i am getting pissed with Zen 3... (still random reboots at idle/low load, cache hierarchy bluescreens and other crap)

[Mussels]

The black screen crashes are IMC/IF related. If you arent stable at 2133Mhz on the RAM with a 1:1 IF, you've just got a faulty CPU and need to RMA it.

You should not need to mess with C states or downclocking or anything at all like that - either you have a 1 in a million incompatibility that hasn't been patched yet, or one common piece of hardware you have is faulty.

 

[Fouquin]

The black screen crashes are IMC/IF related. If you arent stable at 2133Mhz on the RAM with a 1:1 IF, you've just got a faulty CPU and need to RMA it.

You should not need to mess with C states or downclocking or anything at all like that - either you have a 1 in a million incompatibility that hasn't been patched yet, or one common piece of hardware you have is faulty.

But with three different CPUs presumably from different production batches? That's incredibly hard to believe. If AMD had failure rates that high there'd be a high level inquiry and/or recall of those chips.

Again, I'd like to reiterate that on my friend's new 5800X + X570 build from last month we fixed this exact same error with a different brand, and newer SSD. That was literally all it took, just a fresh install on a fresh SSD.

 

[purecain]

Its more than likely the state of the bios agasa update that your on. I havnt updated to the latest bios yet and running a 5950 as it could be a few months before the new agasa code is 100% stable.

I havnt had one crash and am running stable as a rock on an older bios. Good luck and try an older bios agasa version.

 

[dont whant to set it"']

It happens all the time when I take the os drive from one platform to another. Uninstalling the drivers first before "porting" it, may be the key.

 

[Valantar]

Useless post...read before attacking the OP.

"completely rebuilt my PC from scratch, swapped my PSU, GPU, RAM, Mainboard (b550 Tomahawk) and i tried two other Ryzen CPUs."

Lol, "attacking"? Where? How? Please try adding just a tad of nuance to your language, it'll help you actually have constructive dialogue with people. Criticizing, maybe? Nonetheless, a sample size of two and an undiagnosed problem are still not grounds for making sweeping proclamations about an entire architecture. I absolutely understand the OP's frustration, but their way of framing this amounts to shaking their fist at the sky and shouting "Damn you, AMD!", which doesn't help them solve their problem whatsoever.

 

[Mussels]

But with three different CPUs presumably from different production batches? That's incredibly hard to believe. If AMD had failure rates that high there'd be a high level inquiry and/or recall of those chips.

Again, I'd like to reiterate that on my friend's new 5800X + X570 build from last month we fixed this exact same error with a different brand, and newer SSD. That was literally all it took, just a fresh install on a fresh SSD.

interesting and possible

 

[thesmokingman]

But with three different CPUs presumably from different production batches? That's incredibly hard to believe. If AMD had failure rates that high there'd be a high level inquiry and/or recall of those chips.

Again, I'd like to reiterate that on my friend's new 5800X + X570 build from last month we fixed this exact same error with a different brand, and newer SSD. That was literally all it took, just a fresh install on a fresh SSD.

If one had corsair ram, it could give issues on any cpu. And your anecdotal issue, that's on the individual ssd. Did you confirm that the ssd was defective or incompatible? There's an important distinction there. And one case doesn't make it wide sweeping. The ram is more likely an issue especially with new users to ryzens.

 

[Fouquin]

And your anecdotal issue, that's on the individual ssd. Did you confirm that the ssd was defective or incompatible? There's an important distinction there. And one case doesn't make it wide sweeping.

I'm not sure what the purpose of this post is. Read the OP:

but if anybody has another "trick" that worked for him.. please let me know it.

Trick provided. Anecdotal or not, it's a solution that worked for this error code.

 

[GLD]

My rig was working fine, then I upgraded to the newest v.3001 bios and got problems. It wouldn't boot/run with D.O.C.P. enabled, as before. It wouldn't let me downgrade the bios either. It would only run with bios set to default values. I don't OC and would normally just set fan profile and enable D.O.C.P. and it would be stable at 1:1:1. What ended up working for me was also changing two other things in the bios. I changed Vcore and Soc from auto to manual and set them to the default values it showed. It showed 1.1v for the Vcore and I thought that was low, down from the auto 1.4v+, but went with it. I think it was 1.025v for the Soc. It has been a couple weeks now with gaming and benchmarks, with being stable at 1:1:1 AND the cpu is an average 20*C+ cooler!

