PC Drivers & Software

[GoLLuM13]

Hello,
I would like to report an issue that started with AGESA 1.2.0.5 and continues with AGESA 1.2.0.6 and 1.2.0.7 for Ryzen 3000 CPU owners.
Currently, I have BIOS 3801 based on AGESA 1.2.0.3 and SMU version for Matisse (Ryzen 3000) processors is 46.71.0.
On this BIOS version everything works as it should (apart from the fTPM problem with Windows 11 which causes a stutter supposed to be fixed with AGESA 1.2.0.7).
Since BIOS version 3904 (i.e. 3904 - 4002 - 4006 and 4201) the AGESA has gone from 1.2.0.3 to 1.2.0.5 then 1.2.0.6 and finally 1.2.0.7 and the SMU for the Matisse CPU (Ryzen 3000) is increased from 46.71. 0 to 46.72.0 and from this version (AGESA and/or SMU) it is almost impossible to start or restart the computer without Q-Code error, black screen, BSOD, as soon as the DOCP profile is activated or the memory frequency is set manually. The only way to make the BIOS based on these AGESA versions work is to leave the RAM frequency on "default" which is not ideal because in my case for example it means 2133 Mhz.
Of course, reverting to an earlier BIOS like the 3801 (which I'm currently using) fixes the issues instantly.
I did some memtests and other tests, the problem is not with my hardware. I researched various forums (including Asus and MSI) and found that this is a common problem among owners of Ryzen 3000 processors and an X570 motherboard or B550. I even saw a person having the same problem, who solved the problem by changing their Ryzen 3000 pro processor to a Ryzen 5000.
So, I really hope this is not a trick from AMD to let Ryzen 3000 users down, and you will do something about it.
PS: I have already contacted Asus support, they have already offered me several solutions, none have worked, and none will work as long as it is a problem within the AGESA, and it comes from you.

[jawntem]

Make sure that your chipset drivers are up to date, but this seems to be an issue with the 3000 CPU going bonkers. I'd upgrade your CPU to a different model. Also, are you running an XMP profile?

[GoLLuM13]

Thanks for the answer,
As you may know, having a Q-Code or a black screen before OS loading has nothing to do with any driver (chipset included), even though my chipset driver is already up-to-date.

XMP Profile is Intel's terminology, on an AMD platform it's called DOCP and as mentioned, yes ! I tried DOCP profile and I also tried setting up memory manually it's in both cases the source of the problem with BIOSes based on AGESA 1.2.0.5 and newer, only way is to "fix" this is to go back to an earlier BIOS or setting everything RAM related to "Default" which isn't an option for me because whether I'm losing performance or I'm having stutters because of the fTPM implementation in Windows 11 which 1.2.0.7 was supposed to fix.

Upgrade my CPU to a different model ? Why would I buy a new CPU when mine is more than sufficient and perfectly working ? The problem is in the AGESA provided by AMD to the manufacturers, not in my hardware.

Have a nice day

[jawntem]

I first came across this issue with the 3000 chip while watching a YT vid by Jayz. I'll see if I can run down this vid and post a link here pertaining to it.

[ThreeDee]

just a quick note .. 

Only ASUS calls it D.O.C.P. .. every other AMD motherboard manufacturer refers to it as XMP

[michelle_willia]

Yes brother I am used XMP and word press easily and my Pc Different  Different  websites Information to carry Like Gold Information etc .

[michelle_willia]

Make sure that your chipset drivers are up to date, but this seems to be an issue with the 3000 CPU going bonkers. I'd upgrade your CPU to a different model. Also, are you running an XMP profile?

[hawkeye]

I don't think they will fix it but it never hurts to try to open a support ticket. That way you have a higher chance to actually having an AMD employee look into this.
As a workaround I would suggest to manually set the timings for your RAM. There is probably a memory training error on the motherboard on the latest BIOS versions that prevents the PC to boot. I see that you have 4 sticks of 3000 CL 15. I don't know how the ASUS bios works or if it has some auto memory tuning enabled, I would suggest to turn off anything related to that and you manually set them to 2966 CL15-17-17-35 or even higher 3200 CL16-18-18-36 and don't forget to manually set the DRAM voltage to 1,35V. With four sticks of RAM I would also manually set GDM to on (geardown mode) and command rate to 2T.

