Ryzen Owners Zen Garden

[Klemc]

Look, i screened DOCP OFF (+ CPU boost too) and DOCP Tweaked (the most tweaked, by Asus lol) :

EDIT:
it's not a joke the 7800X3D is a "gaming" CPU, it performs low in explorer (10+11 dualboot), that's strange, but gamles work hell fast btw.

Loading apps or opening explorer windows (normal window, save as window...) is like more than 2sec...

... or is it tied to my B650E-E MBoard HDD/SSD/NVMe chip ?.. or did it burn some parts of my motherboard that can't recover full capability (like bluetooth disappearing and needed to purge power for more than one day?

So, since yesterday night it began to give Windows BOOT failures, then BSOD everytime when booted after a short time... what's happening ?..

Also since BIOS 1413 i don't see my 3 SSD appearing in HWMonitor..., and only 2 of the 3 NVMe too !!

//

Clear CMOS button helped (looks loke it) !

... EDIT: NO, another BOOT BSOD , also NVMe 3 and 3x SSD not appearing in HWMonitor, as since BIOS 1413.

BSOD about MEMORY_MANAGEMENT

 

[Mussels]

This all reminds me of how I blew up my AM4 motherboard and nearly destroyed the CPU too.
Lessons learned
1) don't use the motherboard PBO limits
2) the cpu will only try to save itself, not the board
3) it really sucks when motherboard safety features don't work or actually don't exist
4) burning PCB smells really, really, really bad

I've been trying to find your thread for WEEKS now, i couldn't remember your damned user name or thread title to find it!

That was the board that was so amazingly dishonest with the VRM's it just non stop overheated them, but once i lost the link to your thread i couldnt remember the model name

So, since yesterday night it began to give Windows BOOT failures, then BSOD everytime when booted after a short time... what's happening ?..

Also since BIOS 1413 i don't see my 3 SSD appearing in HWMonitor..., and only 2 of the 3 NVMe too !!

//

Clear CMOS button helped (looks loke it) !

... EDIT: NO, another BOOT BSOD , also NVMe 3 and 3x SSD not appearing in HWMonitor, as since BIOS 1413.

BSOD about MEMORY_MANAGEMENT

You on the latest BIOS?
Definitely try lower RAM speeds, as they may have dropped your SoC voltage - if not, you may be experiencing the signs of an impending dead CPU...
Remember that SoC voltage also controls your PCI-E, SATA and some USB ports - so it starting to fail (or being underfed) can result in oddness, especially as voltages can be weirder at low load, than at full load

Look, i screened DOCP OFF (+ CPU boost too) and DOCP Tweaked (the most tweaked, by Asus lol) :

EDIT:
it's not a joke the 7800X3D is a "gaming" CPU, it performs low in explorer (10+11 dualboot), that's strange, but gamles work hell fast btw.

Loading apps or opening explorer windows (normal window, save as window...) is like more than 2sec...

... or is it tied to my B650E-E MBoard HDD/SSD/NVMe chip ?.. or did it burn some parts of my motherboard that can't recover full capability (like bluetooth disappearing and needed to purge power for more than one day?

Looking more closely:

DOCP off, you have 1.025v SoC

Your tweaked settings now shows 1.25v, the max anyone recommends. The danger that's happened has been that boards have sending far more than that voltage through in certain states, like on a hung/failed boot.

It's entirely possible for example that you were totally stable at 6000 before, because the SoC ran up higher (even if it didnt report that in the software/BIOS readings) and now it's capped, 6000 isn't stable.
If you havent updated to one of the new safer BIOSes, i'd be trying a lower RAM speed with tighter timings, to avoid any SoC dropouts. 6000 was the dream for AM5, the sweet spot - not something guaranteed to work for everyone.

Plenty of BIOS settings control NVME/SATA and what is active/disabled at any time, people still get tripped up by drives without an OS not showing in boot menus in UEFI mode with CSM disabled.

 

[Klemc]

So, since yesterday night it began to give Windows BOOT failures, then BSOD everytime when booted after a short time... what's happening ?..

Also since BIOS 1413 i don't see my 3 SSD appearing in HWMonitor..., and only 2 of the 3 NVMe too !!

//

Clear CMOS button helped (looks loke it) !

... EDIT: NO, another BOOT BSOD , also NVMe 3 and 3x SSD not appearing in HWMonitor, as since BIOS 1413.

BSOD about MEMORY_MANAGEMENT

It took one sec to find erros !! Freezed at 6%

 

[Mussels]

You're right, it's errored so hard it's in FRENCH


I'd definitely be trying lower RAM speeds/IF speeds, and seeing if you've been affected - because warranty claims wont be super fast right now.

 

[Klemc]

I've been trying to find your thread for WEEKS now, i couldn't remember your damned user name or thread title to find it!

