• 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.

9600KF Unexpected Overclock

Associate
Joined
15 Feb 2020
Posts
703
Location
Q-Dance HQ
So, it would seem my 9600kf is a little odd. For a while now I was hitting 4.8ghz @ 1.2v and 4.8ghz avx @ 1.248v. I would then have to increase voltage by .05v for every 100mhz to be fully stable so 5ghz is my max before pushing over 1.35v for 5.1ghz+.

After messing around today for most of it, it turns outs cache ratio is the issue. As soon as I go above 45x cache I have to add a lot more voltage to be stable as above. So at 4.8ghz I was at 4.5ghz cache and add 100mhz to both for making me add .05v to be stable.

So i did a little test and kept cache at 4.5ghz and surprize I am stable so far at 5ghz @ 1.264v and 5ghz avx @ 1.308v. I tried cache 4.6ghz and I suddenly needed 1.296v non avx and 1.344v avx to be stable and voltage only increased for each 100mhz cache.

Strange that my 5ghz avx load dropped from 150watts to 135watts. that's a large saving in energy. so to make sure I am fully stable I fired up a few games varying in cpu usage. from most cpu intense first we have warzone, ark survival evolved, killing floor 2 and crossout. All past for 1 hour gameplay each. The gameplay is the main reason for me having to take a whole day testing but so far 5ghz stable far lower then expected
icon_biggrin.gif
5.1-5.2ghz here I come, maybe?
 
Wow buddy you are the hero of the day!

you just revolutionized my 9600KF overclock. I had the identical issue, and was always curious why my 9600KF needed slightly more voltage than a 9900K when overclocking considering the cores are basically identical. Now I know it's because I had my cache running at 4.9GHz.

Lowered that back to 4500MHz and instantly noticed the difference. I originally needed 1.373 volts to achieve a 5GHz OC, and once I changed the cache to 4500, it appears I only need about 1.238v to keep it stable at 5000MHz.

Thank you very much!
 
no idea if your being sarcastic or not (you are) but either way your welcome.
4 posts in 9 years leads me think he is being quite genuine. :)

Glad you discovered that you should leave the cache at 4.5Ghz when overclocking the core and once you've got the core stable only then bump the cache up in increments until unstable. My cache went to 4.9Ghz, without touching the voltage which was 5.3Ghz @ 1.305v.
 
Last edited:
4 posts in 9 years leads me think he is being quite genuine. :)

Glad you discovered that you should leave the cache at 4.5Ghz when overclocking the core and once you've got the core stable only then bump the cache up in increments until unstable. My cache went to 4.9Ghz, without touching the voltage which was 5.3Ghz @ 1.305v.
i miss my 9700k. it randomly died a few months back. booted pc and nothing (00 debug code). sent back to intel and they confirmed the chip refused to detect itself on the mobo so reads error code 00. shame that was a great cpu 5ghz 1.25v. so had to get a 9600kf as that is all that was in stock as intel had nothing else. i did get a refund for the remaining £140 extra as the 9700k would have cost £360. i then upgraded my watercooling.

I will have a 9700k again soon :)
 
Wow buddy you are the hero of the day!

you just revolutionized my 9600KF overclock. I had the identical issue, and was always curious why my 9600KF needed slightly more voltage than a 9900K when overclocking considering the cores are basically identical. Now I know it's because I had my cache running at 4.9GHz.

Lowered that back to 4500MHz and instantly noticed the difference. I originally needed 1.373 volts to achieve a 5GHz OC, and once I changed the cache to 4500, it appears I only need about 1.238v to keep it stable at 5000MHz.

Thank you very much!
I was being honest and actually helped another guy out with your advice as well on another forum.
 
Based on your recent post (cpuz benchmark all welcome) it would seem you lied about not only being at 5ghz @ 1.373v but also at being stable at 1.238v. Your only at 4.8ghz @ 1.354v without avx load in that thread. Try loading cinebench and watch your voltage hit 1.392v+ as that is what you'll get if your not using avx offset.
 