[GoLLuM13]

Thanks for your answers

@hawkeye as mentioned already set up manually as soon as I reach 2800 Mhz problems begin, I don't think that has something to do with the number of sticks, I saw people having the same issue with only 2 sticks. I also rised the ram voltage (just for testing) up to 1.42v which is a lot for just 3000 Mhz. I always set Gear Down Mode to disabled, so it won't turn the 15 into 16

@ThreeDee DOCP is nothing but XMP on an Asus AMD platform and in the end of the day manufacturers are free to call it whatever they want (XMP or A-XMP or DOCP) especially that most RAM kits are "Intel XMP Certified" anyway (just look at the box that came with your RAM) or look at this example anyway that's not the problem

And to answer your other message 4 or 2 sticks (even 1) didn't change a thing for me or other people, I went up to 1.42v for Memory Voltage and same thing. Unfortunately, my RAM has only one XMP/DOCP profile
My RAM sticks work like a charm on 3801 BIOS (based on AGESA 1.2.0.3) 4 sticks with 1.35v, I even managed to OC up to 3600 C16 but it wasn't for a daily use just a test, and no error whatsoever. So I really don't think the problem is in the hardware, it's in these buggy BIOS version (and to be precise the AGESA)

Thanks again to both of you and have a nice day

[ThreeDee]

I went thru some wonky BIOS's on my ASRock x570 Taichi and my daughter's x570 Steel Legend (3700x & 3800XT) where XMP would work .. then it wouldn't and I had to manually set what where the XMP timings and THEN it was fine .. it was annoying is all. But not since running 2000 series Ryzen's have I had issues with running my RAM's rated speed

I now run 5000 series chips on mine, and my daughters.

My wife's setup (ASRock B550m Phantom Gaming 4), when I ran a 3700x on it,  didn't have any memory issues .. but I only ran 2 x 16GB sticks kits.. one set was an ECC UDIMM's that ran at 1.20v 3200 .. the other set was 2 x 16GB 2666 stuff @ 3200 (these were on ASRock's 1.2.0.5 BIOS). I've since upgraded her to a 5600x though with 2 x 8GB 3466 stuff. (the ECC stuff went back into my server and the 2666 stuff is in another build that I'm selling/trading)

As far as DOCP(ASUS)/EOCP(Gigabyte?)/XMP(everybody else) .. I was merely pointing out that ASUS is the only  motherboard company that uses that nomenclature when you were "correcting" @jawntem .. not everybody runs ASUS stuff so they call it what their respective BIOS's label it (most times) but we all know what they mean/are talking about/referring to

Your Statement>>"XMP Profile is Intel's terminology, on an AMD platform it's called DOCP" 

[Flyfisherman]

On my system I have only set the Memory frequency to DDR4-3200MHz
No DOCP (Direct Over Clock Profile). Gear Down Mode and all other memory settings are Auto or at its default settings.
And it is working.

With memory frequency increased (above) the DRAM Voltage is automatically increased from 1.2V to 1.35V (this could vary) after DRAM voltage training.
Changes will not take effect until F10 (Save & Exit) and Re-enter Bios if one would like to check the dram voltage.

Mobo: Asus ROG Crosshair VIII Hero (wi-fi) with AMD X570 chipset
BIOS ver. 4006 (with AGESA V2 PI 1.2.0.6b) I don't care for beta BIOS.
CPU: AMD Ryzen 9 3950X (16 cores)

Ram: Corsair Vengeance LPX 2x8GB 3200MHz CL16
LPX=Low profile 33.5mm height so it fits under my Noctua NH-D15 SE AM4 CPU cooler.
@1.35V 3200MHz Tested Latency 16-18-18-36
Part. No. SKU CMK16GX4M2Z3200C16

GPU: ASUS TUF Gaming AMD Radeon RX 6900 XT OC 16GB
GPU Driver: AMD Adrenalin s/w: 22.4.1 (Driver version: 30.0.15021.1001)
SAM: Enabled and fully functional.
Extended monitor setup: Gigabyte AORUS-FI27Q27" 2560x1440 @165Hz (primary), Philips IPS 24" 1920x1080 @60Hz (secondary).
Chassis: Fractal Design Define 7 with 6 140mm fans: Intake one at bottom, two at front. Outblow: two on top and one at rear.
CPU Cooler: Noctua NH-D15-Se AM4 (two 140mm fans).
Storage: M2 Samsung 980 Pro 2TB NWME, SSD Samsung 950 Pro 1 TB SATA, BPX M2 NWME 480GB
PSU: Corsair RM850X 850W 80 Plus Gold certified.
Mouse: Logitech Gaming mouse MX518
Windows 10 Pro 21H2 with latest Windows optional update.

