News AMD's 'Ryzen Burnout' Fixes Reportedly Plagued With Bugs

[Admin]

A new AMD microcode up is now out that was expected to fix all the issues surrounding recent Ryzen 7000 CPU deaths. But sadly, the microcode up appears to be riddled with bugs.

AMD's 'Ryzen Burnout' Fixes Reportedly Plagued With Bugs : Read more

 

[Deleted member 2838871]

In short, user reports have cropped up revealing dead Ryzen 7000 CPUs that melted in their sockets due to SoC voltages operating at unsafe voltages (which is specified as anything over 1.3v right now), and bugs surrounding AMD's CPU temperature control system, which can cause Ryzen 7000 CPUs to go a beyond its safe thermal specifications.

I have had no further issues since rebuilding with a new CPU and mobo on April 23rd and updating to a new BIOS on April 28th which locked SOC voltage at 1.3v. EXPO has been enabled the entire time.

Still waiting on the AGESA update.

On the bright side, AMD is reportedly testing a new 1.0.9.0 update internally that is rumored to have all of these fixes intact, but this still means Ryzen 7000 users could be left waiting months for a complete fix.

It is what it is. I'm not going to not use my PC.

 

[Deleted member 2731765]

This means that EXPO has been limited and the only way to enjoy faster memory speeds is to use the older AGESA firmware.

As of now, both the "PROCHOT Control" and "PROCHOT Deassertation Ramp" toggles that were available in the oldest versions of AGESA for AM5, are not available in the latest production AGESA.

 

[A Stoner]

This is my first AMD build, and I have to say, if I end up with a burned out computer after spending effectively as much on this one RIG as I spent on my last 3 intel builds, it is going to be a massive black eye for AMD.

 

[setx]

User on Reddit being stupid with BIOS settings now warrants an article? At least read the comments to that post...

Actual users have positive response to new BIOS with AGESA 1.0.0.7: yes, many old overclocking settings are unstable, but after finding new ones it's the same performance with less voltages and power consumption.

 

[Deleted member 2838871]

This is my first AMD build, and I have to say, if I end up with a burned out computer after spending effectively as much on this one RIG as I spent on my last 3 intel builds, it is going to be a massive black eye for AMD.

It's my first AMD build since the Athlon XP 1800+ in 2001 with a GeForce3 and Windows XP.

My CPU and board already failed and as mentioned above were replaced. I'm not salty about it at all... it didn't cost me a nickel. All I lost was a Saturday afternoon of gaming.

This PC is an absolute beast of a machine... and I'm looking forward to many years of use. I could have went Intel like I have for the last 20 years... but nah... AM5 is new and a solid upgrade path with great thermals. LGA 1700 is EOL.

The choice was easy.

 

[InvalidError]

Looks like Zen 4's launch was almost as badly bungled up as the original Zen's.

If you care about long-term reliability, avoid buying first-gen anything as much as possible. Just like cars, avoid buying first-gen new frame, engine, transmission, etc.

 

[TerryLaze]

User on Reddit being stupid with BIOS settings now warrants an article? At least read the comments to that post...

Actual users have positive response to new BIOS with AGESA 1.0.0.7: yes, many old overclocking settings are unstable, but after finding new ones it's the same performance with less voltages and power consumption.

The article was warranted by AMD thinking that it was warranted to make a new AGESA.....

The reddit post and comments just show how much confusion there is because of naming schemes the companies use.
So this person set values manually but because he used PBO instead of PB, which most people will probably not even realise that it's not the same thing, the settings weren't applied and there was no message telling them so.

 

[bit_user]

I'd have expected these sorts of issues to be sorted, by now, but this is exactly why I don't jump on a new platform, right away.

HardwareLuxx reports that it tested AGESA 1.0.0.7 itself with dual channels 16GB, 32GB, and quad channel 64GB kits and had no issues.

The kit might be sold as "quad-channel", but the client processors only have support for two DIMM-level channels. "quad-DIMM kits" would avoid the potential for confusion.

