1700x OC help

I finally decided to try overclocking and could use some advice.
I attempted 3.9ghz and boot into windows everything seems fine.

Performance in pubg got a very noticeable difference.

But after some time I came across some issues. After some time (haven’t noticed any particular interval) OBS will crash while streaming pubg. Someone on OBS forums suggested some issue with ryzen OC and OBS encoding that could do this. Not sure how credible that is.

Eventually pubg started crashing too… but to be honest that’s not exactly anything new because pubg seems to do that pretty randomly between all my friends, and we all have different builds.

Finally windows did crash and I got a blue screen, this was after several hours though of up time.

I decided to run prime95 and came back after a while. I had the ryzen master suite open to monitor temps. When I came back it was 69c which I thought seemed reasonable, but then I noticed that my mouse was not responsive. It seems to me the whole pc crashed and the screen was just stock on whatever the last frame drawn was.

I assume that’s probably the more positive confirmation that the OC is not stable right?

What are you guys getting with this particular cpu and which mobo do you have?

I have a gigabyte ax370 gaming k5 and trident z memory with xmp at 3200
My bios is F22. The most recent on the website as of last week.

normally i push it to where i start seeing problems, usually by running fx. prime95.
then back off 100Mhz. and then start fiddeling with ram.
My current Ryzen i hit 4Ghz “somewhat stable”, but it starts being sortta meh when pushing it,
so i backed down to 3.9 ,then i clocked my RAM ~2800Mhz(from 2400Mhz). And from there just leave it be.
Havent touched voltage or anything.
Basically it doesn’t really matter what temperature your CPU shows, it matters wether or not it throttles while showing
~69 degrees.
IF those 69 degrees were the result of running 6-8 x a throttled 900Mhz. Then you pushed it to far.
Also what cooling solution are you using?

Also, when testing with Prime there is a difference between running it with default parameters or adjusting them accordingly :wink:

Are you trying to achieve something by this overclock, or are you just overclocking for its own sake?

Turn off any kind of energy saving modes, C-states management and Auto settings.
Make sure the memory gets 1.35V. Also keep an eye on the actual clocks, XFR can be a bitch.

Make sure LLC is turned on. I get pretty significant vdroop without it = instant crashes. I run my 1700X at 3.95GHz and 1.375v with level 4 (second highest) and get 1.362 under full 8-core load at that setting.

So, in OP I was at 3.9 @ 1.45v

I reset everything on my motherboard and attempted 3.8 oc without any other setting changes. No voltage increase etc.
I booted up and was able to run cinebench no problem. FWIW stock voltage is 1.35

I ran prime 95 and after about 10 minutes the the system became unresponsive. It didn’t shut down, it just seemed to no longer accept input or draw new frames. This is basically the same as when I was when I first OC.

I added .025 to dynamic vcore, rebooted. Prime 95 crashed. Shortly after I tried cinebench and the whole system froze.

I proceeded to add .5v. Ran Cinebench, it crashed. Prime 95 crashed after about 2 minutes.

Cooler is Corsair H110i AIO, presently just one fan on the radiator thought because of mountain space issue (used an old case and because of the cpu power cables and heatsink at the top of the mobo, I couldn’t properly mount the second fan). I can jerry rig attach it with zip ties if I need to.

I don’t believe my issues are heat related. If it was it, wouldn’t my pcjust shut down rather than freeze?

Mobo settings
Core performance boost - disabled
amd cool and quiet - disabled
SMV mode - disabled
Global CState - disabled
power supply idle control - disabled
opcache control - auto
down core control - auto
smt mode - auto

P.S. at 3.9 1.45v with global cstate enabled I can at least run cinebench. But with it disabled I keep getting errors.
Also, I just confirmed that when the PC freezes, it’s just the video that stops drawing new frames. The most recent time this happened the screen froze but I still had audio in teamspeak and I was still able to hear and speak back.

I would firstlly dial back the core voltage to something like 1.40V to start with.
And then see if that is any stable atall.
Also do you have a decent airflow arround your vrm area?
Because if you want to pump 1.45V into a 8 Core cpu,
then you are definitelly going to need that on your particular mobo.

its the X370 Gaming K5 we talk about here.
And not the X370 Gaming 5.
Those 2 particular boards differ significantlly wenn it comes to the vrm implementation.

1 Like

I’ve tried different voltages. Maybe I just got the short straw. Seems like best I can do is 3.8 at 1.45 :neutral_face:

Wow, that is terrible. Honestly I would maybe even try a different board.