Best regards from Sweden

[Edit] Btw.

XMP (Extreme Memory Profiles) is an Intel tech introduction to automatically set DRAM to intended data rates over and above a base level (for DDR3 it applies to DRAM 1600 and up), DDR4 for DRAM 2400 and up. You enable the feature, it reads the profile off the DRAM sticks and applies it in BIOS.

DOCP (Direct Over Clock Profile), is from ASUS for AMD motherboards and came as the motherboard makers didn't want to pay royalties to Intel to implement XMP on AMD motherboards. It effectively uses the DRAM XMP profile to set up data rates and comparative timings on the AMD motherboards for a variety of data rates.

EOCP (Extended Over Clock Profiles) is Gigabytes version of the ASUS DOCP..

Within the last couple of years, more and more the AMD motherboard makers are simply moving to XMP as it's an industry standard and much easier to keep up in the BIOS than other methods with the proliferation of new DRAM, tighter timings, etc.

[GoLLuM13]

Thanks to both of you

@Flyfisherman I did the same on one of my many tries to make it work with no success, unfortunately.

I tried to manually set memory frequency, timing and ram voltage, as soon as I reach 2800 Mhz problems begin, and it's random, sometimes I can't even access Windows, as soon as I start I have a Q-Code (random Q-Codes) mostly RAM related, but sometimes I have "BIOS Recovery mode" which is totally weird, and sometimes I can access Windows normally but as soon as I restart, back to square one.

And of course If I return to 3801 every problem I'm having disappear like magic.

[Flyfisherman]

Hi,

What part no. on Your Corsair Vengeance RGB Pro 4x8 GB 3000 C15?
Are they in the Corsair memory list for Your exact motherboard and CPU  that they guarantee will work?
Because I can't find your Corsair Vengeance RGB Pro 4x8 GB 3000 C15 in that list above.

If it is this memory kit Corsair Vengeance RGB Pro 4x8 GB 3000 C15 with Part no. SKU CMW32GX4M4C3000C15
then it is only supported by Intel 100, 200, 300, or X299 Series with Core X Series i7-76xx or i5-76xx (codename: Kabylake-X).

Are they not in the Corsair memory compatible list above and not even in the ASUS Memory QVL (Qualified Vendors List) that ASUS has tested, then You have the wrong memory's.
https://rog.asus.com/motherboards/rog-crosshair/rog-crosshair-viii-hero-wi-fi-model/helpdesk_qvl

I recommend to always use a Memory configurator (different manufactures call it differently) from the memory manufactures, to find a memory kit that they guarantee will work on a particular system.

Are the four memory sticks from the exact same memory kit (important with faster memory's and tighter timings)?

Since I have a similar setup with Asus ROG Crosshair VIII Hero (wi-fi), AMD Ryzen 9 3950X, Corsair Vengeance LPX 2x8GB 3200MHz CL16 and have Not had any issues with my RAM, regardless BIOS version except that I haven't tested beta BIOS ver. I find the whole thing strange and I highly suspect that incompatible memory's are the problem, even if it works (perhaps only occasionally) with prior BIOS ver.
The risk of getting corrupt data, problems to read/write to the memory cells are increased and could eventually cause apps & programs even the operating system to crash/hang.

Best regards from Sweden (I'm sorry for my English grammar).
[Edited]

[GoLLuM13]

Hi thanks again for the help,

It is indeed the CMW32GX4M4C3000C15 that I have, but I can assure you it is NOT a hardware issue or incompatibility, I contacted Corsair support before buying them back then in 2019, and they assured me they're compatible, and I also checked the QVL list for Crosshair VIII Hero (Wi-Fi) and they were and still are on the list (check back they're on both PDF for Ryzen 3000 in the link you provided), plus they were and still are working fine since I bought them back then and with all BIOS versions except those with AGESA 1.2.0.5 and superior where I'm forced to lower the speed frequency.

