Single core overclock causing multicore software to crash

Associate
Joined
30 Aug 2017
Posts
38
So heres something thats got me stumped.

I have an i7 7820x with ASUS Prime X299 Deluxe mobo and G.Skill 3200mhz 32gb ram.

With testing i can get a stable overclock of [email protected]. However, as we know some cores are better than others. My core 0 is always 10-15 cooler than my other cores...the intel turbo boost max 3.0 software identifies this as the best core also. So off i went into the bios and made these changes:
1) changed from sync all cores to by core multiplier and voltage adjustment
2) Changed core 1 (asus bios starts at 1 not 0 - so this is core 0) max multiplier 50 and voltage to 1.285v.
3) all other cores remain on auto.
4) windows boots fine - run HWiNFOx64 to monitor temps and multiplier and voltage
5) run cinebench - single core test - and runs fine. Run it multiple times and fine. In HWiNFO i can see core 0 being 100% utilised@5ghz and 1.285v.
6) run the multicore CPU cinebench and it starts and then crashes. sometimes cinebench multicore will run once. the second time (running immediately after) causes crash. By crash i mean the screen freezes and fans die down to normal levels.

so...what the jack?
 
Clearly the Cpu is stable with that core at 5ghz when only that core is used and becomes unstable when it's multicore. Not too hard to understand with FIVR being used and ofcourse the demand on the onboard vrm being way different in each circumstance.

I tested a lot and 2 ratios is the average you can push higher not 5 for one or two good cores.
 
Clearly the Cpu is stable with that core at 5ghz when only that core is used and becomes unstable when it's multicore. Not too hard to understand with FIVR being used and ofcourse the demand on the onboard vrm being way different in each circumstance.

I tested a lot and 2 ratios is the average you can push higher not 5 for one or two good cores.

Ah right - thanks for taking the time to reply. That makes sense!
 
Back
Top Bottom