Based on your recent post (cpuz benchmark all welcome) it would seem you lied about not only being at 5ghz @ 1.373v but also at being stable at 1.238v. Your only at 4.8ghz @ 1.354v without avx load in that thread. Try loading cinebench and watch your voltage hit 1.392v+ as that is what you'll get if your not using avx offset.
LOL someone has black and white thinking. You shouldn't just assume people are lying. Did it ever occur to you that my demo submission was exactly that?

Chip was not dialed in at all. This thread is way old, lots have changed since then. I find it annoying that you are trying to take me to task on this BS when my system isn't even set up to bench yet.
 
See the screen shot? Look at the voltage. Are you happy now? Im telling you the truth, your advice helped me out initially. Two water pumps 10ft of tubing and 1000MHz later, I am in a much different headspace. I've had countless bios profiles since we last spoke, but i was being genuine. Please don't get it twisted. I'm just here for a bit of friendly competition, not very often I see another 9600KF around.

Capturelink.jpg
 
Based on your recent post (cpuz benchmark all welcome) it would seem you lied about not only being at 5ghz @ 1.373v but also at being stable at 1.238v. Your only at 4.8ghz @ 1.354v without avx load in that thread. Try loading cinebench and watch your voltage hit 1.392v+ as that is what you'll get if your not using avx offset.
Im well aware of what voltages I get with what types of workloads I'm shoveling towards the CPU, including what happens in CB.

1640292364653.png
 
i guess thats once way to overheat your cpu lol. what temps do you get at 5ghz running cinebench multi core for one run?
I have a very good cooling system and on cold nights I run a passive chiller that sits outside in the cold in a 5 gallon bucket of ambient temp methanol. The reason I use methanol is because it's good down to -25*F and is less viscous than antifreeze. I have a very effiecent plate heat exchanger that pulls the heat out of my conventional loop and into the 5 gallons of methanol.

As for the CB question, Im guessing my peak temps are about 155*F when running CB (without chiller). And maybe 100*F with chiller attached.
I can do some testing and get back to you. I run 0 AVX offset.

download.png
 
you get around the same temps as me, without chiller of course. sure lets see what results you get after running some tests.
 
you get around the same temps as me, without chiller of course. sure lets see what results you get after running some tests.
What model waterblock are you running?

Here is a quick result at my 24/7 OC of 5.2GHz. 5.3GHz is the limit on benching ambient without chiller.
5.8 is the absolute limit of CPU frequency, with chiller. And I can complete CB at 5500MHz, IIRC.

cb20.jpg
 
not bad. 5.2ghz and getting 77c @ 1.392v. i will upload my cinebench with hwinfo so you can see my clock speed voltage and temps sometime tomorrow.

Edit: i can only do 5ghz safely as to get 5.1 i need 1.360v (75c) and 5.2 i need 1.408v (79c) for me anything above 1.35v or 70c is beyond my safe spot :)
 
not bad. 5.2ghz and getting 77c @ 1.392v. i will upload my cinebench with hwinfo so you can see my clock speed voltage and temps sometime tomorrow.

Edit: i can only do 5ghz safely as to get 5.1 i need 1.360v (75c) and 5.2 i need 1.408v (79c) for me anything above 1.35v or 70c is beyond my safe spot :)

What's the max frequency you have hit with your chip? Have you tried for frequency only? Sounds like you might be a little timid.

80C leaves plenty of headroom. No risk of degradation. At all. Remember that's only a single core most are averaging a good bit less.

But fair warning, my 9600KF does get pretty nasty with liquid to liquid chiller attached. Danger close! :D
 
What's the max frequency you have hit with your chip? Have you tried for frequency only? Sounds like you might be a little timid.

80C leaves plenty of headroom. No risk of degradation. At all. Remember that's only a single core most are averaging a good bit less.

But fair warning, my 9600KF does get pretty nasty with liquid to liquid chiller attached. Danger close! :D
I haven't gone above 5.2ghz as i don't want to risk damaging my cpu nor will i try for the hell of it lol.
 
Back
Top Bottom