And today I wanted to get all software issues out of the equation, so I decided to fresh install Windows 11 after flashing the new BIOS, same result, Black screen - random Q-Codes F2 - F3 ... (BIOS related) I couldn't even install Windows when I managed to pass these errors, telling me that there are a missing file or couldn't load it or bad signature. I flashed back 3801 used the same BIOS settings (manually), I installed Windows flawlessly, so it appears for me, in my case, it's 100% BIOS related nothing to do with Windows and its softwares or my hardware. Everything else is a constant except the BIOS which is the only variable.

Maybe you got lucky with the 3950X, maybe it is because you don't have any disabled CCD/CCX (just making theories, it's probably rubbish)

(Don't worry your english is perfect, I'm not a native english speaker too anyway )

[darkexile666]

I just went through the same issue with my ryzen 7 3700x, but it was on Bios version with AGESA 1.0.0.4 B and above where my ram started having issues where I could not even post. It drove me nuts as my usual ram settings would not work and would only let me through if it is setup on its default 2133 values. And guess what was the issue? Freaking RGB that are connected on the onboard 5v rgb headers. I moved all the RGB since I had like 5 fans and 3 light strips which could explain it to probably adding laod to the 5v rail or something. Since moving it directly to the psu and using a controller instead, it completely removed the booting issues and I am no longer getting any post issues, before it was stuck with the cpu light before moving on to ram. Worth a shot trying it out to eliminate any issues as part of your troubleshooting. 

[GoLLuM13]

Thank you @darkexile666 for the reply

Unfortunately, I don't have any RGB light connected to the motherboard, I connected everything to a Corsair Pro Commander (I have adapters to connect other brands fans) only things connected to the MB are power cables, CPU Fan and Front Caste Connectors

----

I also contacted the support (french branch) and sorry I don't mean to be mean but they were of no use, they don't understand what I'm telling them, even though I detailed everything, I had to repeat the same thing over and over, in the end they told me, in summary (among other unuseful things because, for what it appears, they haven't understood or even read what I wrote) :
"overclocking the RAM is a form of overclocking the CPU and it isn't covered by the warranty. Just stay on 3801 BIOS version. In new versions, everything seems functional in their standards of specificity. Sorry about how it is for you. If you have any other questions, contact us"

Don't they know that activating the XMP profile is also a form of overclocking ? And my CPU isn't covered by warranty anymore anyway, they would have known if they had read what I wrote. I digress.

Anyway, understand, they won't do anything to fix this (predictable). I wasn't expecting much, but I'm still disappointed.

Thank you everybody for trying to help me, you are good people.
Right now I'm pissed off and totally disgusted with AMD, I think I won't recommend them from now on.

[ThreeDee]

It might be an ASUS thing .. just say'n

It's kind of like being mad at Ford because your Bridgestone tires went flat .. 

Doesn't help you I know .. .. and you may disagree .. but I feel your angst is misplaced. AMD provides whatever AGESA codes to manufacturer's and it's up to them to implement them correctly on their respective hardware. 

[GoLLuM13]

@ThreeDee thanks for the reply, don't worry I'm aiming very well. At first, I also thought that was an Asus thing since I'm an Asus user, and I was focused to look the same issue only on Asus products, but I saw it elsewhere. And as I said in a previous message  (the first one I guess) same problems with other manufacturers like MSI it's not just Asus and this is one example amongst others this person >>here<< with the same version of AGESA and same CPU having the exact same problems and the "fix" >>here<< was after replacing the 3900X with a 5800X3D and even admin/moderator on the MSI forums were saying that AGESA 1.2.0.5 is unstable "with certain system configuration". The 1.2.0.6 was supposed to be the fix, but apparently not for everybody as they (MSI) officially state "improve 1.2.0.5 instabilities" and not "fix the instabilities"

And I also analyzed SMU versions inside BIOS (of MSI and Asus) with Ryzen SMU Checker, it appears that it has changed along the AGESA version for Ryzen 5000 for example, so if there were any anomalies with Ryzen 5000 they were probably fixed or at least partially fixed, while the SMU for Ryzen 3000 hasn't changed since AGESA 1.2.0.5 so any introduced anomaly is still here on 1.2.0.7, and I'm the living example of it.