I think Ryzen problems are caused by the bios setting to much voltage by default. Mine was anyways. Try changing some values from auto/default to manual to keep them from getting over juiced.

 

[efikkan]

i completely rebuilt my PC from scratch, swapped my PSU, GPU, RAM, Mainboard (b550 Tomahawk) and i tried two other Ryzen CPUs.
the Problem is 100% an issue with Zen 3/Bios/AGESA.

Just for clarity, you either tested two other Zen 1/2 CPUs that didn't experience the same problem, or they were Zen 3 CPUs and did experience the same problem?
It matters, because one would eliminate the CPU as the culprit.


Beyond trying a completely different SSD or HDD as a long shot, I think you have pretty much narrowed it down to being either BIOS, motherboard or CPU. (But you tested two different motherboards, didn't you?)

I would probably test an unstable computer in Linux too, just to see if it happens there and if it sheds some extra light on what's going on, but this is probably only useful for those who are familiar with this.

What you should do in terms of returning it, contacting the motherboard maker support etc., or just wait a few weeks for a new BIOS depends on the rights you have as a buyer in your country, return policy of the store etc.
To me this smells like a BIOS issue, which I would expect to be eventually fixed. And there are a number of such problems reported by others, which is why I have postponed setting up a Zen 3 system for another few months.

 

[aQi]

Before i add something does the OP has any update on the situation ?

 

[Mussels]

You're using one (or more) incompatible parts at the same time.
Only you can narrow that down

PSU, RAM (stock/XMP), Mobo, GPU, CPU, SSD, OS etc. Find out the part(s) all these faulty systems have had in common, and do a build without it.
Hell we dont even know what BIOS settings you're using, you keep tweaking them (fine) but we cant give advice on anything you do and dont explicitly tell us about.

Yes, people have these issues like you... but theres a lot more WITHOUT the issues, or those that get it and resolve it.

 

[Caring1]

I was going to suggest disabling C states but noticed you already did that in the O.P.
Have you looked for and found ULPS and tried disabling that?

 

[Deleted member 193596]

You're using one (or more) incompatible parts at the same time.
Only you can narrow that down

PSU, RAM (stock/XMP), Mobo, GPU, CPU, SSD, OS etc. Find out the part(s) all these faulty systems have had in common, and do a build without it.
Hell we dont even know what BIOS settings you're using, you keep tweaking them (fine) but we cant give advice on anything you do and dont explicitly tell us about.

Yes, people have these issues like you... but theres a lot more WITHOUT the issues, or those that get it and resolve it.

Here is an update.

i've spent the last 1 1/2 days mostly in the Bios and on Windows installation screens. i tried every single workaround available on the internet and i got it somewhat stable. (see bottom)



Hardware tested:

Corsair RM850X, bequiet Dark Power Pro 11 750W.

Asus B550 Strix F , B550 Tomahawk, X570 Aorus Pro

2x 8GB Corsair vengence LPX 2400 Mhz ,1x8 GB Corsair vengeance LPX 3200 Mhz , 2x 16GB Trident Z Neo 3600 Mhz

RX 5700XT (Sapphire Pulse, Nitro + SE)
RX 580 8GB (Armor OC, Pulse Lite)
Vega 64 Strix OC
GTX 1060 6GB (two models from Zotac and EVGA)
GTX 1070 Zotac
GTX 1080 Ti Gaming OC, Strix OC
GTX 1660 Ti
RTX 2060 KO
RTX 2070 Super
RTX 2080 Super
RTX 2080 Ti
RTX 3070 (Gigabyte Gaming OC, MSI Ventus 2x)
RTX 3080 (TUF OC, FTW3 Ultra)
RTX 3090 Gaming X Trio

Ryzen 7 5800X, Ryzen 5 5600x, Ryzen 7 3800X, Ryzen 5 3600X

Fresh Windows 10 Installations on the following SSDs:
Samsung 970 Evo Plus 250GB
Samsung Evo Plus 500GB
Three! different Crucial MX500 1TB
two Samsung 870 QVO 4TB
One other Crucial MX 500 500GB.


i tried every single Bios completely Stock, With and without XMP, C States, Virtualisation (IOMMU), Fast Boot and PBO/Curve Optimizer.
I tested every single board with all ram kits for several hours with Memtest,Prime 95, OCCT and tons of Gaming.

