So, 18 months ago put my 7920x under water, direct die. Pushed to all core 4.9 on 1.325v. Had pushed it to 5.0 on 1.375 for a few top scores, but daily at the 4.9 1.325. Ran stable at 4.9 for 6 months and got random crashes. Windows going to sleep or Lightroom, which has known issues with x299. Couldn’t tell if windows but dialed the oc back to 4.8 and good again mostly. Proceeded to repeat the issue down to 4.6 after some random crashes again. None ever in a video game. Then a few weeks ago I was getting constant crashes when raiding and then simply loading into WOW. Before it was random locks. Not repeatable. Now, every single time I would crash.

So went back to stock on my oc and was good for about a week. Then repeat. Crashed while gaming. So put all cores down to 3.7 and offset +.1 and it’s been stable. But the computer doesn’t behave right. When gaming I get pauses in the game every 3-5 seconds. It lags randomly and I expect a freeze but comes out of it.

So, before I drop $700-800 on a new x299 chip that’s behind ryzen, wanted to confirm it looks like my first time of true OCing was a learning lesson as you shouldn’t leave the vcore locked at 1.325 and you need to use offsets. Does my pc behavior match that of a cpu that’s been cooked long term?

Its seems i am not allowed to link to my parts listing.  If you check my account, i have 2 posts, this one and the one with my build in it. 

Also what has made you determine it's CPU causing crash and not memory, hard drive, or system/driver issue causing crash?

sparetimepc
So is this cooled with an AIO cooler or s custom loop? What are your temps when crashing? Have you made sure your radiator is not plugged with dust and has severe air flow restrictions? Have you checked your thermal paste to make sure it has not degraded? I ran my 7980xe for almost 3 years all cores at 4.9 with 1.395v and all I ever done was blow out radiator and reapply thermal paste once a year.

Custom loop, using direct die and LM.  I was able to get some prime95 stress tests off at stock and they cores were never above 55c.  Based on that, I assume the cpu is getting solid cooling still, though my loop does need a cleaning. 
 
Were you manually locked at 1.395v?  Cause I was at 1.325 and fear that was what might have fired it.  
 

sparetimepc
Also what has made you determine it's CPU causing crash and not memory, hard drive, or system/driver issue causing crash?

 
Great question.  Been leaning towards that because the period dropping of the OC has helped with stability until i am now down at 3.7.  I am due for a new windows refresh as I am at 18 months on this one.  Will try that today.  I have been lazy and not checked memory, kept meaning too but forgetting.  Will do that now before the new windows.  Just confirm its none of the sticks causing the problem. 

rjohnson11
If you happen to have a spare hard drive around disconnect your primary system boot drive. Install windows fresh onto the spare hard drive and make sure BIOs settings are at default. See if the freezing goes away. 

Yea, going to reinstall windows.  I did update my bios as my first attempted fix to see if my bios and windows were fighting.  
 

Damn, thats cooking.  
 
So, might be the ram.  I pulled all but one stick, put the OC up to 4.4 and xmp back on.  No other tweaks to OCing, booted into windows and was stable.  Hit it with some cb15 (was causing crashes before), no issues.  Loaded up wow and no issues and the period pauses were gone.  Back to smooth.  My FPS also were back at my monitors max.  Before it was sitting at 10 to 20fps lower than the max before.  Even when i cranked the graphic setting down.  
 
So, running through the ram sticks one by one with Memtest86.  See if any throw faults.  Using the same slot for the test, so if they each pass, then add them all back and test them in their slots.  

If you set your ram to xmp and it's booting fine I would just leave it alone except maybe bump the voltage a bit, like if it's 1.35 volts set to 1.4 or 1.4 go to 1.45 it's not gonna hurt anything. Then don't do anything else to the memory and just raise your CPU little by little. Changing both memory and CPU overclock settings at the same time makes it harder to figure out where the issue is.