And I won't talk more about the AMD support and the lack of desire to escalate the problem to a competent technical team. BTW, one of the latest thing they told me to do, was to decrease RAM voltage while I'm going higher in frequency. So, just to please them I tried and no success of course. They're still focused on my RAM while it is obviously not my RAM the problem, it worked since the first BIOS version until the 3801 and still working like a charm on 3801, problems started with 3904 and AGESA 1.2.0.5. And when you have FA Q-Code it means "Invalid recovery capsule" it's BIOS related, not RAM related.

To finish this I repeat, thank you everybody, any suggestion is welcome, and in this case I know I'm aiming very well when I'm talking about AMD and not Asus.

[ThreeDee]

fair enough

[Wimma]

Hello there,

so, I just updated my BIOS to 4021 singe I ordered a 5800X3D which should arrive tomorrow. Currently I‘m running a 3900x. Because I wanted to game tonight I re-configured my BIOS to my standard BIOS setting that I used before with my 3900x and  BIOS version 3602. With 3602 I never had any problems with DOCP whatsoever. Now with 4021 or higher I can‘t set DOCP to its default profile anymore. Well actually I can but there are some weird things happening. First, I get a DRAM error when booting; the system gets stuck on the DRAM debug LED. When I reboot the system posts just fine, which is weird in itself. Secondly after it boots just fine and everything is working nomally iCue says it no longer receives information from DRAM and the RGB gets stuck (CORSAIR Veangeance RGB Pro 32GB 3200). I am currently in the process of reverting to BIOS 3602 to see if that fixes the issue and gets everything back to how it was. I will report back tomorrow with the 5800X3D installed and running 4021 or higher to see if DOCP is possible with that configuration. If so, we know it is a problem with Ryzen 3000. If not, well I have a problem because I don‘t want to have a 3080 with a 5800X3D running on 2133mhz RAM…

[Wimma]

So, like I suspected, this is an issue with Ryzen 3000 chips. Installed my 5800X3D yeaterday and no issues at all… hope this gets fixed for the 3000 users soon

[ThreeDee]

Have you tested with just 2 sticks in slots A2/B2? 

Do you have access to other RAM to test with on the newer BIOS's from ASUS?

Have you tried enabling DOCP and then try bumping up the voltage from the default 1.35v to 1.36v .. and so on up to 1.4v..? (I've read that some with Corsair RAM have had to up their RAM's voltages to even be stable at default XMP/DOCP settings)

When I ran a 3700x on my wife's old B350 Prime Plus .. there were 2 DOCP settings .. I forget how they were worded now, but I think one was standard and that's the one I had use for her 3200 stuff to be stable. (It's been awhile we run all 5000 series Ryzen chips now and ASRock motherboards of whom uses XMP labeling instead of ASUS's DOCP)

[renesdk]

I'm on an MSI X570-A PRO with a AMD Ryzen 7 3700X and this is my exact experience also.

• On the AGESA 1.2.0.3 it is completely stable.
• On AGESA 1.2.0.5 it started doing weird BSOD's at random times.
• On AGESA 1.2.0.6c it seems to be running fine. I could restart the computer no problem, however when I shut down the PC and powered it back on, I would get a BSOD during first boot to do a repair of the Windows install, after every single power off.

Going back to AGESA 1.2.0.3 fixed it all and everything was completely stable.
All drivers have been updated to the latest drivers from AMD.

I'm running  a 32 GB Corsair Vengeance DDR-3200 kit (2x16 GB) and it's a supported model by MSI on their supported list.

I was experiencing the issues both with and without XMP enabled.

[GoLLuM13]

@renesdk sorry to hear that and sorry to tell you that AMD won't do anything about that, at least in my case, they were stuck with the so called incompatibility with 4 sticks of RAM while I told them it worked since the first BIOS version.

Have you tried to set manually your RAM frequency at 2133 Mhz or leaving it on auto? Do you have the same symptoms with these settings? 

[renesdk]

Well, if they won't even work on fixing it (I also have 4 sticks of RAM), they have just lost a customer on the next upgrade I do.
This has worked since launch of the platform and is supported on the MSI matrix of RAM for the motherboard.

It does not work on default speed with the new BIOS versions either. Also the performance loss going to 2133 over 3200 is not only measurable, but can also be seen in FPS, especially when streaming at the same time.
So basically I'm stuck on AGESA 1.2.0.3 with the issues with the TPM on Windows 11, then. That's just great.

