Ryzen 7000X3d is burning out.

[jos]

16 hours ago, Mark Kaine said:

Didn't have any issues* with 2200G, 3600 or 5800x3D...

 

 

these new chips though,  too hot for my tastes, not sure about the intel situation but im not gonna upgrade for a few years anyway

 

*edit: except some ram shenanigans which wasn't too bad though (i blame it more on infamous corsair "Vengeance ")

Bad memory controller is what had kept me away from this generation. I was planning upgrade from 3000 series, and ram is priority for me.

 

 

 

[leadeater]

8 hours ago, Fasterthannothing said:

It's getting absurd at this point. Atleast for the most part Intel just works.

Absurd how? I can't think of any serious issues that really have anything to do with AMD specifically and a lot of the "issues" are also "these apply to Intel too things" like 2DPC limitations etc.

[Mihle]

8 hours ago, Fasterthannothing said:

It's getting absurd at this point. Atleast for the most part Intel just works.

If it's only Asus mobos, that it seem might be the case, it's Asus fault and no AMDs. But if it isn't, then it's something else.

[GoodBytes]

Correction, MSI boards are affected as well, or they are taking preventative measures:

https://www.anandtech.com/show/18825/msi-addresses-cpu-voltages-on-am5-motherboards-for-ryzen-7000x3d-processors

MSI released a fixed BIOS.

 

MSI locks down the BIOS further, preventing offset voltage to be specified.

 

I am not sure if that is really the issue, and MSI is playing safe, or what, as all reports so far, at least, publically, has been ASUS boards and one report from ASRock board.

[Quackers101]

Would be understandable, when these 3D chips was locked or more restrictive for many reasons, hopefully it doesnt hit non 3D chips or cause issues between them like performance over safety or the other way around.

[Fasterthannothing]

7 hours ago, Mihle said:

If it's only Asus mobos, that it seem might be the case, it's Asus fault and no AMDs. But if it isn't, then it's something else.

Except it is AMD considering MSI and other boards are causing the same issues 

[pas008]

19 minutes ago, Fasterthannothing said:

Except it is AMD considering MSI and other boards are causing the same issues 

link them please

[GoodBytes]

8 minutes ago, Fasterthannothing said:

Except it is AMD considering MSI and other boards are causing the same issues 

MSI seems like it is preclusion, I did not see any reports... at least not publicly. Considering that they say that they are disabling Offset Voltage (Positive), then that means that people where boosting voltage to probably force an OC.

ASRock has tendency of not following specs limitations and likes to make everything avail. We still don't know if the user has done some big OC and boosted things.

ASUS might have already fixed the problem, I am not seeing any statements, beside them removing all old BIOSs, and keeping the last one and the latest Beta one.

 

Another possibility is that there is no actual problem. Users just never updated the BIOS to the X3D supported version and using the out of the box BIOS.

I mean, how many YouTubers and guides says that BIOSs should never be updated unless you have a problem. While when the motherboard is new, it is EXACTLY the right time to update it, as if it fails, you can simply return/exchange it with ease at the retail store. Also, motherboard manufacturer should have a robust BIOS update system, not your problem if it craps out. Your RMA to the manufacture is costly to the manufacture (the call + warranty replacement), so they'll be pushed in making a good implementation. It is very rare to see a BIOS update fail with Dell's, HP and Lenovo's. If these OEMs can make it, so can ASUS, ASRock, MSI, and so on.

 

[jos]

3 hours ago, GoodBytes said:

Correction, MSI boards are affected as well, or they are taking preventative measures:

https://www.anandtech.com/show/18825/msi-addresses-cpu-voltages-on-am5-motherboards-for-ryzen-7000x3d-processors

MSI released a fixed BIOS.

 

MSI locks down the BIOS further, preventing offset voltage to be specified.

 

I am not sure if that is really the issue, and MSI is playing safe, or what, as all reports so far, at least, publically, has been ASUS boards and one report from ASRock board.

https://imgur.io/a/1oNS9DC

Asrock it is shown as 7700X not one with 3D vcache 

 

[GoodBytes]

5 minutes ago, jos said:

https://imgur.io/a/1oNS9DC

Asrocknit is shown as 7700X not one with 3D vcache 

 

Well, that is a case of 1, so far. And again, we have no idea what the user did.

Maybe he tried extreme OC and not saying it, or not understanding the impact of high voltages.