 

[Deleted member 2838871]

Looks like Zen 4's launch was almost as badly bungled up as the original Zen's.

If you care about long-term reliability, avoid buying first-gen anything as much as possible. Just like cars, avoid buying first-gen new frame, engine, transmission, etc.

I usually do that... but not in this case. Still... no regrets here. Even with growing pains the time to jump ship was now.

The thermals are great too.

 

[Zerk2012]

All I lost was a Saturday afternoon of gaming

Most people have to order parts and would of been down much longer.

 

[Deleted member 2838871]

Most people have to order parts and would of been down much longer.

Oh I know... mine died on a Saturday and I ordered replacements for next day Amazon delivery... was back running Sunday.

Then RMAed dead parts back to Newegg on Monday... and got replacements a week later... which I returned to Amazon to get the $1000 refunded that had been in limbo for a week.

Better than waiting a week for replacement parts but I know a lot of people wouldn't/couldn't do that.

 

[Amdlova]

Happy who got an intel

 

[Kamen Rider Blade]

Quad DIMM kits always lowers the maximum Memory Frequency, by how much, that depends on the memory controller and MoBo support.

 

[TJ Hooker]

The article was warranted by AMD thinking that it was warranted to make a new AGESA.....

The reddit post and comments just show how much confusion there is because of naming schemes the companies use.
So this person set values manually but because he used PBO instead of PB, which most people will probably not even realise that it's not the same thing, the settings weren't applied and there was no message telling them so.

The reddit user wasn't using PB or PBO. They just set a fixed frequency and voltage, mistakenly thinking they were setting maximum limits. Setting static values overrides the normal dynamic frequency/voltage behavior. Same thing happens with Intel CPUs, IIRC.

Manual overclocking of Ryzen 7K also overrides the default temperature target of 95 C, allowing the CPU to get up to the actual TJmax of 115 C before throttling/shutting down, which explains the high temps the reddit user was seeing.

In other words, everything behaved as expected, the 'issue' is purely user error.

 

[waltc3]

The reddit user wasn't using PB or PBO. They just set a fixed frequency and voltage, mistakenly thinking they were setting maximum limits. Setting static values overrides the normal dynamic frequency/voltage behavior. Same thing happens with Intel CPUs, IIRC.

Manual overclocking of Ryzen 7K also overrides the default temperature target of 95 C, allowing the CPU to get up to the actual TJmax of 115 C before throttling/shutting down, which explains the high temps the reddit user was seeing.

In other words, everything behaved as expected, the 'issue' is purely user error.

I have to agree with this completely. Reddit has gotten so rife with people who barely know the difference between a ram chip and a mouse that it's no fun anymore because it's a bunch of ignorant kids who write dumb posts as opposed to using search engines to sort through their configuration problems and figure them out and *solve* them. Invariably they make incorrect, erroneous diagnoses and leave it at that. It's a sad fact, but so many should be on consoles (computers for dummies) instead of open-ended computers that assume some intelligence and eagerness to learn on the part of the operator.

I've read far more accounts of people who love their systems and aren't having any trouble at all! I wonder why. It's a sad time we live in, where ignorance is extolled and stupidity is considered a virtue. It's really so very tiresome.

 

[TerryLaze]

The reddit user wasn't using PB or PBO. They just set a fixed frequency and voltage, mistakenly thinking they were setting maximum limits. Setting static values overrides the normal dynamic frequency/voltage behavior. Same thing happens with Intel CPUs, IIRC.

Manual overclocking of Ryzen 7K also overrides the default temperature target of 95 C, allowing the CPU to get up to the actual TJmax of 115 C before throttling/shutting down, which explains the high temps the reddit user was seeing.

In other words, everything behaved as expected, the 'issue' is purely user error.