That was the board that was so amazingly dishonest with the VRM's it just non stop overheated them, but once i lost the link to your thread i couldnt remember the model name


You on the latest BIOS?
Definitely try lower RAM speeds, as they may have dropped your SoC voltage - if not, you may be experiencing the signs of an impending dead CPU...
Remember that SoC voltage also controls your PCI-E, SATA and some USB ports - so it starting to fail (or being underfed) can result in oddness, especially as voltages can be weirder at low load, than at full load


Looking more closely:

DOCP off, you have 1.025v SoC
View attachment 294164

Your tweaked settings now shows 1.25v, the max anyone recommends. The danger that's happened has been that boards have sending far more than that voltage through in certain states, like on a hung/failed boot.
View attachment 294165


It's entirely possible for example that you were totally stable at 6000 before, because the SoC ran up higher (even if it didnt report that in the software/BIOS readings) and now it's capped, 6000 isn't stable.
If you havent updated to one of the new safer BIOSes, i'd be trying a lower RAM speed with tighter timings, to avoid any SoC dropouts. 6000 was the dream for AM5, the sweet spot - not something guaranteed to work for everyone.

Plenty of BIOS settings control NVME/SATA and what is active/disabled at any time, people still get tripped up by drives without an OS not showing in boot menus in UEFI mode with CSM disabled.

I had first 1406 (backflash before received CPU), that is the version listed on MB CPU QVL to be the first for X3D compatibility (earlier, it listed ALL), then 1409 and 1413 applied !!

I disabled CPU Turbo, DOCP is disabled, i can't use lower speeds !!

 

[Mussels]

enable DOCP, XMP, EXPO, or whatever they call it - then lower the RAM speed before saving and rebooting

That sets the timings and DRAM voltage correctly, which can help even at lowered speeds - and should get a safe SoC voltage that prevents any dropouts, on the newer BIOSes

 

[Klemc]

enable DOCP, XMP, EXPO, or whatever they call it - then lower the RAM speed before saving and rebooting

That sets the timings and DRAM voltage correctly, which can help even at lowered speeds - and should get a safe SoC voltage that prevents any dropouts, on the newer BIOSes

Yo're telleing me NEW BIOSes are worst that old one ?

DOCP i'll set on MANUAL, then i set his freq to 4800 ?

//

Well, RAM problems again, i want RAM all this stuff i buyed brand new and they burned to death

 

[Mussels]

Yo're telleing me NEW BIOSes are worst that old one ?

DOCP i'll set on MANUAL, then i set his freq to 4800 ?

Well, RAM no problems now, but i want RAM allthis stuff i buyed brand new and they burned to death

The new BIOS should have the safety features to prevent your CPU exploding from a too high SoC voltage, but that too high voltage could also have made your memory overclock more stable on the old bios.
You want to find what's stable on the latest BIOS, since exploding CPU's are not cool.


Yeah, DOCP on and then lower the RAM to 4800 and see if things behave well - you can obviously raise it higher after testing, but it's a good way to start ruling out problems.
If its stable at 4800, then you can raise it higher until you start seeing issues again. You may find that 6000 has issues but 5800 is fine, or 5600 lets you drop latency values and get the same performance anyway.


DDR5 6000 was *never* guaranteed for AM5, it's the overclocking dream - and some motherboards made it possible by using unsafe voltages, it would seem.

 

[Klemc]

My MBoard is under return window on Amazon (10 may), but CPU i buyed at TopAchat andbill is from 7 april, so i'll see, RAM is from Amazon so they will take care of it i can hope.

 

[Mussels]

My MBoard is under return window on Amazon (10 may), but CPU i buyed at TopAchat andbill is from 7 april, so i'll see, RAM is from Amazon so they will take care of it i can hope.

Yeah all you do is wait until you have parts on hand you can test with.

As an example @ir_cow is our motherboard/RAM reviewer here, and he hasnt been able to get high speed RAM working on every AM5 motherboard either - it's something that's just not always possible, no matter how skilled you are.

You may have damaged hardware, but that's not something anyone but you can determine - sounds like it is, since you couldnt get 4800 working reliably.
Have you inspected the CPU and socket for burn marks or bulging yet?

 

[Klemc]

Yeah all you do is wait until you have parts on hand you can test with.

As an example @ir_cow is our motherboard/RAM reviewer here, and he hasnt been able to get high speed RAM working on every AM5 motherboard either - it's something that's just not always possible, no matter how skilled you are.

You may have damaged hardware, but that's not something anyone but you can determine - sounds like it is, since you couldnt get 4800 working reliably.
Have you inspected the CPU and socket for burn marks or bulging yet?

Recently, not after updating to 1409 & 1413. I will, since i have to dismout all.

 

[Mussels]

Those BIOS wont have cause damage - the most they'll do is cap the SoC voltage for safety which could make a previously stable ram overclock, no longer stable.

 

[Klemc]

Those BIOS wont have cause damage - the most they'll do is cap the SoC voltage for safety which could make a previously stable ram overclock, no longer stable.