From my reading from the original poster of the issue on reddit, the system was scorching hot, even his AIO was hot, when the system was idle. In his case, he noted that he does have X3D chip but was using an older BIOS which didn't supported X3D, as he had better memory support, he could use EXPO setting, while the newer one, he could not without stability issues. (This was an issue I faced. In my case 1410 version of ASUS board did fix everything, and so many online, but some people remain affected, but anyways, different topic. Just explaining the why, of that user)

 

So, I guess, one should check their temps, but he wasn't home, and sleep was disabled.

[jos]

8 minutes ago, GoodBytes said:

Well, that is a case of 1, so far. And again, we have no idea what the user did.

Maybe he tried extreme OC and not saying it, or not understanding the impact of high voltages.

From my reading from the original poster of the issue on reddit, the system was scorching hot, even his AIO was hot, when the system was idle. In his case, he noted that he does have X3D chip but was using an older BIOS which didn't supported X3D, as he had better memory support, he could use EXPO setting, while the newer one, he could not without stability issues. (This was an issue I faced. In my case 1410 version of ASUS board did fix everything, and so many online, but some people remain affected, but anyways, different topic. Just explaining the why, of that user)

 

So, I guess, one should check their temps, but he wasn't home, and sleep was disabled.

I have heard in many videos, if processor runs hot, it will thermal throttle and eventually shuts down. To prevent damage. So it can be design flow in AMD part right? Also another possibility is the thermal run away is faster than sensors picked it up. It should not have happened unless short circuit, even with some buggy firmware. Asus should never allow such quick deltas in voltage changes.

 

[HenrySalayne]

27 minutes ago, jos said:

I have heard in many videos, if processor runs hot, it will thermal throttle and eventually shuts down. To prevent damage. So it can be design flow in AMD part right? Also another possibility is the thermal run away is faster than sensors picked it up. It should not have happened unless short circuit, even with some buggy firmware. Asus should never allow such quick deltas in voltage changes.

 

You are absolutely correct: it could be anything...

 

43 minutes ago, jos said:

https://imgur.io/a/1oNS9DC

Asrock it is shown as 7700X not one with 3D vcache 

 

Why is the load mechanism missing?

[Napper198]

29 minutes ago, HenrySalayne said:

Why is the load mechanism missing?

probably because the person ran into the same problem as me and realized that their cooler isn't actually compatible with AM5 and decided to go the janky route of keeping the CPU in with the cooler mounting pressure. NGL I thought about doing the same for a second but instead I decided to run with a wraith spire at 100% and just did nothing more than windows installation until I waited for parts.

I'll let you decide which way was worse.

[HenrySalayne]

1 hour ago, Napper198 said:

probably because the person ran into the same problem as me and realized that their cooler isn't actually compatible with AM5 and decided to go the janky route of keeping the CPU in with the cooler mounting pressure. NGL I thought about doing the same for a second but instead I decided to run with a wraith spire at 100% and just did nothing more than windows installation until I waited for parts.

I'll let you decide which way was worse.

That would be a plausible explanation, but:

 

I see no reason to unscrew the loading mechanism. The cooler obviously mounts to the stock backplate. ¯\_(ツ)_/¯

 

[porina]

tldw

Happening on various mobo manufacturers, not just Asus

Not limited to X3D models

Common factor seems to be EXPO ram - suspects are SOC and ram voltage. Core seems unrelated.

 

[leadeater]

6 hours ago, jos said:

I have heard in many videos, if processor runs hot, it will thermal throttle and eventually shuts down. To prevent damage. So it can be design flow in AMD part right? Also another possibility is the thermal run away is faster than sensors picked it up. It should not have happened unless short circuit, even with some buggy firmware. Asus should never allow such quick deltas in voltage changes.

 

That is only for internal die temperatures, this is a substrate or contact point issue so those protections won't do anything since they wouldn't actually be going above thermal limit points. Either it's a manufacturing fault bonding the die to the substrate or a defect in the substrate itself or a contact/mounting issue.

 

Either way it's not because the "CPU temperatures" are getting too high.