From the reply to the reddit OP that probably triggered setx.
Also PB IS the normal dynamic frequency/voltage behavior just like turbo boost is for intel.
Precision Boost = normal turbo (like power limits enabled)
Precision Boost Override = Overclock (like power limits unlimited)

You did not limit the CPU; you set the CPU to manual frequency/voltage mode. This disabled Precision Boost in favor of locking the CPU to your chosen frequency and vcore.

 

[TJ Hooker]

From the reply to the reddit OP that probably triggered setx.
Also PB IS the normal dynamic frequency/voltage behavior just like turbo boost is for intel.
Precision Boost = normal turbo (like power limits enabled)
Precision Boost Override = Overclock (like power limits unlimited)

I'm not sure what point you're trying to make. Yes, PB is the regular dynamic frequency/voltage behavior, and PBO is raising the power/frequency/voltage/current limits to allow PB to boost further and/or longer than it could under stock limits (sort of an auto-OC).

Setting a static frequency and voltage (manual overclock) disables PB/PBO, which is what that redditor did (without realizing).

 

[zx128k]

You can still get a good memory overclock within the 1.3Volts SoC limit. I have seen people get upto DDR5-6300 and have a safe SoC. So 6000-6200 could be possible still for some.

You can't just set memory to DDR5-6400 and let SoC be 1.41 volts.

 

[sitehostplus]

In short, user reports have cropped up revealing dead Ryzen 7000 CPUs that melted in their sockets due to SoC voltages operating at unsafe voltages (which is specified as anything over 1.3v right now), and bugs surrounding AMD's CPU temperature control system, which can cause Ryzen 7000 CPUs to go a beyond its safe thermal specifications.

I have had no further issues since rebuilding with a new CPU and mobo on April 23rd and updating to a new BIOS on April 28th which locked SOC voltage at 1.3v. EXPO has been enabled the entire time.

Still waiting on the AGESA update.

On the bright side, AMD is reportedly testing a new 1.0.9.0 update internally that is rumored to have all of these fixes intact, but this still means Ryzen 7000 users could be left waiting months for a complete fix.

It is what it is. I'm not going to not use my PC.

I've been running my 7950x3d system stock 24/7 for weeks now, and have zero problems. Just don't overclock it and you should be OK.

 

[blacknemesist]

SoC voltage just go through the limit with AGESA 1.0.0.6 if DOCP is enabled and CPU can go past 95ºC on my 7700X. AGESA 1.0.0.5 had the thermal limits in place but never noticed the SoC which was probably as bad as it is now if I enable DOCP.
Only AGESA 1.0.0.7 bios is beta so i rather have DOCP off. Asus mobo and older bios is not available so no 1.0.0.5 to at least keep the thermal protection.

Literally can't win with this damn build.

 

[ikissfutebol12]

So I guess I'm confused with who exactly is at risk here. Is it only overclockers or even at stock settings? I know some articles places originally only suggested the 3D chips in the series, but now it sounds all 7000 CPUs.

 

[blacknemesist]

So I guess I'm confused with who exactly is at risk here. Is it only overclockers or even at stock settings? I know some articles places originally only suggested the 3D chips in the series, but now it sounds all 7000 CPUs.

It's all 7000 CPUs. The current AGESA 1.0.0.7 limits the SoC voltage to a maximum of 1.3v but in my case it is still a beta bios so if something goes wrong the warranty is void. But another issue is thermal protection via throlling at 95ºC not being triggered which is less of an issue if you limit the EDC via PBO and/or give the boost a negative value so it boosts lower than it would normally do thus giving you more control over what temperature the CPU can reach..

 

[TJ Hooker]

it is still a beta bios so if something goes wrong the warranty is void.

Do you have a source indicating that using a beta BIOS voids your warranty? I have not seen that suggested before.

 

[Deleted member 2838871]

I've been running my 7950x3d system stock 24/7 for weeks now, and have zero problems. Just don't overclock it and you should be OK.

All I had done was enable EXPO. Didn’t touch CPU at all. Imagine losing a processor and board using a feature that AMD advertises. 😂

It is again enabled.