i5 6600K overclocking advise to 4.6Ghz

[Crofighter88]

Hi all

i just upgraded my pc and need some info.
i've been reading many threads, But never found the answer i was looking for

i have :
Intel i5 6600k Skylake
Gigabyte z170x GAMING 3
16gb DDR4 @ 2666mhz
cooler : H100i (with sp120 Quiet edition fans)
GTX 980 Strix (overclocked it)
250Gb SSD + 2x1TB HDD (no raid)
coolermast 850w power supply

the stock cpu is 3.5Ghz , boost 3.9Ghz

So this is what i did to overclock it :

i adjusted the multiplier to 44 and and it is stable without adjusting the voltages
i used prime95 for couple hours, and had no issues

now i wanted to try to go to 4.6 /4.7 if possible

when i tried 4.6 , i got a blue screen
so i adjusted the vcore to 1.350 . got in windows and tested it with Prime95
i stopped prime95 after half-hour because one core stopped working.. (rest was still testing)

i"ve read that u need to increase the vcore to go further
so i raised the vcore to 1.4v
when running prime95 , one core still stopped working

Im kinda afraid to push it over 1.4V since i want to use my CPU daily


So what do i need to do get it stable?

When i put the multiplier to 45 with 1.4V , i have a stable overclock

The max temperture i had on my CPU is 63 C (145F)


grtz

Cro

 

[philipew]

after alot of reading i decided to stay on 4.4Ghz
the only thing i had to do is put the mulitplier to 44
the max CPU Vcore tested is 1.249 .. so thats pretty good.
had my cpu tested with prime95 overnight, had no crashes.
so i have a pretty stable OC
thx for advice all

the funny thing is , intel says that skylake is built to stand 1.52v ???
altough i never read one article that someone says they used this..

Well... the really funny thing is that Intel produce CPU chips which they want to sell. The absolute theoretical maximum voltage is in fact 1.60 V but they take off a 5% 'safety' margin and arrive at 1.52 V. In reality, around 1.40 V is probably the practical maximum. The limiting factor is mostly the overall package temperature because the current increases with the voltage and therefore the resulting power... and heat (temp).

If you had something like a cooling loop passing through a bath of liquid nitrogen, 1.60 V may just be fine. But with conventional "retail" cooling, either air or water, 1.45 V is likely to cause the chip to become too hot already over time (particularly when the ambient temp. goes from 22 C to 27 C for example). In that case it will eventually shut down to avoid damaging internal components. I had this happening at 4.7 GHz after two hours on a Prime95 heat test when it reached 85 C at 23 C ambient.

That's why I remained at 4.6 GHz and optimized my air cooling, short of replacing my single fan budget CPU cooler with a more effective one at double the price like a good ($) Noctua unit for example. Here is a picture of what I did in my rig: http://imageshack.com/a/img923/3196/PjaqTS.png

 

[jrizzz]

1.4v is encroaching on the dangerous end of voltages, there is nothing you can do to get it stable at 4.6 without raising voltages to dangerously high levels

Not necessarily true, As long as max temps allow you to run a higher voltage, you are fine. Also, increasing power levels and playing around with LLC may help you find better stability under load.

oc.net also has a very helpful and informative guide to help you in your overclocking endeavors: http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

 

[Joe_180]

I must have won the cpu lottery, mines clocked at 4.9Ghz @ 1.28v. Enough for me!

 

[philipew]

I must have won the cpu lottery, mines clocked at 4.9Ghz @ 1.28v. Enough for me!

Are you absolutely sure that you are looking at (quoting) the CPU VCore voltage though. There are a lot of voltages in there. Perhaps you could post a picture of what CPUID HWMonitor shows (it is a free app.) the top, under "Voltages", first line: CPU VCORE (Value, Min, Max). I just don't believe this unless you are using liquid nitrogen in a lab but even then, it would have to be more than 1.28 V and would only be stable for a few minutes under load. It's like saying I won the car lottery, mine clocked at 490 Miles/H.... Wouldn't you have doubts? I definitely do.

 

[Themastererr]

My friend used that ASUS one click app and got 4.6Ghz with it on his 6700k. Lol...It seems stable enough.

 

[Joe_180]

My friend used that ASUS one click app and got 4.6Ghz with it on his 6700k. Lol...It seems stable enough.

4.6 is plenty, I was just pushing to see how far it would go, I've scaled mine back to 4.7 Ghz for peace of mind.

 

[philipew]

1.4v is encroaching on the dangerous end of voltages, there is nothing you can do to get it stable at 4.6 without raising voltages to dangerously high levels

Not necessarily true, As long as max temps allow you to run a higher voltage, you are fine. Also, increasing power levels and playing around with LLC may help you find better stability under load.

oc.net also has a very helpful and informative guide to help you in your overclocking endeavors: http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

I agree, in the end, it is the temp that limits your frequency (GHz) NOT the CPU voltage. Your rig will shut down automatically when the temp goes over about 85 C (if you don't change the default max temp) to protect the CPU, NOT when your voltage (CPU VCore) goes over 1.60 V, or even over 1.90 V..

