That makes the whole thing even harder, there's like no information on Google on how certain boards should react to certain overclocks, which boards are good, bad or ugly. I don't even know what some of the settings in MSI's bios mean, there's some like RAM termination Ohm (resistance) value, nobody knows what that is.
I don't touch what I don't know, maybe it's the key to get the board stable, who knows? 
What really annoys me, once I'm past that "bios boot barrier" the whole system runs so freakin fine and smooth, the whole day not the slightest issue. Then I reboot and am back in OC hell.
There has been an abysmal lack of accurate and useful coverage of this Ryzen launch.
There are some useful reviews of the various mainboards build quality but it's slow in coming. The buggy BIOS are still being developed by each manufacturer, but you hit the nail on the head when you called out MSI's VRMs for the OC. The DRAM compatibility is still being worked on with about 30 new models added to the QVL list today for my make/model X370. Shopping for RAM deals has been a nightmare because currently if isn't on the QVL it might not work at all.
I suggest Buildzoid teardowns and identification of the parts and check his YouTube link also:
Just a quick update: Do not buy this board!
MSIs support is worse than the board itself.
Quote:"Your question has responded" <- LMAO title of MSIs support email. Almost as good as "The API not find" <- an actual error message from one of MSIs software tools..
One mail they tell me the RAM is incompatible (it is not), next they want me to test another CPU cause it could be the CPU. What is it now, guys?
I've asked nicely for a replacement and also nicely if this could be a "Ryzen Beta Bios" issue - they didn't want to give me a warranty replacement and also didn't want to answer that question. But all their beating round the bush tells me that it's probably not a single production failure but a general problem of that board.
Be warned.
Could be happenstance, but I've been seeing similar reports of less than optimal performance from this board, particularly the VRM. Also, bad Quality Assurance as too many users are reporting their product was DOA, or "POST not find.".
Memory overclockability and compatibility is mainlly a bios thing.
It doesnt really have that much to do with the memory VR.
Because DDR4 memory in general doesnt really pull much power.
So its highly unlikely that the memory vrm is limmiting you.
By chance I found this reddit today, maybe it was already known to some of you: https://www.reddit.com/r/Amd/comments/66ggjd/psa_about_ryzen_and_memory_compatibility/
It took me about 3 minutes to read this post and use the info to fix my problem*. Something I failed to do in 2 weeks before practically operating a "black box", also the support of MSI failed to do in the same amount of time, as well as the tech support of my hardware dealer.
Also it dawned on me that in the MSI BIOS what they call "CPU NB Voltage" probably isn't the NB Voltage, but the SoC Voltage. Couple of things tell me that:
1.) It caps at exactly 1.2V like the SoC Voltage.
2.) It doesn't actually raise the NB Voltage if the HWMonitor reading is correct
3.) The hint text in the BIOS vaguely says it's can "stabilize" overcocking
4.) Turning this Voltage too high has an destabilizing effect, as well as turning it too low (as described in the reddit for the SoC Voltage)
5.) MSI cannot translate
Overall a very disappointing experience with both the MSI product as well as their support. A waste of money and time on my side.
*(If anyone's interested in the details just tell me so please)
LOOOOOOOL
Which makes sense since there's technically no North and Southbridge anymore. What used to be the Northbridge was moved into the CPU with the Core Series from Intel and the Southbridge is now the PCH. So yeah, makes sense even though the wording is wrong 
Also pretty sure they just ran their UEFI through Google translate 3 times.
My understanding that the chipset is the Southbridge. Just a matter of semantics I'm sure.
I agree CPU NB Voltage is the SOC.
Most people still refer to it as the Southbridge for easier understanding, but it's technically the PCH now, so yes, it's basically just sematics at this point.
Yeah a slight SOC voltage bump might improve stability wenn it comes to memory overclockings.
X99 in the beginning basiclly had a similar issue with memory.
DDR4 is still pretty finnicky, on X99 you had to tweak SA voltage to get things stable on some motherboards.
Hi All,
This isn't specific to Linux so please point me in the right direction if I'm not in the right place. I've got this board and I'm able to get a good OC at 3.8ghz with 1.35v and that's been setup in my BIOS. The problem I'm having is that windows doesn't recognize the OC and I have to manually enter those values into Ryzen Master for windows and CPU-Z etc to recognize the overclock. Any ideas what may be causing that? I'd rather like the BIOS handle the overclock, but right now RyzenMaster seems to be the only thing that allows the OC to 'stick' and I have to reapply those OC settings each time the PC restarts.
I'm happy to provide any additional detail.
Well i would say safe to OC profile in the bios, and then go to stock settings.
Boot to windows and do a Cinebench R15 run.
Then go back into the bios, load the saved OC profile, and do the Cinebench run again.
That way you could determing if it might be just a software readout bug or not.
That is so bizzare.... I ended up having to reset the BIOS using the CMOS jumper to get it back to stock settings. That worked and I ran cinebench. Got a score of 1399. Went back to the BIOS, reloaded my OC settings, and rebooted into Windows... this time CPU-Z and HWinfo both reported the higher clock speed without having to use RyzenMaster. Ran cinebench again... 1636. Then rain AIDA 64 for 30 minutes to validate stability... PROFIT!
I may play around with it some more once I get my new RAM but thanks very much for your assistance!
Had the exact same issue in the beginning. BIOS showed 3.8Ghz, HWMonitor showed 4-6 cores at 3.2Ghz and 2-4 cores at 3.8Ghz.
Only with Ryzen Master I was able to oc all cores.
After couple of reboots and oc try-outs it fixed itself. My less-than-educated guess on this phenomena is that AMDs XFR didn't really turn off until targeted with Ryzen Master.
Ryzen automatically turns off XFR unless you do a workaround to keep it enabled:
"When the "OC Mode" is activated the SMU will disable the voltage controllers, which among other things disables the automatic voltage offsets. This can create an illusion that the power consumption is heavily increasing due to the use of "OC Mode". While technically it is accurate, it is more of a consequence than the actual reason. A vast majority of the increased power consumption comes from the now disabled automatic negative voltage offsets, which causes the actual CPU voltage to increase anything between 50 and 150mV. Because of this behavior, it is advised that the user doesn't increase the CPU voltage right away (when overclocking), but only upon actual demand (as usual).
One of the major down sides of the "OC-Mode" is that upon activation both Turbo and XFR will be disabled as well. Basically, this means that unless you are able to reach at least the default MSCF / XFR frequency on all cores, then you will essentially be losing single threaded performance compared to stock configuration."
While I'm interested in having XFR on top of a manual overclock, you run some risk if you already pushing higher voltages as XFR Boosts have temporarily bumped voltage past 1.45v. I don't know how high this would go if you were say already at 1.4375?
Hmmm, good info, but keeping my voltage at 1.36 should give me enough headroom to be fine with XFR enabled correct?
I currently have it disabled.
You have headroom, so I see no reason why you shouldn't take advantage of this feature.