Question Tried to overclock my ryzen 1700x and pc shut off

MechaNintendoMaster

Distinguished
Jul 22, 2010
268
0
18,810
6
I have:
Ryzen 1700x
Msi b450 pro carbon ac (second to lastest bios) Fractal design focus g (140mm intake, 120 exhaust)
Cooler: Deepcool game max 400 (or something)
PSU: Seasonic focus gold 750 watt
GPU: EVGA GeForce 980 Win 10.

I looked a lot of guides and videos and I thought these settings would be okay.

Clock: 3.8ghz CPU vcore: 1.35 to start. My motherboard called it ‘override mode’ to enter your own value. There was other voltage things to mess with but I didn’t. RAM: Set to an XMP profile at 2933 or something like that. I was running this before overclocking so I knew that was already stable Load line: Set to level 3 (because an msi guide said it)

Story: I didn’t touch the C states settings (should I have?) and the cool and quiet feature was already off. Windows could boot just fine. I ran prime95 small fft for a few moments and when returning to the room the pc had shut off. It wasn’t a black screen it was actually powered off. I was using Hwmonitor (yes I’ve heard that Hwinfo 64 is better by this point but that’s what I had then) and no temperatures seemed bad unless they scaled up dramatically in five minutes. Event viewer didn’t have error other than the one when the pc doesn’t shut off correctly. I know it’d be better if I had exact numbers but that’s all I can say.
 
Yes, your OC is unstable. That's what happens.

This is a good guide: https://forums.tomshardware.com/faq/cpu-overclocking-guide-and-tutorial-for-beginners.3347428/

But just a point to make. Reading the guides is no use if you just dial in settings and hope for the best. When OC'ing you start off low and work up. Follow the guide/process. Test at each stage. This will give you stable, repeatable results. Take your time doing it.
Start at 3.6 with 1.3v. Run Prime5 small FFTs. It needs to run for a min of 20-30 mins to get the processor up to temp. If the temps is too high, or there is not enough voltage Prime will have failures. Then you go back and bump the voltage one notch at a time, and test again. It's time consuming. But can be worth the effort when done correctly.

Edit: I'd leave the LLC on it's default setting. Changing it up to higher settings is really for crazy OC's. No need to mess with it right now, as it can over volt the CPU, which can make getting a stable OC difficult.
 
Reactions: MechaNintendoMaster

mamasan2000

Honorable
Jan 13, 2014
1,538
19
12,665
227
You have an air cooler. 3.8 Ghz should be AIO-cooler territory (water). Your CPU might have overheated.
Unless you turned on Logging in Hwinfo, it doesn't save the temps etc. So what you are seeing is what the highest temp is since you started Hwinfo. Not since last reboot or since forever.

My 1700 non-X at 3.8 Ghz and 1.34V gets up to 70-80 C with a corsair 240mm radiator. Air cooler wouldn't be enough in my case.
 
You have an air cooler. 3.8 Ghz should be AIO-cooler territory (water). Your CPU might have overheated.
Unless you turned on Logging in Hwinfo, it doesn't save the temps etc. So what you are seeing is what the highest temp is since you started Hwinfo. Not since last reboot or since forever.

My 1700 non-X at 3.8 Ghz and 1.34V gets up to 70-80 C with a corsair 240mm radiator. Air cooler wouldn't be enough in my case.
Your voltage is quite high for 3.8 with your chip. As a result your temps are bit high too with that cooler.
 

MechaNintendoMaster

Distinguished
Jul 22, 2010
268
0
18,810
6
Today I tried for 3.6 with 1.300v and ran the prime small fft test for a few minutes. I stopped it because my temperatures were in the 80s quickly. Was that okay? I didn' think so since it had just started.

Edit: And for reference. Running the tests at stock only puts me at around 60c.
 