[Kisai]

9 hours ago, GoodBytes said:

 

I mean, how many YouTubers and guides says that BIOSs should never be updated unless you have a problem. While when the motherboard is new, it is EXACTLY the right time to update it, as if it fails, you can simply return/exchange it with ease at the retail store. Also, motherboard manufacturer should have a robust BIOS update system, not your problem if it craps out. Your RMA to the manufacture is costly to the manufacture (the call + warranty replacement), so they'll be pushed in making a good implementation. It is very rare to see a BIOS update fail with Dell's, HP and Lenovo's. If these OEMs can make it, so can ASUS, ASRock, MSI, and so on.

 

Because it's always been bad advice to flash the BIOS if you don't know what the hell you're doing. Dell has always made it easy to flash the BIOS, and even recover from it. Something that most desktop systems do not do. Dell even pushes BIOS updates via windows update. Most desktop motherboards DO NOT do this because because a system built by a third party is not a known quantity, unlike a laptop.

 

What if flashing the BIOS on a desktop results in the CPU or RAM (over)clocks bricking the device? What if the RAM worked fine under the original training, but won't work at all? There's a lot of ?'s all over the place for pushing a firmware update to a desktop, that you really do want to do a BIOS update via Windows Update only to prevent some critical issue (the last two cases I can think of this happening was MELTDOWN/SPECTRE-related.) The firmware updates were pushed out and various systems lost up to 15% of their performance for the BIOS mitigation. But the issue was critical and Dell has a lot of installed systems.

 

Your AsRock/Asus/MSI/Gigabyte system, you have to realize is not typically something you find in an enterprise environment. These companies tend to trade on bells and whistles, and aren't concerned with compatibility with each other, let alone standardizing on firmware.

 

It's always been "possible" to flash the firmware of one board to another manufacturer's board with the same chipset, but that's the problem in itself. When a manufacturer has 15 boards, and it's competitor has 15 boards, there is a risk of actually downloading the wrong firmware. The only thing that prevents this from happening is the firmware updater actually checking the name of the firmware already in use. People also mod firmware to put boot-logo's or add boot rom data to them.

  

Laptops, mobile devices and IoT stuff, you are never supposed to flash the BIOS unless the manufacturer has indicated something fixed by that BIOS applies to your situation. The consequences will often be dire if you screw it up, but continuing to use the stock firmware might also have consequences that aren't immediate. When you call Dell with a problem, one of the first things they make you do before sending it to them becomes an option, is "update the firmware"

 

Desktop's because of the variety of hardware you might encounter, often has rules like "unplug absolutely everything", which means for some people, this can mean unplugging thousands of dollars of (picky) hardware to do so. If it's not broken, you are not going to do this.

 

[leadeater]

43 minutes ago, Kisai said:

Because it's always been bad advice to flash the BIOS if you don't know what the hell you're doing

That's like really old advice and it wasn't even correct or good back in the day either. This came about from older motherboards or software utilities that didn't do any proper basic checks and would just fire down a firmware file regardless if it was compatible or not. Today and for a decade or more every possible supported way to do a BIOS firmware update actually checks the file integrity and compatibility.

 

The current correct advise is ALWAYS update the BIOS firmware to the latest. A lot of fixes do some through especially early platform life so if you want to be able to use the latest and greatest fast RAM it'll likely only work on the latest firmware. That's just how it is now days. Installing it within the first week might not be the best advice but running the latest possible in general is, just take a measured approach as to when to install it.

 

Higher end desktop motherboards, so all OEMs excluded, have dual BIOS ROMs so if you corrupt the firmware during an update you can toggle the physicals switch and change to ROM2 and fix ROM1 etc.

 

Also Spectre and Meltdown didn't really have anything to do with BIOS firmware and it's not really a BIOS mitigation. These were CPU microcode updates that you could do within the OS if you wanted, or drop the microcode files in the relevant Linux filesystem location for your distro and it would automatically update at next boot. The BIOS updates that were released simply contained the CPU microcode updates and replaced the existing ones and any CPU installed in the motherboard would get checked and updated if it was below the version contained in the motherboard firmware. There were a lot of different ways to get the microcode updates to the CPUs, either way that's where the fix was. BIOS update isn't necessary to get it done.

[CommanderAlex]

Can we cook on top of AMD processors now??? This video is getting a little old. 

 

[Bitter]

Anyone else remember this era of AMD?

 

[Arika]

damn, people are are lenient when it comes to AMD...

[GoodBytes]

1 minute ago, Arika S said:

damn, people are are lenient when it comes to AMD...