Everything is completely Stable und every circumstance (0 Errors in Memtest, 0 Errors in P95 (version 30.3) Zero errors in OCCT and Zero Crashes in Games.


As soon as i let it sit on the desktop or watch a youtube video the PC just reboots or freezes.

This Happens with the 5800X the borrowed 5600X on all Mainboards.

BOTH ZEN 2 CHIPS ARE WORKING FINE ON ALL BOARDS WITH ALL RAM KITS.

i am almost 100% Sure that it is a Power issue with the extremely low VID that the CPUs have at idle (as low as 0.2V)



///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

What did i to get it stable?

1.Manual OC with 1.33125V at 4.65Ghz (no more .2V at idle and ZERO crashes/reboots)


And a feature called Powersupply Idle Control ( Advanced/AMD CBS/ CPU Common Options)

it is at "Auto".
and Auto is "low current idle"

i changed that to typical current idle.

what happened?

The VID and VCore reported from the Nuvoton NCT6798D does NOT go below 0.9V (SVI2 TFN Voltage at idle)


i am Sorry if some sentences sound weird, off or "mad". i am German, i mostly translate my slightly pissed mood German into english and it is 2:30 AM right now.


I will post updates again if it crashes or freezes with the current settings which are the following:

Stock, Latest Beta Bios Version 1216 (Patch C ) on a B550 Strix F.
DOCP enabled (2x16GB Trident Z Neo 3600 CL16
PSU Typical Current Load instead of Low current load.

 

[Space Lynx]

I'm glad you found a fix. I'm just opposite as you, as soon as I OC I get a crash no matter what. Leave it at stock no crashes, lol. AMD is fun like that. I think it will improve with future BIOS though so I am not worried/no rush.

Can't say I have had any issues myself, I had some in the beginning, but it was because I was overclocking. I'm not overclocking now and I just enable XMP for the ram. 0 issues and 0 crashes, and I been gaming 7+ hrs a day, watching youtube daily, etc.

Waiting for new BIOS before I do any oc'ing.

 

[Totally]

After googling the issue is a software/hard drive issue. Just clean install and be done with.

 

[Caring1]

Here is an update.

i've spent the last 1 1/2 days mostly in the Bios and on Windows installation screens. i tried every single workaround available on the internet and i got it somewhat stable. (see bottom)



Hardware tested:

Corsair RM850X, bequiet Dark Power Pro 11 750W.

Asus B550 Strix F , B550 Tomahawk, X570 Aorus Pro

2x 8GB Corsair vengence LPX 2400 Mhz ,1x8 GB Corsair vengeance LPX 3200 Mhz , 2x 16GB Trident Z Neo 3600 Mhz

RX 5700XT (Sapphire Pulse, Nitro + SE)
RX 580 8GB (Armor OC, Pulse Lite)
Vega 64 Strix OC
GTX 1060 6GB (two models from Zotac and EVGA)
GTX 1070 Zotac
GTX 1080 Ti Gaming OC, Strix OC
GTX 1660 Ti
RTX 2060 KO
RTX 2070 Super
RTX 2080 Super
RTX 2080 Ti
RTX 3070 (Gigabyte Gaming OC, MSI Ventus 2x)
RTX 3080 (TUF OC, FTW3 Ultra)
RTX 3090 Gaming X Trio

Ryzen 7 5800X, Ryzen 5 5600x, Ryzen 7 3800X, Ryzen 5 3600X

Fresh Windows 10 Installations on the following SSDs:
Samsung 970 Evo Plus 250GB
Samsung Evo Plus 500GB
Three! different Crucial MX500 1TB
two Samsung 870 QVO 4TB
One other Crucial MX 500 500GB.


i tried every single Bios completely Stock, With and without XMP, C States, Virtualisation (IOMMU), Fast Boot and PBO/Curve Optimizer.
I tested every single board with all ram kits for several hours with Memtest,Prime 95, OCCT and tons of Gaming.

Everything is completely Stable und every circumstance (0 Errors in Memtest, 0 Errors in P95 (version 30.3) Zero errors in OCCT and Zero Crashes in Games.


As soon as i let it sit on the desktop or watch a youtube video the PC just reboots or freezes.

This Happens with the 5800X the borrowed 5600X on all Mainboards.

