• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

I9 9900k

The - or + will depend on you CPU VID, you might have a really nice low VID chip. Mines is higher, so I subtract

When you're done you should find the total needed is less than the static you had to set for manual Vcore in bios before. And bonus it now steps down.
 
Run Cinebench at your stable static voltage, what's the VID and Vcore? e.g 1.32v & 1.28v

Stick the VID value in 'Additional Turbo CPU core voltage'. e.g 1.32v Boot it, whats the Vcore on load?

Add or probably subtract some e.g 0.04v in the 'Offset Voltage' to get to the static you had earlier.

Fine tune by adjusting offset for lower Total Vcore if you can.

So, just to check. I input voltage in Additional Turbo CPU core voltage using the + sign. Then test.

Then.. I possibly reduce the vcore with the - offset like the picture above (which shows auto)

Not sure how you increase it with + though as you already used it?
 
Run Cinebench. What's the Vcore as its running and what is the VID reported?

Remember VID is what the CPU is calling for, the Total Adaptive is what we will give it. Based on the VID and stable load voltage using the offset value. The offset will -or+ the entire VID stack.
 
Last edited:
Run Cinebench. What's the Vcore as its running and what is the VID reported?

Remember VID is what the CPU is calling for, the Total Adaptive is what we will give it. Based on the VID and stable load voltage using the offset value. The offset will -or+ the entire VID stack.

Not testing right now sorry,

Hard to get your head around.

I find the stable voltage. I read the VID from this voltage. I input it in additional turbo CPU core voltage. Then I change the offset mode sign to either + or - and input a small amount in there (like 0.04v) to get the stable voltage using adaptive settings.

That's how my mind is understanding it so far.

Any closer?
 
Experimenting stress tests:

Prime stable settings: 0 crashes for weeks.

Aida64 and realbench stable for hour or two: multiple crashes and errors playing bfv.

This is why i question realbench or aidia runs of a couple of hours as from personal experience this is not an indicator of being "gaming stable"
 
Experimenting stress tests:

Prime stable settings: 0 crashes for weeks.

Aida64 and realbench stable for hour or two: multiple crashes and errors playing bfv.

This is why i question realbench or aidia runs of a couple of hours as from personal experience this is not an indicator of being "gaming stable"

I mentioned this before. BF V multiplayer is probably one of the best stress tests out there right now for stability.

I was in the same boat as you. Could pass stuff left and right but then crashes in BF V. In my case, needed to add .01vcore on the 9900k

Keep an eye on temps also. BF V will push cpu and vrm temps up there as well.
 
Excellent work. It's all a learning experience. Only just started getting into this OC business recently. Thanks again. I'll see how it all goes.

This is my settings for Adaptive for my 5.2ghz OC, 1.264v idle, then 1.284v when load applied. :)

b3e88725-a414-4608-b82f-282b21bf445a-original.jpg

8c85c83f-9b3e-4f72-af7a-b378e053b693-original.jpg
 
Have any of you looked at the bios cpu prediction? I am wondering where to start for voltage. Mine says:

Non AVX req for 5000MHz 1.279v.
Max non AVX Stable 4960 MHz.

Not really sure what the difference between the two is? Seems like two different readings.

I am guessing at starting at 1.3v with LLC 6 like der8aur's guide states. Trying to use LLC to get close to 1.279v as possible under load.

I think I am going to set 1.3v as a limit because it gets way hot over that voltage!
 
Have any of you looked at the bios cpu prediction? I am wondering where to start for voltage. Mine says:

Non AVX req for 5000MHz 1.279v.
Max non AVX Stable 4960 MHz.

Not really sure what the difference between the two is? Seems like two different readings.

I am guessing at starting at 1.3v with LLC 6 like der8aur's guide states. Trying to use LLC to get close to 1.279v as possible under load.

I think I am going to set 1.3v as a limit because it gets way hot over that voltage!

The predictions are just that, it will also depend on your cooling as well. Its all about those minute adjustments and so on.
 
I mentioned this before. BF V multiplayer is probably one of the best stress tests out there right now for stability.

I was in the same boat as you. Could pass stuff left and right but then crashes in BF V. In my case, needed to add .01vcore on the 9900k

Keep an eye on temps also. BF V will push cpu and vrm temps up there as well.

Arma III another good one for cpu gaming stability.
 
I mentioned this before. BF V multiplayer is probably one of the best stress tests out there right now for stability.

I was in the same boat as you. Could pass stuff left and right but then crashes in BF V. In my case, needed to add .01vcore on the 9900k

Keep an eye on temps also. BF V will push cpu and vrm temps up there as well.

Yes i saw this. My temps are higher in bfv than aida64.
I think bfv uses avx? Really hits the cpu when loading between maps too. Luckily my temps are good. I just like to experiment.
I was thinking of dropping prime from my test suite. By using aida and realbench as my guides, i was hoping to have a stable gaming system. But it didnt take long until my pc crashed in game.
 
I got my 9900K a week ago from Amazon US and thought to share my OC results here. Seems like a got a pretty good chip.

- Lowest voltage (after vdroop) pass on cinebench.
5GHz all cores 1.133v
BRd6zNU.jpg

5.3GHz all cores 1.306v
kENw9lT.jpg

- 5.2GHz/4.7GHz cache at 1.296v under load prime95 26.6 small ffts for one hour with no whea errors.
IeBDRET.jpg

- 5.3GHz OC. It can pass one hour of prime95 custom 1344k and AIDA64 (CPU, cache) at 1.363v 5.3GHz/4.7GHz cache with no whea errors
yS6eett.jpg
This chip should be 6.9 R15 ln2 given scaling. Its similar to the one I used for Intel launch demo I did in Manchester on Tuesday. Loop 6850-6900 cine all night with 64-65 cache.
 
would have loved too mate but no time this side of Xmas but thank you very much for the offer, maybe sort something in the new year if your up for it? And besides it may give me chance to try on DICE at the very least as it's been nearly 2 years!
Yeah come in. I will have ln2. We can do some records for you to submit if you wish. Maybe Dagmar will also be up for benching. I will ask him.
 
Not exactly. He is in contact with me and we are doing afew things together now. Not really benching as such. I was going to ask him to join for fun. I am not benching either as such but now amd then it's fun.
 
Not exactly. He is in contact with me and we are doing afew things together now. Not really benching as such. I was going to ask him to join for fun. I am not benching either as such but now amd then it's fun.

Just sent him a message via "whatsapp"...........when you mentioned him i realised we hadn't seen each other since Cambridge a few years ago.............either you or Dag lol.
 
Back
Top Bottom