Last edited:

MechaNintendoMaster

Distinguished
Jul 22, 2010
268
0
18,810
6
Update. Tried for 1.275 and got lower but still near 80c temps and idled at 40ish . I’m going to retool my cooling set up soon with some better case fans and reapplication of thermal paste. Until then I’m holding off.
 

mamasan2000

Honorable
Jan 13, 2014
1,538
19
12,665
227
Try some real-life benchmarks, programs you use. Handbrake maybe.
Stress tests aren't that indicative of real life. I can't run OCCT for more than 5 mins before CPU overheats but if I run Handbrake and stress all cores to the max for an hour, temps will be 10-15 C cooler. Every time.
Stress tests are usually way over the top. Good for stability testing, not so good for temps in general.

If you game, games use at most 50% of my CPU (Ryzen 1700). Way lower temps.

Paste and casefans wont do much, at best 5C if you have no or really poor cooling.
If temp bothers you, start looking at watercooling, AIOs to start with.

Air coolers can only cool so much, once you go over that barrier, efficiency of air cooler drops off a cliff.
Temps rising like mad.
Relevant info
View: https://youtu.be/cFICTlMZWiY?t=926
 
My CPU isn't stable under lower voltage. Have to use LLC level3 or PC blackscreens under load.
Cinebench R20 crashes instantly.
Have you tried offset in stead of LLC. play around with a higher voltage with a minus (-) offset to bring it to what you want. I find i don't have to change LLC at all by using the offset.

I've my 1600x at 3.9 1.3v Prime small ffts stable for 8 + hours. Temps will stay steady at about 72-75c.
 
Update. Tried for 1.275 and got lower but still near 80c temps and idled at 40ish . I’m going to retool my cooling set up soon with some better case fans and reapplication of thermal paste. Until then I’m holding off.
It may just be the cooler isn't enough for your CPU. It is a budget cooler, and as such, you won't get phenomenal performance. At 1.275 even with a moderate OC, you should not really be hitting 80c+ IMO.
 

mamasan2000

Honorable
Jan 13, 2014
1,538
19
12,665
227
My AIO just died which made my PC and OC unstable, including memory OC. I have Corsair 3000 Mhz kit running at 3333 Mhz.
CPU reached 85 celsius and started throttling or PC hanged. Problems all around.
So if you are worried about temps, look out for those symptoms. Hanging, blackscreen, bluescreen, unstable.
The AIO was old, 5+ years. It was time.

Here is my memory OC settings for anyone interested. I don't see a lot of people posting Hynix CJR settings.
https://www.reddit.com/r/overclocking/comments/c6wzxk View: https://www.reddit.com/r/overclocking/comments/c6wzxk/cmk16gx4m2b3000c15_3333_mhz_asus_prime_x470pro_r7/
 

MechaNintendoMaster

Distinguished
Jul 22, 2010
268
0
18,810
6
It may just be the cooler isn't enough for your CPU. It is a budget cooler, and as such, you won't get phenomenal performance. At 1.275 even with a moderate OC, you should not really be hitting 80c+ IMO.
I haven’t gotten notifications either! Anyway that 80c temp was strictly using the prime smallfft which is supposed to generate more heat than normal. I don’t think normal usage would have hit that high but I was cautious regardless. Do you really think my cooler isn’t that good? Maybe my thermal paste sucks.

I haven’t messed about much because things came up and then in the midst of this I got completely unrelated GPUs issues grabbing my attention away from Overclocking. If I get something going I’ll update. Thanks for any help so far.
 
No worries mate. Let us know how you get on. Yes, Prime will give a lot of heat. But that's the point. it give you max temps at a specific voltage. If it doesn't fail prime workers, then it's pretty damn stable.

Yes, you can use other software to generate the load, but when I OC i want to know the max temp, per given voltage. Those who say you leave performance on the table testing so stringently, are not OC'ing their systems right. They will suffer random crashes and instability.

Do it right, and you will avoid any headaches. If you need any further help, you can PM me :)
 

ASK THE COMMUNITY

TRENDING THREADS