Adventures in 1700 Overclocking. Somewhat success

The UEFI definitely plays a role, I could get 4.05 stable on my 1600x with 2.3 on my Taichi but 2.4 will only just net me 4.0. At the moment I think UEFI updates are more tailored toward RAM frequency, once that is more or less sorted we should see other things take focus and improve. 

I love how much Ive learned, and have gotten good guidance from this forum, YouTube, and elsewhere. I'm thinking I simply have a BIOS limitation at the moment...rather than a chip that cant overclock higher.

But I'm hopeful for whatever AsRock has in store for us soon in terms of BIOS Updates

Preface: I'm writing this in a way that new overclockers can understand. Which would have helped me last week as I'm still learning this myself. My CPU is a Ryzen 1700 with an AsRock x370 Killer SLI/ac motherboard.

I personally define a successful overclock as one that is stable through numerous benchmarks, and allows for auto downclocking and downvolting when the CPU is not in use

Note the following before reading on:

Pstates - Think of them as power states that allow for more customizable overclocks. If I used the simpler "OC Tweaker" menu in my BIOS, and made a simple two-click change of my CPU frequency and voltage...Ill end up with a Ryzen CPU that ran my overclock at its highest frequency all the time. Pstates allow for control of downclock on your CPU, thus increasing energy savings and CPU life span. You will see several Pstates in your advanced settings.

FID - Frequency value in each of your Pstate settings (uses hex values)
DID - Divider value in each of your Pstate settings (uses hex values)
VID - Voltage value in each of your Pstate settings (uses hex values)

For a little help with these hex values, check out the description of the Ryzen overclocking guide that Youtuber Tech City did a few months ago. The video can be helpful too...although its fairly dated now given the numerous BIOS updates we've had since.

https://www.youtube.com/watch?v=52Tw-wcT7o4

------------------------------------------------------

Here are my findings:

3.7ghz @ ~1.3v was a success

I get a 3.7ghz stable and stress tested overclock with proper downclocking to 2.7ghz and 1.5ghz by doing the following:

-when going into Pstates, I change Pstate P0 to Custom, and ONLY change the FID value in order to up my frequency to 3700mhz. If I touch the DID or VID values at all, I end up with Windows getting stuck at 1.5ghz downclock, and a max 2.7ghz boost. Since I don't change DID or VID, the voltage in P0 gives me 1.1875 volts. In order to get the voltage I need, I go to OC tweaker and set the Vcore offset to +130mV (BIOS HW Monitor will now see total CPU Vcore around 1.3v)

-ALL OTHER Pstates are left at AUTO. Changing the other Pstates (usually P1 and P2) never got my overclock to be recognized properly, with correct downclocking. I would either get stuck at 1.5ghz, stuck at 3.7 ghz, or get 1.5ghz downclock with a 2.7ghz boost. So for now its best to only edit Pstate P0 and leave the rest on Auto

I can run Cinebench multiple times with no issues and temps in the 40s or 50s. I max out at 60 degrees Celsius in AIDA64 stress test, about 70 or slightly less in prime95, and about the same in Intel Burn Test.

----------------------------------------------------

3.8ghz and up was a failure

I am unable to test 3.8ghz or higher overclocks due to what I believe is currently a BIOS limitation as I don't get close to a chance of getting proper voltage or even stressing high temps.

-If I try the same method as above, but simply change my FID in Pstate P0 to 3800mhz, my system crashes during my second Cinebench test. At 45 Celsius, temps aren't the issue and clearly I need more voltage, as HWiNFO monitor shows me getting the same 1.3v or less of juice, which I set for my 3.7ghz overclock. And here's where things start to suck.

-If I instead set my Vcore offset in BIOS to +200mV, in order to give myself the extra needed voltage for my 3.8ghz overclock....the system will not load Windows. It will POST, show the AsRock logo, and then the monitor goes off, my CPU fan ramps up, and Windows never loads.

-I can get back into BIOS if I reset, and check to see that my Vcore has indeed increased to 1.35 or so thanks to the increased offset. But I cannot load Windows. VERY STRANGE, considering I could load Windows just fine with the lower offset, and thus lower voltage.

-So I figured, let me try setting the Vcore offset back to Auto, and instead trying to set the voltage in my P0 pstate. Once I changed the VID to allow for me to have 1.35v or 1.36volts, I can load Windows. Remember, I am still leaving my other Pstates at Auto.

Windows and Cinebench list a 3.8ghz CPU in my specs, however, Ryzen Master, HWiNFO, and other monitoring programs read my Max CPU frequency as 2.7ghz. When I run benches, this is exactly what I max out at. And thus all my scores go down compared to my earlier successful 3.7ghz overclock.

-Finally I decided to set my Pstates manually. Changing the FID(frequency), DID(divider), and VID(voltage) values for Pstates P0, P1, and P2. I set P0 to 3.8ghz with 1.35v, P1 to 2.7 with 1.2v, and P2 to 1.5ghz with 0.875v.
And guess what? YET AGAIN, my PC gets either stuck at 1.5ghz, or only boosts to 2.7ghz.

Make what you will of my findings and please chime in.

Edited by CocoaThumper - 04 Jul 2017 at 1:50pm