[SOLVED] i9-10900K first time OC 51x 1.38V High LLC or 1.36V but Turbo LLC!? and what for 52x?>>>2. Bios Screenshots questions too - any suggestions pls?

May 1, 2021
3
0
10
Hello Local Pro!
My present situation - very first time OC. Would you look at my BIOS settings screenshot and correct it please, what to do with CPU OC and in the end about DDR4 Crucial Ballistics opinions will be very welcomed and appreciated:

Didn't find how to attach screenshots here - My Bios album screenshots https://photos.app.goo.gl/hoQXznyRTvv1FgxQ8

a) clear some statements like: As Bulldozer said, important is only Voltage under loading!??
For example to set Fixed voltage as 1.4V for 51-52x and then with normal (High) LLC under loading it will drop till 1,28-1,3V and will be still ok for CPU and long life?
or better use 1.35V with Turbo LLC? Or it all depends on drops off amplitude up and down - turbo with 1.36V is ok but not with 1.4V?


My config now :
mb Gigabyte Vision G Z490
i9-10900K at 51x (48x cache) stable.
AIO NZXT 360mm


Main open questions:
a) CPU - I can't make it work at any voltages close to 1,25-1,3 as many reviewers-overclockers have. May they mean it after stress test drops by VR OUT?

For now I have setup: Fixed Vcore set = 1.380V (VR OUT in HwInfo shows 1.378V) LLС-level High or 1.35 with Turbo LLC !?
In CPUID = 1.380V and with Drop-off under loading in Cinebench20 - VR VOUT = going down till 1.289V and still working ok!
Same time if I set the same in Bios - can't load windows with 1.28-1.3V value in BIOS.

With stress tests- in Prime95 in SmallFFTs with Disable AVX2 was stable even with 1.35-1.36V only with TURBO LLC, what is it no good - because of Drops and returns amplituda too high for Turbo LLC?
Getting temperature in Cinebench20 with this Voltage till 94C :-( I have good enough working 9 Fan=6 in, 3 out middletower.)
Because of 1.380V?


Is it more proper check CPU with PRime95 SmallFFts NO AVX2?

Tried btw for a short period - 1.4V for 52x could be not enough to load windows and run stress-tests.
=====================================================================================================================
========================================================================================================================

b) I Have no idea how to set up a recommendation by everybody "Override Voltage" at Gigabyte (Adaptive too).
How am I supposed to use it after FIXED CORE for example 1.380V at 51x and High LLC?


c) So what are my options for 51-52x with fixed Voltage code, Override(Adaptive) Voltage.

PS. BTW Can't find SVID to switch it OFF as in ASUS mb instructions. Does it exist in Gigabyte mb?
PS2. CFG Lock OFF is ok?
========================================================================================================================

PS. About OC DDR4 Crucial Ballistix:

DDR-4 2x16Gb Micron b-die Crucial Ballistix Game 3200 16-18-18-38 Standard XMPm now OC -> 3500 15-18-18-36 (trfc 560) with 1.39V.
TCMD = 1T and Trefi auto. (or 3600 18-18-18-38 but it is slower?).

Why slow memory test or is it a Gigabyte mb typical issue? After updates for some betas speeds fell down - from 40 to 30 Gb/s.


I read many reviews about good DDR4 OC but this is the max without errors in TestMem extreme profiles etc. VCCIO + VCCSA set till 1.3V nothing makes it stable.
And still in AIDa64 memory bandwidth are lower than in Cache-memory test around as 41-4-38 Gb/s and latency 63.3nc

Thank you for any answers and sorry so many questions to clear out.
 
Solution
The reason you didn't figure out how to attach screenshots here, is because you can't. You CAN however, link to them "in thread", as follows.


As for the rest, I'll have to look through your post further to see if I have anything to offer.
The reason you didn't figure out how to attach screenshots here, is because you can't. You CAN however, link to them "in thread", as follows.


As for the rest, I'll have to look through your post further to see if I have anything to offer.
 
  • Like
Reactions: Webbmaster
Solution
May 1, 2021
3
0
10
The reason you didn't figure out how to attach screenshots here, is because you can't. You CAN however, link to them "in thread", as follows.


As for the rest, I'll have to look through your post further to see if I have anything to offer.
thank you! Looking forward.
 
So, what are you actually trying to achieve, SPECIFICALLY?

Because, you should know that for MOST CPU samples and users, there just IS NOT any headroom for overclocking on modern CPUs, to a degree that makes it worthwhile. And the extra costs associated with the necessary hardware and cooling can't be justified by the gains you'll generally see.

Delidding might help some, but the truth is that almost all of these modern CPUs are already pushing the boundaries of what they can support, in order to show sufficient gen vs gen improvement to entice users to buy into the product at all. Aside from some minor gains from increasing all core boost, on applications or games that are well optimized for multithreaded performance, there really isn't much of anywhere to go on latest gen Intel OR AMD processors. Especially if it's a flagship model.
 
May 1, 2021
3
0
10
Overall - trying to make as less as possible Voltage in Fixed mode (cant figure out yet about Override-Adaptive modes :-( ), same time with less LLC Level,
with High LLC is barely enough for 1.38 for stress tests (just for example) , Prime95 still freeze PC.
with Turbo LLC is was ok 1.36 and Prime59 not Freezes ! Looks easy with TURBO BUT

Drops and highs could be worse with TURBO LLC - that's I am trying to AVOID or make it less this part of OC negative side effect

So I desire to make this Voltage numbers less ...
 
Really the only major problem with using a Turbo or very high LLC is that you end up with substantially higher thermal response (Temperature).

Personally, when using a fixed mode overclock I like to use a middle ground LLC, slightly less than high OR at the most, high, but not Very high, Turbo or "7, 8, 9 or 10" (For systems that use a 1-10 numerical scale for LLC settings. I like to stay around 3 for systems that have a 1-5 numerical scale for LLC) and use a slightly higher vcore. Split the difference in other words. A slightly higher vcore with a middling LLC tends to raise idle temperatures a little, which is NOT a problem so long as your peak temperatures are still within the recommended range, but it also tends to REDUCE the maximum thermal response by not having an extremely high LLC.

That also helps to minimize the drops and highs. And the bottom line is, if you are already getting instability where you're at, then REDUCING voltage is not going to happen. Using LLC does not reduce voltage, it increases it during load conditions. If you can't get a CPU, any CPU, stable, with a tolerable amount of voltage at a given frequency, then it's likely the overclock is simply too big for your specific CPU sample. Just because somebody else could do it with the same CPU does NOT mean you can. Every CPU is completely different. Every motherboard is completely different. Every graphics card is completely different. They are of course MOSTLY the same, if they are the same model, but they will all have individual characteristics that make for mild to significant differences between them which is why you can't just plug in numbers based on the model. You have to do the work to determine what is best for YOUR particular sample.

I have three links for you. They should ALL be very helpful if you can read AND understand what is being offered in each of them.





https://www.tweaktown.com/guides/9225/intel-core-i9-9900k-kf-overclocking-guide/index.html


Granted, the last one is for Coffee lake, but there are very few changes between Coffee lake and Comet lake in terms of overclocking.