I almost knew back when I went with AMD I would regret it. I've never had a single AMD based system that didn't have major issues. They really need to work on their software side of things. That's where they are really lagging behind.

Oh well, and I thought AMD had improved since I had my ATI Radeon 9700 Pro, I guess not.

[GoLLuM13]

@Matisse thanks for the feedback, I almost always disable fastboot (in bios aswell in Windows) and obviously no success with this
I don't know if I have precisely "fast memory boot" option, but I'll look into it thanks

@renesdk I'm pretty sure I won't buy AMD anymore, whether it's for me, my family or my clients and I won't advice anyone to buy any of their products.

And yes we're stuck, and told them that, and they seemed like they didn't care. I'll translate you their latest message. You'll see that even though I detailed everything, talking about Q-Codes related to BIOS they're stuck with memory incompatibility and false claims and some rubbish explanations to avoid doing anything about it.

"Hello Mr xxx,

Thank you for your detailed explanation.

Although your memory modules are listed as QVL by the manufacturer, we remind you that by using 4 RAM memories on the same motherboard, it cannot reach the same speeds that it would reach by using only 2.

As described in the motherboard manual, if 4 memories are used at the same time, the maximum speed without overclocking is 2666 MHz.

This was demonstrated by the tests you performed using only 2 RAMs (reaching 3600 MHz, although that is also OC) than using 4 (reaching 2800 MHz, also above spec).

We also remind you that by using 2 sticks of RAM, the maximum frequency officially supported by the memory controller of your CPU is 3200 MHz.

Therefore, we do not offer support for the configuration you are trying to use, as it is considered overclocking and the equipment performs well when within specification.

If using an older BIOS version (3801) makes your computer stable, you can still use that BIOS version. Since these speeds are not supported, it may stop working or have issues at any time.

If you have any other questions about your motherboard's memory QVL list, we advise you to contact Asus support.

If you have any further questions, please don't hesitate to contact us."

To sum up, my platform worked since day one, Asus tells me on their manual and/or QVL that I can reach 3000 and above even with 4 sticks, but AMD denies it. They claim that maximum speed without OC is 2666 which is false because my NANDs are rated for 2133, They say that I reached 3600 with two sticks and 2800 with 4 which is false too, I reached 3600 with 4 sticks on old BIOS versions, but in the new ones I can't go above 2733 whether with 1 or 2 or 4 sticks (they don't read carefully what we write to them) They try to tell me that the official maximum frequency speed is 3200 while I have 3000 Mhz sticks ? And this is just to put the next sentence, so they won't help "we do not offer support for the configuration you are trying to use" and as you can see they told me to stick with BIOS 3801 (AGESA 1.2.0.3) while I told them and they're supposed to know there is an issue with Windows 11 and fTPM with this AGESA version. And to end "politely" they tell me to contact Asus if I have any question about QVL list (RAM again) while I never asked anything RAM related

[renesdk]

So basically what they are saying it that this config was never really supported (4 RAM sticks at full XMP speed), but worked anyway, and now they've made some updates to the AGESA that changes that and we can now not run full speed on the RAM anymore?

Wow, just wow.

[GoLLuM13]

@renesdk Yes something like that, for them as long as it "works" at 2133 or 2666 it's perfectly functioning and nothing has to be done.

So it appears to me that all their slides during presentations about FCLK 1:1 (up to 1800Mhz), fast memory speeds ... etc are just marketing and false claims

[Matisse]

1. MSI B450 Tomahawk Max, agesa 1.2.0.6c
2. 4x8 Corsair Vengeance 3200 MHz
3. Ryzen 5 3600

On startup AND reboot, my PC shuts down and then turns on if X.M.P. profile is enabled.

I solved the problem by simply disabling "fast memory boot" in the BIOS settings.

[herhey]

@GoLLuM13 

I have the exact same problem and the only solution I found was to roll back to BIOS v.3801.

