News Ryzen Burnouts Force Asus to Add Overclock Warnings to EXPO Features

[Admin]

Asus is one of the first motherboard vendors to add a notice stating that memory technologies like XMP / EXPO are considered overclocking, with clear warranty implications/

Ryzen Burnouts Force Asus to Add Overclock Warnings to EXPO Features : Read more

 

[domih]

In other words, CPU and motherboard manufacturers want to be able to use XMP and EXPO as marketing tools while denying the customer support of these functions as soon as they become a $$$ problem (e.g. CPU and/or motherboard destroyed.)

French have an expression for that: "They want the butter, the price of the butter and the ass of the farmer's wife".

The conclusion for end users is simple: (1) do NOT use XMP or EXPO. (2) Protest against the practice as it is now borderline false advertising.

 

[Sergei Tachenov]

I've been discussing this message on Reddit recently, and we've noticed in the discussion that nowhere it says anything about warranty. It's more along the lines of "not guaranteed to work" rather than "may void warranty". And there's a huge difference. One thing is being denied RMA if if EXPO simply doesn't work, another thing is being denied RMA if your CPU blows up because you've turned EXPO on.

 

[TerryLaze]

Overclocking is always overclocking even if it's not clearly stated and it will always void your warranty, that second part is always stated somewhere both on the cpu site as well as on the mobo site and in the manuals as well.

 

[btmedic04]

I've been discussing this message on Reddit recently, and we've noticed in the discussion that nowhere it says anything about warranty. It's more along the lines of "not guaranteed to work" rather than "may void warranty". And there's a huge difference. One thing is being denied RMA if if EXPO simply doesn't work, another thing is being denied RMA if your CPU blows up because you've turned EXPO on.

Unfortunately by saying overclock/overclocking, it will give assus enough wiggle room to deny any rma since they apparently would rather shove more voltage into vsoc instead of actually optimizing the bios.

 

[geoiv]

In other words, CPU and motherboard manufacturers want to be able to use XMP and EXPO as marketing tools while denying the customer support of these functions as soon as they become a $$$ problem (e.g. CPU and/or motherboard destroyed.)

French have an expression for that: "They want the butter, the price of the butter and the ass of the farmer's wife".

The conclusion for end users is simple: (1) do NOT use XMP or EXPO. (2) Protest against the practice as it is now borderline false advertising.

Yeah, I had a thread going on LTT forum about that. I really wish reviewers would first test CPUs/RAM at JEDEC, before they show off their potential overclocking capabilities.

 

[Venivik]

My ROG Strix X670E-A motherboard says 6400mhz ram speed right on the box.

I don't give a damn what they say, if I can't run my Ram as advertised and the board blows up.. they're going to replace the board or I'm going to file a lawsuit regardless.

This is no different than companies continuing to use don't tamper stickers even though they are completely illegal to be used as a reason to deny somebody warranty coverage.

Do motherboards advertise overclocked CPU speeds? Nope. They're going to have to stop advertising Expo speeds before denying warranties.

Don't be scared of Corporations, never forget you're the hand that feeds.

 

[Deleted member 2838871]

I agree with all of the above.

I built mine last week with parts ordered from Newegg... and did nothing but turn EXPO on. 3 days later both cpu and mobo blew up. Within 24 hours the news had broke that it was a widespread problem.

I ordered the exact same parts from Amazon for next day delivery... and computer was back up and running. Then I did an RMA with Newegg citing the news of exploding parts and they sent me a new cpu and mobo that I just received yesterday and then returned to Amazon.

$1000 in limbo for a week but it was better than having no PC for a week. The whole thing is laughable because AMD uses EXPO in their advertising and you better believe I am going to use it.

I don't much care about OCing the CPU but I want the advertised ram speed. Newegg made the RMA easy... and I'm glad I didn't have to go through AMD or MSI.

 

[DookieDraws]

I've been discussing this message on Reddit recently, and we've noticed in the discussion that nowhere it says anything about warranty. It's more along the lines of "not guaranteed to work" rather than "may void warranty". And there's a huge difference. One thing is being denied RMA if if EXPO simply doesn't work, another thing is being denied RMA if your CPU blows up because you've turned EXPO on.

Mine didn't at first, but it CLEARLY does now after updating the BIOS. I have the ASUS ROG Strix B650E-F Gaming WiFi motherboard, running BIOS 1409, which was posted April 21. As of today, the latest stable BIOS is 1414. There are some Beta BIOS', but I won't touch a beta BIOS. And I like to wait a bit before updating.

Just built this AM5 PC in mid January. After reading that AMD had confirmed that DDR5-6000 will be the "sweet spot" for Ryzen 7000 CPUs, that's exactly what I purchased. Even went on AMD's forums and discussed this BEFORE buying my RAM.