People are waiting for info

[Arika]

1 minute ago, GoodBytes said:

People are waiting for info

exactly, something only AMD seems to be afforded in much more often than a certain blue themed company.

[Kisai]

4 hours ago, leadeater said:

That's like really old advice and it wasn't even correct or good back in the day either. This came about from older motherboards or software utilities that didn't do any proper basic checks and would just fire down a firmware file regardless if it was compatible or not. Today and for a decade or more every possible supported way to do a BIOS firmware update actually checks the file integrity and compatibility.

I always updated the BIOS for hardware I ran. Or new hardware for clients, but only if there was a reasonably assured no-fail way to do it. I've remotely updated firmware for systems thousands of miles away, and THAT always makes me nervous.

 

I've also had hardware FAIL after updating the firmware, such as with that Xeon/DDR2 system I had, which the gigabyte P45 board just kill itself, despite being "quad bios", every few boots it would screw up the memory training and then when that failed it would try to restore the BIOS. Eventually one day I didn't catch it when it did it, and it destroyed the BIOS. Fortunately I still had the P35 board from the previous crisis.

 

4 hours ago, leadeater said:

The current correct advise is ALWAYS update the BIOS firmware to the latest. A lot of fixes do some through especially early platform life so if you want to be able to use the latest and greatest fast RAM it'll likely only work on the latest firmware. That's just how it is now days. Installing it within the first week might not be the best advice but running the latest possible in general is, just take a measured approach as to when to install it.

 

I'd say the correct advise is to UPDATE the BIOS if you intend to upgrade the CPU or replace the RAM at some point. This is typically why I would not buy a MB after a change in PCIe or DDR family, because if you've ever seen the motherboard revision history for companies like ASUS, you'd know that some times there's like 4 versions of the same board and they only differ in the firmware it came with and several missing headers.

 

4 hours ago, leadeater said:

Higher end desktop motherboards, so all OEMs excluded, have dual BIOS ROMs so if you corrupt the firmware during an update you can toggle the physicals switch and change to ROM2 and fix ROM1 etc.

That doesn't save it. That only recovers from a situation where the BIOS was stable before. If the BIOS, CPU or RAM were not stable, such as when you put a CPU or RAM module that is newer than the MB validation date in it. Like I mentioned in the previous paragraph. If you are intending to upgrade it at ANY point, you would want to update the BIOS before you do. So if you've updated the BIOS every time there was a new BIOS, you should be OK to install a newer CPU or RAM for that board. However if you bought it from the retailer, who maybe had it in the back of their warehouse for 11 months, and you put a new CPU on it, good luck updating the BIOS, you'll probably have to RMA it for a newer version.

 

4 hours ago, leadeater said:

Also Spectre and Meltdown didn't really have anything to do with BIOS firmware and it's not really a BIOS mitigation. 

It was pushed thru windows Update. I've seen it on more than a dozen occasions where a firmware update pops up in Windows Update, and they were all around the same time. 

https://www.dell.com/community/Laptops-General-Read-Only/Bios-Update-pushed-through-Windows-Update/td-p/5116085

 

You'll note the date aligns with those mitigations.

 

 

Just look at how many warnings there are

https://www.dell.com/support/kbdoc/en-ca/000124211/dell-bios-updates

That is as intimidating as hell for someone who is asked "did you update the BIOS?"

 

 

At any rate. One would presume that a whitebox system would be updated to the BIOS needed to run the CPU, correctly, but not everyone who builds PC's, actually knows what they are doing.

 

Hell the "PC simulator" game on steam/epic completely loses the big picture when it focuses on RGB and colored cables and doesn't ever get into firmware or driver issues.

 

[HenrySalayne]

4 hours ago, Arika S said:

exactly, something only AMD seems to be afforded in much more often than a certain blue themed company.

Yes, you are absolutely right...

6800X dies cracking - only a dozen people affected, issue has nothing to do with AMD in particular, wordwide outrage

7900XT(X) issues with the stock cooler - number of affected people unclear, issue resolved within a month, worldwide outrage

AM5 boards are expensive - worldwide outrage

Z790 boards are expensive - ... nothing.

 

6 hours ago, CommanderAlex said:

Can we cook on top of AMD processors now??? This video is getting a little old. 

Most people don't realize it, but even the IHS has roughly double the power density of a hot plate. The dice itself have 5 to 10 times the power density of a hot plate. So cooking on a processor shouldn't be a problem.