! AMD’s new horror! Say some YouTubers. But to learn more about this problem and understand why at first glance everything seems different than it actually is…

www.igorslab.de

 

[Zubasa]

Also this shouldn't apply to GN.
Twitter 上的 GamersNexus："@ThatGuyJD98 @dainluke @Buildzoid1 @HardwareBusters Yes, there were at least 3 locations measured for the ASUS board." / Twitter

 

[mama]

Putting aside the questions as to whether the measurements were taken correctly or whether there is compliance with the AMD specification, Ari tested one board. That isn't a scientific methodology. One apple does not an apple tree make. As an aside, my board has been solid. I have monitored the SOC voltage with HWInfo64 since the reports emerged and it has remained below 1.28V throughout.

 

[asdkj1740]

am i reading this right, basically saying some Youtubers are doing it wrong

asus also addresses the issue today. bing=director of asus rog team
so for the short answer, yes they are all wrong about the new bios still not properly locked soc voltage at 1.30v max, and failed to understand which point of soc voltage amd has been talking about.

 

[Outback Bronze]

Im starting to feel sorry for AM5 users here.

All these new bioses and disabling EXPO etc.

Why the hell didnt AMD release the damn CPU being well tested instead of customers being the lab rats.

IMO very unprofessional.

 

[Bomby569]

Also this shouldn't apply to GN.
Twitter 上的 GamersNexus："@ThatGuyJD98 @dainluke @Buildzoid1 @HardwareBusters Yes, there were at least 3 locations measured for the ASUS board." / Twitter
View attachment 296486

Buildzoid seem to disagree with that statement

 

[N3utro]

That's the difference between real pros who know what they are talking about and wannabe "experts" who brag a lot on youtube for views and ad money on which they rely on to survive like gamer nexus.

 

[Zubasa]

Buildzoid seem to disagree with that statement.

Look at the time stamps.
Buildzoid had concenrs on how GN measured the voltage. Either way he states that GN's findings are correct.

GN replied in the same thread with how they measured multiple spots on the Asus boards including the back of the socket.

Also both of them are in the Buildzoid's twitter post refuting Hardware Buster's claim that Gigabyte's beta bios did not lower SOC voltage.
This what Gigabyte is refuting.
ASUS is NOT the ONLY ONE: Gigabyte - EXPO and SoC Voltages Before & After the BIOS Update - Hardware Busters (hwbusters.com)

GN DID NOT made statements of Gigabyte' bios, but how Asus' bios at the time did not lower SOC Voltage.
Asus later released 1610 which users report that SOC voltage is lowered.

 

[Bones]

I know Buildzoid knows alot about such things - He, like anyone else isn't perfect but at the same time I do trust his take on it as being an honest observation on his part.

Just know at least some of this is nothing more than PR from the board makers scrambling to cover their asses because if they don't, they stand to lose big. To be fair, as to how much is and isn't, that's subjective to all the variables in the hardware itself AND the one's on both sides of it saying it's "This" or "That", be it Buildzoid, GN, corporate reps/testers/engineers and so on.

I do know one thing, if the board makers don't get it right it's just going to continue, leading to even more controversy and even they know that's the truth of it.
I'm glad I don't have a DDR5 based build to worry about all this.

That's the difference between real pros who know what they are talking about and wannabe "experts" who brag a lot on youtube for views and ad money on which they rely on to survive like gamer nexus.

I can't agree 100% with you on this.
It's true the Youtubers don't have access to all the equipment the makers do in their lab or experience in doing it for a living but at the same time it's not like the YT guys are totally dumb either - Sometimes they nail it hard when they do and at this point the evidence of these problems proves it's fact, not fiction.
You'd think the "Experts" for the board makers would have gotten it right but so far, no good.

The one thing that can't be disputed is there is a problem and ATM it does lay at the feet of the board makers themselves regardless of who's testing what and how it's being tested.

 

[Bomby569]

i trust the media less and less, especially Ytubers that talk to much, to soon, and love drama because that's what drive views. GN is not above that, clickbait, crazy faces thumbnails, hyper on the drama. I trust Buildzoid, the man really knows a lot of this stuff and we doesn't care about clicks, you just have to watch his content for this to be apparent. Lots of salt, lots and lots of salt.

 

[Six_Times]

If its not lying about core counts, (bulldozer) or boards bricking, now this...

 

[ZoneDymo]

Im starting to feel sorry for AM5 users here.

All these new bioses and disabling EXPO etc.

Why the hell didnt AMD release the damn CPU being well tested instead of customers being the lab rats.

IMO very unprofessional.

I'm sorry but it seems you don't have a clue what this is about

 

[Outback Bronze]

I'm sorry but it seems you don't have a clue what this is about

Please forgive my ignorance. Just venting frustration over the whole situation.

 

[Slizzo]

My AM5 board has been perfect so far, 7950X3D still kicking from launch day.

I don't use EXPO though, I have been using the Memory Try It! profiles that MSI provides.

 

[judojedi]

YouTubers are no different than the MSM. My gigabyte board and 7800x3d have been doing just fine with expo enabled. Not once has the soc gone over 1.21v.
This is mostly hype

Im starting to feel sorry for AM5 users here.

All these new bioses and disabling EXPO etc.

Why the hell didnt AMD release the damn CPU being well tested instead of customers being the lab rats.

IMO very unprofessional.

There is nothing to feel sorry about, 95% of the cpu's are doing just fine, mine included

 

[Kapone33]

The most humorous thing about this thread is Gigabyte used to be the worst when it came to pumping voltage. If your cooling was good enough Gigabyte was the snappiest AM4 boards but also had the most infuriating quirks and blue and black screens. Now Asus has egg all over their face for trying to justify the premium we pay for boards that don't even equate to TRX40 but cost in some cases double.

 

[ir_cow]

Why the hell didnt AMD release the damn CPU being well tested instead of customers being the lab rats.
IMO very unprofessional.

Remind you of Zen 1 by any chance

 

[user556]

Belated comment: One way to test accuracy of the internal SVI voltage measurements is to operate the CPU at lowest power level while measuring the same voltages externally. This way the socket, interposer and die volt drops are minimised so the external and internal readings should be a very close match.