Although I've never done it myself, but plan to do so, I was always under the impression that undervolting was safe.

 

[hotaru.hino]

A thing that annoyed me is AMD kept touting PBO as a nifty feature and whatnot... even though anything that happens to the processor while PBO is in use is not covered under warranty. If this were Intel, there would've been torches and pitchforks, but for some reason I never really saw this with AMD.

Yes of course you could say "but how does AMD know I used PBO?" It doesn't matter. The fact the company is proudly advertising a feature that isn't supported if poop hits the fan is a problem.

 

[bigdragon]

EXPO and XMP are huge selling features for motherboards and RAM. All of the RAM I've bought in at least the past 10 years has required the use of EXPO/XMP profiles to get the advertised performance. The process of using those profiles is just a selection menu. There's no guesswork. No trial and error. No memtest or stress testing. Motherboard and RAM vendors advertise EXPO/XMP in lists that confer an understanding of tested compatibility.

I don't agree with what Asus is doing here. If Asus boards are causing physical damage to AMD CPUs, then the motherboard is supporting something it clearly doesn't actually support! This is a flaw with the product and its marketing. The same is true for any other motherboard vendor advertising capabilities their boards cannot safely support. We're not talking about manual memory timings and voltages common to overclocking -- we're talking profiles and QVL publications! I'll concede that damage could happen with manual overclocking, but profiles that are supposedly tested should be select and forget.

 

[hotaru.hino]

I think we're pointing the crosshairs at the wrong people. As far as the motherboard manufacturers are concerned, as long as they deliver the voltages required by the memory module and the signal integrity is fine from the DIMM slot to the CPU socket, they've done their job. Of note, JEDEC officially goes up to DDR5-8000 (of which at the time of this post, there were only five listings on PCPP for such memory).

Whether or not the CPU itself can manage that? I don't think that's for motherboard manufacturer to really care about. After all, if AM5 is supposed to be supported for as many generations as AM4 (if not more), then Zen 5 or Zen 6 CPUs may be able to handle it at that point.

Considering AMD needed to issue microcode updates to tame power delivery, this tells me they were really lax with their validation of what the CPU can handle. That or they didn't issue clear enough limitations to board makers.

 

[Kamen Rider Blade]

Maybe we should stop using ASUS MoBo's and use one of the other vendors.

GigaByte, MSI, ASRock all exist along with a bevy of smaller vendors.

If ASUS isn't honoring their end of the bargin, why buy from them?

 

[Sergei Tachenov]

I think we're pointing the crosshairs at the wrong people. As far as the motherboard manufacturers are concerned, as long as they deliver the voltages required by the memory module and the signal integrity is fine from the DIMM slot to the CPU socket, they've done their job.

Please note, however, that this issue is mostly about the SOC voltage, and it is not required by RAM in any way. It's just the motherboard manufacturers choosing some values they believe are safe and stable. And Asus decided to err on the higher side here.

So at least both AMD and motherboard vendors are at fault here. As AMD should have researched this and told motherboard vendors what voltage to use, and motherboard vendors should have double-checked with AMD that the voltage they chose is safe.

Also note that this all happens with SOC voltage set to Auto, which, combined with the fact that EXPO profiles don't have any value for that voltage, is very confusing to the user. You enable a setting that should only affect RAM, and then your motherboard suddenly decides to pump more voltage into the CPU? This may make sense to more technically inclined users, but not to an average gamer.

 

[ET3D]

"Force Asus"? Really?

That's possibly the worse response ASUS can have to this. Nothing in that text tells the user "your CPU might burn if you use this feature". It's an extremely weak attempt for ASUS to cover their ass.

If a memory advertises EXPO and the motherboard advertises support for that memory speed (even if that speed is marked as overclocked), then I think it should be fully expected to work without causing any damage to anything.

 

[Ogotai]

I really wish reviewers would first test CPUs/RAM at JEDEC,

might want to go look at the reviews at anandtech.com, as this is what they do for their tests

 

[TerryLaze]

A thing that annoyed me is AMD kept touting PBO as a nifty feature and whatnot... even though anything that happens to the processor while PBO is in use is not covered under warranty. If this were Intel, there would've been torches and pitchforks, but for some reason I never really saw this with AMD.

Intel also advertises XMP and their tool XTU, and also doesn't really advertise that it's voiding the warranty, they do state it semi-hidden still easy enough to find if you are a thorough reader.

Intel users are just as baffled as AMD users are now when you tell them that what they do due to advertising isn't covered by warranty.

Yes of course you could say "but how does AMD know I used PBO?" It doesn't matter. The fact the company is proudly advertising a feature that isn't supported if poop hits the fan is a problem.