My specs are:

- Asus ROG Crosshair VIII Dark Hero

- Ryzen 7 3700x (no overclocking)

- 32GB (4x8) Thermaltake Toughram Z-One RGB (3200mhz)

- Asus RX 5700 XT 8g Rog Strix

[GoLLuM13]

@herhey thank you for your feedback and sorry to know that you're having the same issue, you probably know by know that AMD won't do anything about it, unless we're a large group to ask it

In the meantime like you said you can rollback to 3801 and if you're lucky you won't suffer much of the stuttering on Windows 11, or try a slower frequency speed and lose some performance

[goodplay]

Funny thing is a user a couple years ago (posted over 20) on the same issue with same ram modules. during a 12hr memtest faults appeared. In the end ram was replaced, problem solved.

You are the only one currently reporting this issue, so ? 

[herhey]

@goodplay I also have the exact same issues and on Reddit there are many other users reporting the same problems.

[herhey]

@GoLLuM13  Can you repost it on https://linustechtips.com/forum/5-cpus-motherboards-and-memory/  formun and on Reddit ?

[GoLLuM13]

@herhey Sure I can always try, especially now that my last message has been deleted (is this some kind of censorship ?)

[herhey]

@GoLLuM13 Please let me know as soon as you repost on those forums so I can reply and follow.

[GoLLuM13]

@goodplay I am certainly not the only one reporting this error, in this thread only you already have other people with different RAM modules

I already did multiple cycles of memtests without any error and obviously if that was memory modules that was faulty why would it work on all BIOS versions up to 3801 (AGESA 1.2.0.3) and not on 3904 (AGESA 1.2.0.5) and later ?? I couldn't install Windows 11 after several attempts on the latest BIOS and yet the second I flashback the 3801 version it installs flawlessly on the first attempt (like it should work) ?? FA Q-Code (BIOS related) on the latest BIOS versions but never got them and still don't got them on 3801 ?? No it's not a RAM problem, it's not logical, it's a BIOS (AGESA) problem

And the most astonishing part is people fixing this issue by changing their CPU with a Ryzen 5000

And since you want to see more people here, you're well served :

MSI Forum :
>>Here<< the person describing the problem and >>here<< fixing it by changing the CPU
>>Here<< a moderator saying that 1.2.0.5 is unstable and 1.2.0.6 should be the fix (but it's not)

Asus Forum :
Two persons on a TUF B550 with same symptoms, the first one has only two RAM sticks (btw the whole thread is about the same problem) : the first >>here<< and the second >>here<<
Strix X570 E-Gaming, another one swapped his CPU from 3700X to 5900X to "fix" the problem (and there's more from other people in the messages after it) : >>here<<

Reddit :
Two other people >>here<<

An article where they say that AGESA 1.2.0.5 introduced problems, amongst them CPU Performance, memory instability, overclocking errors and more : >>here<<

As you can see, there's many, many more, and I haven't given you all what I saw. And there's probably people that just don't know yet thet have the problem if they didn't updated their BIOS because they didn't know how to, or didn't feel the need to, especially if they're still on Windows 10

[GoLLuM13]

@herhey Done !
Here on LTT Forum : https://linustechtips.com/topic/1431452-ram-issues-depending-on-agesa-version-1205-to-1207/

And here on Reddit (not really familiar with reddit, so I did on 2 different subrredits) :

1- https://www.reddit.com/r/Amd/comments/uquyu1/ram_issues_depending_on_agesa_version_1205_to_1207/
2- https://www.reddit.com/r/AMDHelp/comments/uquyqz/ram_issues_depending_on_agesa_version_1205_to_1207/

[GoLLuM13]

I just found an article on Tom's Hardware saying that AGESA 1.2.0.5 introduced problems, amongst them CPU Performance, memory instability, overclocking errors and more. Also keep in mind while reading that article, SMU version for Ryzen 3000 CPUs hasn't changed since AGESA 1.2.0.5. Which means any issue introduced in 1.2.0.5 is still here.

Toms's Hardware article here : https://www.tomshardware.com/news/agesa-1205-bugs-amd-ryzen
SMU version number evolution : https://i.imgur.com/L2Ny6Gs.png

[GoLLuM13]

So now they decided to change the title of my post here, introducing errors (just like they did with AGESA)

It went from "RAM Issues depending on AGESA version (1.2.0.5 to 1.2.0.7)" to "my memory overclock / XMP no longer stable since updating my BIOS from ageesa 1.2.0.3 to 1.2.0.7"

first thing first, even though you consider XMP as an overclock I didn't really overclocked, it's "AGESA" not "ageesa" and it's unstable in all versions from 1.2.0.5 to 1.2.0.7 included