Computer shut off during stress test.

[Ace McPlane]

Hi,

I stress tested the CPU overnight with no problems the first time.

I then overclocked my memory and tested it with memstest86+ with no errors.

I ran my prime95 stress test again after the memory overclock and it shut down the computer at some point.

I have another one running now where I increased the ram's voltage by a little bit to see if that fixes it, but in the mean time...

do you guys have an idea of what to do next?

I had this issue before the last time I overclocked and I can't remember for the life of me what fixed it. I think more voltage to the CPU even though it's the change in the memory settings that caused the failure.

 

Anyways, I logged my hwinfo data and have a spreadsheet, so if you want to look it over and see if you can find a clue I didn't, be my guest:

 

ASRock z390 taichi, i7 9700k, asus strix gtx 1080 oc gaming edition, 3200mhz Corsair vengeance pro RGB, 500gb samsung 970 evo, etc.

 

5.1ghz overclock on CPU, 1.28 vcore, cache 4.6ghz, level 2 LLC, ram: 3600mhz 1.38v 18 19 19 39.

 

SPREADSHEET Download: prime95 log 9700k oc test 5.1ghz cache 4.6 1.28v llc level 2 ram 1.380.CSV

[Silver47]

Once overclocked thats a pretty power hungry system. Is your psu up to the task?

I dont know about oc'ing the 9th generation parts, but 1.28v seems like not much for a 5.1 GHz overclock. Is that definitley enough?

Why only 4.6 GHz on the cache though? Roman (der8auer) says a general good rule of thumb is +/-300 MHz from your core frequency, no less.

[Ace McPlane]

6 minutes ago, Silver47 said:

Once overclocked thats a pretty power hungry system. Is your psu up to the task?

I dont know about oc'ing the 9th generation parts, but 1.28v seems like not much for a 5.1 GHz overclock. Is that definitley enough?

Why only 4.6 GHz on the cache though? Roman (der8auer) says a general good rule of thumb is +/-300 MHz from your core frequency, no less.

I have a corsair ax760 so a 760w 80+ platinum unit. 

Should be more than enough.

That 1.28v is before the load line calibration and all that. What I actually set in the bios.

It's actually reading at around. 1.375 in hwinfo.

 

Oh, and I actually got the 4.6ghz from a der8auer video. He said that's about average.

I can technically get it to 4.7ghz, but the CPU temps get a bit high with the voltage I need for that to pass so I rolled it back 100mhz and used that extra headroom to overclock the ram instead.

[Ace McPlane]

Nevermind, someone turned it off on me assuming something was wrong because it was loud without asking me first.

 

[xg32]

7 minutes ago, stateofpsychosis said:

Nevermind, someone turned it off on me assuming something was wrong because it was loud without asking me first.

 

might as well bump it up to 5.3 and use an avx offset.

[Ace McPlane]

21 minutes ago, xg32 said:

might as well bump it up to 5.3 and use an avx offset.

Nah I already have an avx offset of 4 and trying to boot at 5.2ghz crashes. My temps are already in the 80’s at 5.1 as well so I’m out of headroom. I got bigger fish to fry now anyways. Windows is full of freaken errors in the reliability history to do with the audio driver. I’m at the point where I’m reinstalling windows for god sakes. It’s not my day.

[Ace McPlane]

Update: I needed to fix another issue where windows explorer was constantly using 20% of the CPU even during stress tests and I needed just a little bit more vcore. It's at 1.3v now. Stable after 5 hours, but still running it. Should be fine now. *knock on wood.