Because a lot of people have no idea that it's not warrantied they just openly tell the phone operator during RMA the settings of their bios...

 

[Deleted member 2838871]

Because a lot of people have no idea that it's not warrantied they just openly tell the phone operator during RMA the settings of their bios...

I honestly didn't know either. What I do know is when I was reading up on the Ryzen chips the "AMD EXPO blah blah" was plastered all over the web page.

At least mine blew up on day 3 and all I had to do was return to Newegg without having to deal with AMD or MSI.

 

[kerberos_20]

that warning is useless if asus sells overpriced broken mainboards which overvolts vsoc when you enable expo, not even OCP works

View: https://www.youtube.com/watch?v=kiTngvvD5dI

its worth watching

 

[RichardtST]

Yeah, I'm afraid that ASUS is going the way of Samsung and losing touch with their consumers. "Gimme all your money and we'll give you all our flashy crap that breaks as soon as you use it." No, XMP and EXPO cannot be considered overclocking. It states pretty much everywhere, rather blatantly, that it is all supported and blessed by everyone involved. I've had really good luck with ASRock motherboards lately. I think I'll stick with them. This is a cheap cop-out by ASUS. They've lost my dollars. I build custom for my company, and every little bit of "safe" speed I can get pays for itself. Can't have people sitting around twiddling thumbs waiting for compiles!

 

[A Stoner]

How would I know if my processor has been damaged but not to the point that it is currently failing due to this 'bug? I have a Ryzen 9 7950X3D processor and am using 6400 XMP/EXPO memory modules that request a 1.400 volt for the memory. I just updated to the newest BIOS and it still allows the 1.400 volts on the memory. I think this is a different voltage than what is being claimed to cause the damage?

 

[RichardtST]

How would I know if my processor has been damaged but not to the point that it is currently failing due to this 'bug? I have a Ryzen 9 7950X3D processor and am using 6400 XMP/EXPO memory modules that request a 1.400 volt for the memory. I just updated to the newest BIOS and it still allows the 1.400 volts on the memory. I think this is a different voltage than what is being claimed to cause the damage?

According to the GN video, If you are on an ASUS mobo, you have already damaged it and are just lucky that it has not yet died. The OK-to-not-OK voltage difference is so tight as to be utterly ridiculous. Even as a percentage. Basically, if you have a 7xxx3D chip, TURN OFF EXPO immediately. Sorry. That's just what it is. AMD (and asus!) screwed the pooch on this one. That said, if you are on a non-asus board, and you can see your memory voltage with expo turned on is 1.35 or less, then you are probably just fine. I wouldn't risk it, but that's just me... Either way, I'd still be quite pissed.

 

[TJ Hooker]

Please note, however, that this issue is mostly about the SOC voltage, and it is not required by RAM in any way. It's just the motherboard manufacturers choosing some values they believe are safe and stable. And Asus decided to err on the higher side here.

SOC voltage is the memory controller voltage. It makes sense it may have to be tweaked for running high speed RAM.

 

[kerberos_20]

SOC voltage is the memory controller voltage. It makes sense it may have to be tweaked for running high speed RAM.

no, soc is for whole soc, that bigger chiplet, most mobos stick with 1.2v, asus pushing 1.4v
as for expo you just need higher VGGD IOD when IF goes near 2GHz, which is still around 1.2v (slightly below vsoc)

 

[TJ Hooker]

no, soc is for whole soc, that bigger chiplet, most mobos stick with 1.2v, asus pushing 1.4v
as for expo you just need higher VGGD IOD when IF goes near 2GHz, which is still around 1.2v (slightly below vsoc)

Sure, Vsoc may not be just the memory controller, but it includes the memory controller, yes?

"SOC voltage - system on a chip voltage; responsible for the voltage related to the memory controller." https://www.techpowerup.com/review/amd-ryzen-memory-tweaking-overclocking-guide/2.html

"Specifies the SoC/Uncore voltage (VDD_SOC) in mV to support memory and Infinity Fabric overclocking.” https://dlcdnets.asus.com/pub/ASUS/mb/Socket AM5/ROG CROSSHAIR X670E EXTREME/E20466_ROG_CROSSHAIR_X670E_Series_BIOS_manual_EM_WEB.pdf?model=ROG CROSSHAIR X670E EXTREME

"Allows you to specify the SoC/Uncore voltage (VDD_SOC) to support memory and Infinity Fabric overclocking" https://download.asrock.com/Manual/Software/AMD X670/Software_BIOS Setup Guide_English.pdf

Also, given that VDDG cannot be higher than Vsoc, wouldn't that mean you might need to increase Vsoc just to get the VDDG you want?

The person I responded to said Vsoc is not related to RAM in any way, that is clearly not the case.