cancel
Showing results for 
Search instead for 
Did you mean: 

When are you guys going to fix Adaptive Mode OC for Maximus X Hero?

Vitoss
Level 7
I'm forced to use manual mode, it has a precise and solid voltage, example: I set 1.310v in BIOS with LLC 5 and my idle voltage is 1.312v and full load is 1.280v, a nice vdroop as it should be, but if I used adaptive mode, my voltage fluctuates a lot, just by opening a browser tab it jumps to 1.375v from 1.312 and the load is inconstant, keep changing from 1.265v to 1.360v, I can mitigate that using an AVX offset but it is still broken and nothing like manual mode, I heard that there are other boards that are doing fine using adaptive mode.
7,900 Views
27 REPLIES 27

Raja
Level 13
Which UEFI version and how are you setting Adaptive voltage? Might want to share screenshots of the BIOS settings.

Raja@ASUS wrote:
Which UEFI version and how are you setting Adaptive voltage? Might want to share screenshots of the BIOS settings.


Using 1101, also a lot of people have this problem (I asked one time in OCnet), I just set additional turbo voltage to 1.310v just like for manual mode and leave the offset on auto, it doesn't matter if I apply negative or positive offset, the voltage fluctuates the same under load/idle. Also I tested with SVID to Best-Case Scenario and set AC/DC Loadline to 0.01 and no changes at all.

Vitoss wrote:
Using 1101, also a lot of people have this problem (I asked one time in OCnet), I just set additional turbo voltage to 1.310v just like for manual mode and leave the offset on auto, it doesn't matter if I apply negative or positive offset, the voltage fluctuates the same under load/idle. Also I tested with SVID to Best-Case Scenario and set AC/DC Loadline to 0.01 and no changes at all.


Any changes to the cache or other parameters?

I'll put the Hero on the bench tomorrow and see if I can replicate. Last time I checked, this had all been patched.

Raja@ASUS wrote:
Any changes to the cache or other parameters?

I'll put the Hero on the bench tomorrow and see if I can replicate. Last time I checked, this had all been patched.



Well, I set cache to 48x (min and max), Ring Down Bus AUTO and just changed the short and long power (something like that, it's in the same screen of AC/DC Loadline) to 4095 or something close to that, the maximum value, also set 255.50 for another thing related to cache I think, just disabling power save features and assuring I will have always the maximum performance. To test the voltage fluctuation I've disabled C-States, SpeedStep and SpeedShift and left only Turbo Boost enabled, and also set Windows power plan to 100% for minimum state while using all of this with adaptive mode, so I could monitor the minimum and maximum voltage values with HWInfo64 while browsing, idleing and full load stressing.

Raja
Level 13
What happens when you leave the cache ratio and all related settings on auto? Cache voltage and vcore is shared, so changes to the cache settings can affect the requested voltage. You shouldn't need to set all these power settings manually, either. Start at defaults (clear cmos), only set vcore and the cpu ratio (you can try per core and sync all cores), then check voltages.

Raja@ASUS wrote:
What happens when you leave the cache ratio and all related settings on auto? Cache voltage and vcore is shared, so changes to the cache settings can affect the requested voltage. You shouldn't need to set all these power settings manually, either. Start at defaults (clear cmos), only set vcore and the cpu ratio (you can try per core and sync all cores), then check voltages.


Just did a Clear CMOS and left everything AUTO, set 50x sync all cores and the same 1.310v for adaptive mode, still have fluctuations between 1.265~1.392v, stress testing and web browsing.

Vitoss
Level 7
One interesting thing I noticed, I tried with SVID Best-Case Scenario and AC/DC 0.01 and for 4.5Ghz core and 45x min and max cache with the same 1.310v voltage it was solid 1.312v and 1.296v for loading/stress, didn't noticed spikes over 1.312v, but then I tried 5.0Ghz and the same 45x for cache, now I had voltage jumps to 1.360v, 1.344v just by opening Google Chrome tabs, so seems like a certain frequency using THE SAME voltage varies when using adaptive mode.

Vitoss wrote:
One interesting thing I noticed, I tried with SVID Best-Case Scenario and AC/DC 0.01 and for 4.5Ghz core and 45x min and max cache with the same 1.310v voltage it was solid 1.312v and 1.296v for loading/stress, didn't noticed spikes over 1.312v, but then I tried 5.0Ghz and the same 45x for cache, now I had voltage jumps to 1.360v, 1.344v just by opening Google Chrome tabs, so seems like a certain frequency using THE SAME voltage varies when using adaptive mode.


yepp i noticed the same thing and buged me i needed 1,36v for 5 ghz and i would get spikes up to 1.392v .... on battlefield 1 ia/dc 0.01 loadline = 6
went back to manual vcore 1.36v with speed step enabled ( cpu drops down to 0.800 mhz just not the voltege ) - a lil power saving lol .

i wish it would not go past your set voltege in 0.016v steps like if it jumped to 1.372v that wouldnt be to bad but she jumps 1 - 3 svid stacks 0.016 v x 3 when running games
Rig # 1 - 14900Ks SP-118 | 87 MC @ 6.0 GHZ | 5.2 R | 4.7 E | DDR5 48GB @ 8,600 c36 | Strix RTX 4090 | PG27AQN 1440P 27" 360 Hz G-Sync ULMB 2

Rig # 2 - 14900K-SP-120 | 89 MC @ 5.9 GHZ | 5.2 R | 4.7 E | DDR4 32GB @ 4,300 c15 | Strix RTX 3080 | Aoc 1080P 25" 240 Hz G-Sync

bass junkie xl wrote:


i wish it would not go past your set voltege in 0.016v steps like if it jumped to 1.372v that wouldnt be to bad but she jumps 1 - 3 svid stacks 0.016 v x 3 when running games


Try LLC 4 with SVID Behaviour set to 'Best Case'.