BOTH ZEN 2 CHIPS ARE WORKING FINE ON ALL BOARDS WITH ALL RAM KITS.

i am almost 100% Sure that it is a Power issue with the extremely low VID that the CPUs have at idle (as low as 0.2V)



///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

What did i to get it stable?

1.Manual OC with 1.33125V at 4.65Ghz (no more .2V at idle and ZERO crashes/reboots)


And a feature called Powersupply Idle Control ( Advanced/AMD CBS/ CPU Common Options)

it is at "Auto".
and Auto is "low current idle"

i changed that to typical current idle.

what happened?

The VID and VCore reported from the Nuvoton NCT6798D does NOT go below 0.9V (SVI2 TFN Voltage at idle)


i am Sorry if some sentences sound weird, off or "mad". i am German, i mostly translate my slightly pissed mood German into english and it is 2:30 AM right now.


I will post updates again if it crashes or freezes with the current settings which are the following:

Stock, Latest Beta Bios Version 1216 (Patch C ) on a B550 Strix F.
DOCP enabled (2x16GB Trident Z Neo 3600 CL16
PSU Typical Current Load instead of Low current load.

Looks like we came to the same conclusion, that Ultra Low Power States were the culprit.

 

[natr0n]

Try using...Typical current idle

It seems to have solved some random reboots I got. Using an older model psu.

 

[thesmokingman]

1.Manual OC with 1.33125V at 4.65Ghz (no more .2V at idle and ZERO crashes/reboots)

.2v is the sleep voltage. All you're doing is preventing your sleeping cores from using minimal voltage. The cpu is designed to put cores to sleep at that voltage, shrugs.

On a personal experience level, I have had random reboot crashes before and that was generally fixed by raising dram voltage. My dram sticks are 16gbx2@4000mhz which I run at different speeds so I tend to have to give it a lil more voltage than what dram calc suggests.

*For ex. dram calc susgests I run my dimms at 1.37v when I run 3600mhz at cas 14 latency. However real world anything under 1.39v gives me random weirdness. Yea, ya can't take dram calc's suggestions at face value but as a starting point.

 

[Deleted member 193596]

.2v is the sleep voltage. All you're doing is preventing your sleeping cores from using minimal voltage. The cpu is designed to put cores to sleep at that voltage, shrugs.

The Cores may be designed to put to sleep.

But the burger flippers in the AMD software development team aren't able to even make a CPU functional at stock.


Btw. Ryzen Master still says that cores are at their sleep state.
Plus they drop down to 0.000w at idle ( at least two or three at the same time)

 

[Space Lynx]

The Cores may be designed to put to sleep.

But the burger flippers in the AMD software development team aren't able to even make a CPU functional at stock.


Btw. Ryzen Master still says that cores are at their sleep state.
Plus they drop down to 0.000w at idle ( at least two or three at the same time)

works fine for me at stock

 

[kruk]

And a feature called Powersupply Idle Control ( Advanced/AMD CBS/ CPU Common Options)

it is at "Auto".
and Auto is "low current idle"

i changed that to typical current idle.

This is probably the fix that worked. The random restart bug has been there since the first gen, but I was under the impression that it mostly plagued Linux users (and was fixed in future Zen iterations). Until Powersupply Idle Control setting appeared, affected users could also disable the C-states (but it affected boosting) - it's weird that it didn't work in your case.

 

[Deleted member 24505]

Switch to intel, might not be as fast or have the kudos Amd has now but it still is and always has been stable as a mountain. I have had zero crashes or problems with mine since i built it in march last year.

 

[Deleted member 193596]

Switch to intel, might not be as fast or have the kudos Amd has now but it still is and always has been stable as a mountain. I have had zero crashes or problems with mine since i built it in march last year.

I am coming from intel...
Stable or not.. i am getting sick of running a CPU that pulls more power when loading a game than a 5950x in prime95.

AMD deserves the money and support. Intel not.

 

[Deleted member 24505]

Tbh I don't even really consider power use. If it bothered me that much I'd have a i3 or Ryzen 3. Imo the most important thing in a pc above everything else is stability. You could have the most powerful hardware available but if it's not stable, what use is it. Also some people have 850watt+ PSU's then whine about power use.
Well I hope you manage to get it actually usable.
So you gave intel your money and support for a long time, now you switched to amd and say intel deserve neither, bit hypocritical imo