F5F Stay Refreshed Hardware Desktop System fails immediately when time slows after high demand

System fails immediately when time slows after high demand

System fails immediately when time slows after high demand

W
wewerock2002
Junior Member
2
05-13-2020, 12:02 AM
#1
Hey everyone, I moved from my 7700k to a 10700k and ran into a strange problem. After starting Windows Energy Management with Balanced settings and running Prime95, everything worked fine until I stopped the test. Then my system crashed. It’s running all cores at 5 GHz, which is a bit over the head... I didn’t see any AVX offset or LLC 3 issues, and my Vcore is at 1.26V. I tried setting Windows Power Management to Ultimate performance, which stopped the crash under heavy load. The temperatures don’t seem to be the issue either. Anyone know what’s going on?
W
wewerock2002
05-13-2020, 12:02 AM #1

Hey everyone, I moved from my 7700k to a 10700k and ran into a strange problem. After starting Windows Energy Management with Balanced settings and running Prime95, everything worked fine until I stopped the test. Then my system crashed. It’s running all cores at 5 GHz, which is a bit over the head... I didn’t see any AVX offset or LLC 3 issues, and my Vcore is at 1.26V. I tried setting Windows Power Management to Ultimate performance, which stopped the crash under heavy load. The temperatures don’t seem to be the issue either. Anyone know what’s going on?

E
EHB2112
Member
136
05-13-2020, 07:34 AM
#2
This only occurs when OCed, right? That means you probably need to set up an LLC. You might need a bit more voltage during load, so I’d suggest switching to an LLC 2, which could require a slight increase in voltage for stability.
E
EHB2112
05-13-2020, 07:34 AM #2

This only occurs when OCed, right? That means you probably need to set up an LLC. You might need a bit more voltage during load, so I’d suggest switching to an LLC 2, which could require a slight increase in voltage for stability.

M
Manthing12
Junior Member
4
05-15-2020, 06:24 PM
#3
Only when OCed agrees is LLC2 considered higher than LLC3. Likely you think LLC4 would be correct, but it still does so... even at 1.3V, which is odd.
M
Manthing12
05-15-2020, 06:24 PM #3

Only when OCed agrees is LLC2 considered higher than LLC3. Likely you think LLC4 would be correct, but it still does so... even at 1.3V, which is odd.

I
iTzCheTTo
Member
80
05-16-2020, 02:09 AM
#4
The higher the value, the smaller the corrections needed. I’m accustomed to greater significance with more compensation. Consider LLC5—it addresses transient response concerns. This explanation is detailed.
I
iTzCheTTo
05-16-2020, 02:09 AM #4

The higher the value, the smaller the corrections needed. I’m accustomed to greater significance with more compensation. Consider LLC5—it addresses transient response concerns. This explanation is detailed.

S
StephanKruger
Member
226
05-16-2020, 09:29 AM
#5
Thank you for the update. The 5V from LLC works well in prime 95, and the temperature drop is a positive improvement.
S
StephanKruger
05-16-2020, 09:29 AM #5

Thank you for the update. The 5V from LLC works well in prime 95, and the temperature drop is a positive improvement.