Note that I am working on this now. I will test at a fixed 1.40 VCore and an LLC of 3. the chip may work with lower voltages, but tinkering with that would take far too long. I'll do at least 1 test run at 'small' settings. Once I find the highest achievable, I'll try 1.4375 and see if i can get it any higher. I will test 6,8,and 12 cores, although i can test 4 cores if requested. After trying a few diff freq combinations and finding the best one, I can tune the voltage and do extended testing.
EDIT: I will also edit this post with links to the full size images of each test result.
Oh and another edit for
@The Stilt, The only limitation I've had thus far with overclocking my chip has been the number of amps pushed through the package. I can be at 1.4V with an aggressive LLC and have the chip throttle due to hitting a hard limit of 272 amps. Attempting to disable prochot stops appears to stop the throttling, but in some cases it causes system lockups and I worry about why prochot is fired off to begin with. Temps look normal, including VRMs, so I'm at a loss in this regard. Then again, MSI...
Another Update: After Mark's post, I decided to lower the core count to 4. Also my board is doing funky things. I think it's linked to Ryzen master, so I am testing 4.3GHz currently at quad core, but might need to reset CMOS before I continue testing. (this happens ALOT on msi boards. If i set a voltage in BIOS it doesn't stick, so i end up doing it in Ryzen master, then more stuff breaks, then the system stops rebooting, etc. They really need to get more developers on the BIOS stuff, the last BIOS build failed to properly read CPU temp, causing fan profiles not to work right)
Current Status: Done for now. Really need to reset CMOS to continue testing. Until then I will leave you with a teaster momentarily.
4 cores, 4.25 GHz:
https://s3.amazonaws.com/anandtech/425@4.PNG
6 cores, 4.2 GHz:
https://s3.amazonaws.com/anandtech/42@6.PNG
Reminder: Take voltages/temps with a grain of salt. I'm not testing minimum voltages/moderate LLC settings here. The goal is to find the highest 'usable' speed for a given core variant.