• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

4770K OC Advice Needed

Astralogic

Gawd
Joined
Apr 24, 2012
Messages
581
Here is what I currently have:

Vdroop Offset (LLC) = +100
Ring Ratio = 34
Ring Voltage = 1v
VCCIN = 1.8
Vcore = 1.29

So far I have one hour of stability with OCT.

So far I've been pretty much just increasing the vcore. This CPU is crap for overclocking :(

I don't want my vcore to be at 1.29 though it's too high, I might just forget about OCing this cpu :(
 
Yeah unfortunately the 4770K are terrible overclockers. Out of the 4 we tried only one would go over 4.2. Voltage doesn't even help much and they run hot as hell. Delidding does help but it doesn't seem to make enough difference to justify it.
 
OK I've managed to get to 4.2, I'm stable at 1.23 on the vcore. Everything else is at default/auto.

Should I try to get my ring ratio up to 42? If so do I just keep increasing the ring voltage until I'm stable? Or do I need to increase the VCCIN as well?
 
you trying to overclock a haswell like a sandy, is different methods.

reset it all to default and only touch adaptive voltage thats it, leave the rest alone.
 
You mean leave all voltages on auto, set the ratio to 42, ring ratio to 42, then set the voltage modes to adaptive?

Won't that mean more volts then necessary will go into the CPU?
 
Yeah unfortunately the 4770K are terrible overclockers. Out of the 4 we tried only one would go over 4.2. Voltage doesn't even help much and they run hot as hell. Delidding does help but it doesn't seem to make enough difference to justify it.

What this guy said. I have an i7 4770K, they just run way too hot. In the end I gave up and left it at 4.1 because I still had temps going near 90's after awhile of IBT using a NH-D14. On the flipside, I built a computer for my friend and the devil's canyon seem to fare way better in general with OCing. It got to 4.5 ghz in 1 try and stayed a lot cooler compared to the haswell.
 
no, I didnt even touch the ring ratio either.

just set adaptive to something like +0.025 or +0.05 that will be a lower vcore than you used. Plus setting a vcore statically it bumps it also when not in turbo, adaptive only increases when in turbo.
 
Last edited:
you trying to overclock a haswell like a sandy, is different methods.

reset it all to default and only touch adaptive voltage thats it, leave the rest alone.

OK so I tried this and it crashed after about 2minutes. Also the vcore on full load was 1.28.

So, that method is useless. I'll go back to the normal method (didn't expect it to work to be honest).

So now my ratios are both on 42, LLC at +100, vcore on 1.23, I will attempt to find stability at the lowest voltage.
 
So I read that the max safe ring voltage is 1.2 but I'm not stable at that. Does this mean I need to increase my VCCIN to support the ring and core voltage?
 
OK so I tried this and it crashed after about 2minutes. Also the vcore on full load was 1.28.

So, that method is useless. I'll go back to the normal method (didn't expect it to work to be honest).

So now my ratios are both on 42, LLC at +100, vcore on 1.23, I will attempt to find stability at the lowest voltage.

I missed a 0 on the voltages sorry, should have been 0.025 not 0.25.

haswell has on chip voltage controller, messing around with all separate voltages is obselete now.
 
Your missing my point, reducing the voltage wont make it stable, that will only increase it's instability. I assume you are talking about LLC? (vdroop offset on my MSI board)

You're the only person I've ever come across saying this. Can someone else please confirm this is a valid OC method?
 
At 4.2 to 4.3 you don't need to worry about setting the uncore. Leave it alone for now. Do not try to match it to your cpu freq. Does not matter at these speeds.

Also, don't worry about using the offset method until you get stable. Set VCORE to 1.25. VRIN to 1.85, Ring to 1.2 - 1.25 and believe it or not, mine wouldn't act totally right until I gave the system agent voltage a little bump to .20. Even then it does weird shit like not wake from sleep right occasionally and just other weird shit. I can run stability all day. I am water cooled though. Even then, I see 80's sometimes. Gotta watch it.

I don't want to single anybody out but saying there is only one voltage to mess with on these chips is at best misinformed.

http://www.overclock.net/t/1401976/the-gigabyte-z87-haswell-overclocking-oc-guide
Don't know if you've seen this. It is for Gigabyte boards but the list of max voltages is helpful.
You can go higher than you are going.
 
Your missing my point, reducing the voltage wont make it stable, that will only increase it's instability. I assume you are talking about LLC? (vdroop offset on my MSI board)

You're the only person I've ever come across saying this. Can someone else please confirm this is a valid OC method?

this method is shown and demonstrated by the JJ guy from asus.

The original I watched was very long, here is short version on youtube.

https://www.youtube.com/watch?v=4Pilo1E6kXQ

in the short video he sets the voltage for adaptive, on the long video he uses offset (like I did). Also on the long video he explains why the older methods are no longer ideal.
 
I got the long video on my hdd as I did download it but sadly couldnt find it on google, will post it if I find it.
 
I'm going to follow this tutorial as I'm failing to achieve stability modifying things my own way: https://www.youtube.com/watch?v=7mLP3pGvnDU

looks reasonable he is using adaptive. Not sure about the extra stuff he does after tho.

Bear in mind not every cpu is equal, you may just have a bad chip.

My cpu is stable at 4.3 with 1.05v, which I have been told is a good cpu.

haswell4.3.jpg
 
Well, I give up on this CPU, it's really bad. I'll just use the auto overclock like you say. It gave me 4ghz at 1.22v...

I'll take what I can get :/
 
Sometimes the Silicon Lottery really bites. I have had K versions that would scream and others that were lame ducks. Luck of the draw.
 
Back
Top