07-27-2018 02:57 AM - last edited on 03-06-2024 07:57 PM by ROGBot
07-27-2018 07:15 AM
07-27-2018 07:39 AM
I didn't observe spikes above the max voltage assigned by Adaptive mode.
07-27-2018 11:39 AM
07-27-2018 12:08 PM
feedmeink wrote:
I had similar issues but with a different chip and board. When in adaptive it ignores the voltage limit you set. To resolve this i had to offset each core individually using negative increments. Basicly I would run a stress test without avx functions and watch my voltages in hwinfo64. If one core was receiving more voltage spike than another at full load, I would either clock down the core or adjust negative offset. For some reason I was having stability issues offsetting more than - .09. If that was the case I would reduce my core by 1 and offset at a lower number. It is a pain to do this, but all my cores run at less voltage spread evenly across all cores. No voltage spikes. My motherboard allows me to set a max turbo voltage when in adaptive mode, so I set that to my goal. I have a hunch your per core voltage is all over the place. Hope this helps. It would even out per core temps as well.
07-27-2018 12:27 PM
dmt.taran wrote:
metadist
Assigned by You as turbo voltage, or you mean svid?
07-27-2018 12:41 PM
metadist wrote:
Assign by SVID.
Actually, looking over my setup, I also see a strange 0.05V constant, but in a different way.
So, I have:
SVID best case,
CPU core/cache voltage -Adaptive mode,
Additional turbo mode CPU core voltage 1.35
offset - Auto (0)
LLC - level 4 (or more)
Then I see that when all cores are running at my highest ratio (51), then core voltage is 1.35+0.05 = 1.40.
No spikes, just constant overshoot. I can reduce it if I set the Offset in Adaptive mode. But I too have to wonder - where does this 0.5 come from?
Best regards.
Artem
07-27-2018 01:01 PM
07-27-2018 01:45 PM