Not sure but... AGESA 1.0.07 is waiting btw.

I can't have MBoard replacement, they gave me my money back, for CPU it's harder, they want to test it so i will have to directly deal with AMD, for RAM it's tier Amazon reseller but they've been contacted.

After this three parts refurbished, i can buy another CPU MB RAM couple, for the time in wait, i can use my z390 with 3060

But hell, 7800X3D looked good, what will be my alternatives.

 

[Arco]

Had to clear CMOS at 6 AM in the morning, very fun NOT.

Ram training messed up hard, almost got it fixed so .

The problem with being on the bleeding edge is that it cuts you .

 

[Klemc]

Those BIOS wont have cause damage - the most they'll do is cap the SoC voltage for safety which could make a previously stable ram overclock, no longer stable.

OKay, so, i tested DIMM A2 and stick passed, i moved it to B2 and it not reports errors but PC freezed !!

MBoard DIMM broke ?.. how can it be linked to X3D thingie ?

 

[A Computer Guy]

I've been trying to find your thread for WEEKS now, i couldn't remember your damned user name or thread title to find it!

That was the board that was so amazingly dishonest with the VRM's it just non stop overheated them, but once i lost the link to your thread i couldnt remember the model name

I still question if the CPU was damaged by that mishap but it's been working fine so far except I can't run 128GB ram reliably (reboots when idle for a period of time regardless of ram speed, voltage, etc...). I'm due for my 1yr fluid change this month and was going to swap my 5950x's between my daily and backup pc's to see if that other CPU can run 128GB ram on the Taichi or not.

 

[Klemc]

BURNout !!

 

[Arco]

BURNout !!

NOOOOOOOOOOOOOOOOOO.

 

[tabascosauz]

So, since yesterday night it began to give Windows BOOT failures, then BSOD everytime when booted after a short time... what's happening ?..

Also since BIOS 1413 i don't see my 3 SSD appearing in HWMonitor..., and only 2 of the 3 NVMe too !!

//

Clear CMOS button helped (looks loke it) !

... EDIT: NO, another BOOT BSOD , also NVMe 3 and 3x SSD not appearing in HWMonitor, as since BIOS 1413.

BSOD about MEMORY_MANAGEMENT

If you run for long enough on unstable memory profile you will need to reinstall Windows. Once the Windows install incurs enough damage, it basically is unfixable by any means, even if you fix your mem profile to be stable. Unending boot failure/mem management BSOD is a pretty common symptom.

BURNout !!

What's the context here? Just died all of a sudden?

All the picture shows is a few scattered bent pins in the socket. The catastrophic failures so far that GN is investigating have had visible damage on CPU pads.

The issue so far is AM5-wide and not X3D specific, so if you *really* don't want to deal with any of this I don't think AM5 has anything left to offer...

 

[Klemc]

If you run for long enough on unstable memory profile you will need to reinstall Windows. Once the Windows install incurs enough damage, it basically is unfixable by any means, even if you fix your mem profile to be stable. Unending boot failure/mem management BSOD is a pretty common symptom.



What's the context here? Just died all of a sudden?

All the picture shows is a few scattered bent pins in the socket. The catastrophic failures so far that GN is investigating have had visible damage on CPU pads.

The issue so far is AM5-wide and not X3D specific, so if you *really* don't want to deal with any of this I don't think AM5 has anything left to offer...

And how a pin can be bent ? So the CPU let's play MGS3 MGS4 on RPCS3, then all of a sudden under not heat pins bent under a mounted CPU and fails to RAM ??

Sorry, perhaps. I had CPU Turbo OFF all this time, and DOCP Auto, that's enough done by me to not feel culprist.

EDIT, let's know i never used under load this setup before RPCS3, never more than 10 min and one time only, i had no monitor (it died, had accident during ship...).

 

[3x0]

And how a pin can be bent ? So the CPU let's play MGS3 MGS4 on RPCS3, then all of a sudden under not heat pins bent under mounted CPU and fails to RAM ??

Maybe they were bent from the beginning and slowly killed the CPU. Can you take a picture of the underside of the CPU?

 

[Klemc]

Well

 

[3x0]

Doesn't look like the CPU pads are damaged. Unlikely to be the VSOC issue

 

[Klemc]

Doesn't look like the CPU pads are damaged. Unlikely to be the VSOC issue

What about that ?

 

[tabascosauz]

What about that ?

Doesn't look normal, but also doesn't look completely trashed either. Pins don't usually bend themselves, judging from lack of damage on CPU I'd lean more towards user accident

Not all pads/pins are absolutely necessary to CPU functionality. You can check the pinout to see what the bent pins are, and what the pads correspond to. The problem with only some of the DIMM slots working sounds like a bent pin problem.

If your CPU exploded like the others there would probably be visible pin/pad damage in a larger area, not a single pin. And in the vicinity of a main rail like Vcore that actually delivers substantial power.