 

[L0stChild]

I saw 1.35v.. freaked me out.
u shouldnt need that much. but i am surprised on ur temps under load with 54c.. thats low for 1.355
-just to give reference on what i have on my rig: 4.6ghz with 1.272v. (also a non-k)

wow there's a lot to read but at first glace:
what speed did u start with? or did you jump to 4.6 from the getgo?
have u tried overclocking just the cpu only and left the memory settings on auto/xmp? ...

i just looked at the screen shots from ur hwmonitor... i swear that looks like 1.7volts on ur vcore..

btw 850watts... thats yeah thats not needed lol

 

[philipew]

I saw 1.35v.. freaked me out.
u shouldnt need that much. but i am surprised on ur temps under load with 54c.. thats low for 1.355
-just to give reference on what i have on my rig: 4.6ghz with 1.272v. (also a non-k)

wow there's a lot to read but at first glace:
what speed did u start with? or did you jump to 4.6 from the getgo?
have u tried overclocking just the cpu only and left the memory settings on auto/xmp? ...

i just looked at the screen shots from ur hwmonitor... i swear that looks like 1.7volts on ur vcore..

btw 850watts... thats yeah thats not needed lol

I suppose that you refer to my i5 6600K Intel CPU VCore set manually to 1.355 V in the BIOS. It is really not much at all. You can't have "4.6ghz with 1.272v. (also a non-k)" on your rig - it doesn't make any sense even though your rig seems to be an (irrelevant) 6400:
1) A "non-k" i5 6600 Intel CPU (assuming this is what we are writing about going by the thread title) runs at 3.3 GHz (https://www.ple.com.au/Products/620470/Intel-Core-i5-6600-Skylake-33GHz-6MB-Retail-Box). It is 2.7 GHz for an i5 6400. That's precisely why we buy a more expensive "K" version, to go over the stock-standard 3.3 GHz via overclocking... duh.

2) "i swear that looks like 1.7volts on ur vcore" seems to indicate that you are not looking at the CPU VCore voltage when you mention 1.272 V. My CPUID HWMonitor shows (I am still looking at it while writing this) under "Voltages" (first line of figures at the very top) CPU VCORE 1.356 V (Value) - 1.344 V (Min) - 1.356 V (Max) - in fact, "1.7volts" as you write would probably blow up the chip (even in liquid nitrogen) as Intel stipulates an absolute maximum of 1.52 V - ...just saying

3) No I didn't jump to 4.6 GHz straight away but went from 3.5 to 4.2 GHz, then progressively to 4.3, 4.4, 4.5, 4.6 and 4.7 GHz before coming back to the stable 4.6 GHz - because it also has to be stable after hours of running under Prime95 with no error on any of the four cores

4) Overclocking the memory and the CPU are two totally different and separate things - one has no influence on the other but I did the CPU before the memory (which didn't change anything for the CPU as expected) - I also bought that memory specifically to overclock it to the max (from 2133 MHz to 3200 MHz) with an Command Rate (CR) of 1T and very good settings (15 17 17 28) under 1.30 V (I run it under 1.34 V now to ensure even more stability over time)

5) My Power Supply is a Corsair RM650x 650W 80PLUS Gold Modular - I don't know where you plucked "850watts" from

6) It takes roughly 50 mV (0.050 V) per 100 MGhz for a 6600K, so working backward, to go from 4.6 GHz to 4.5 GHz you drop from your 1.272 V to 1.222 V, to 1.172 V for 4.4 GHz, to 1.122 V for 4.3 GHz, to 1.072 V for 4.2 GHz, to 1.022 V for 4.1 GHz, to 0.972 V for 4.0 GHz, to 0.922 V for 3.9 GHz, to 0.872 V for 3.8 GHz, to 0.822 V for 3.7 GHz, to 0.772 V for 3.6 GHz, to 0.722 V for 3.5 GHz, to 0.672 V for 3.4 GHz, and finally to 0.622 V for 3.3 GHz the stock-standard frequency of a "non-k" i5 6600 Intel CPU. If it were a car, it would truly be running on an "oil rug's fumes" - an absolute feat of technology, way beyond simply defying the universal (as "of the universe") laws of physics - well provided you are not from some parallel universe of cause - Mulder and Scully anybody?

As for "just to give reference", be sure to refer precisely to CPU VCore voltage when you mention "voltage". At 4.6 GHz, my CPU VTT voltage is only 1.152 V, and my VID voltage is only 1.012 V. When I type 1.355 V for CPU VCore in the BIOS, that's what I get and it's relatively low, so the temp is bound to be low because it comes from heat dissipated as power (W) and Power W = Voltage V x Current A - I think you just got "L0st" somewhere... (!)

 

[L0stChild]

i got zero clue whats going on.
so i cant have 4.6ghz with 1.272v.
ok i must be dreaming

http://valid.x86.fr/cache/screenshot/qw5v7x.png
http://www.userbenchmark.com/UserRun/2031292
would a i5 6400 have those results.
i bet i run faster than your 6600k


yeah and i got lost in that wall of text of whatever u wrote or pasted
was commenting on the OP and realized its from feb.

https://postimg.cc/image/gi9oetv0p/

 

[philipew]

i got zero clue whats going on.
so i cant have 4.6ghz with 1.272v.
ok i must be dreaming

http://valid.x86.fr/cache/screenshot/qw5v7x.png
http://www.userbenchmark.com/UserRun/2031292
would a i5 6400 have those results.
i bet i run faster than your 6600k


yeah and i got lost in that wall of text of whatever u wrote or pasted
was commenting on the OP and realized its from feb.

https://postimg.cc/image/gi9oetv0p/

... Completely L0st, I know... ;-)

 

[L0stChild]

i got zero clue whats going on.
so i cant have 4.6ghz with 1.272v.
ok i must be dreaming

http://valid.x86.fr/cache/screenshot/qw5v7x.png
http://www.userbenchmark.com/UserRun/2031292
would a i5 6400 have those results.
i bet i run faster than your 6600k


yeah and i got lost in that wall of text of whatever u wrote or pasted
was commenting on the OP and realized its from feb.

https://postimg.cc/image/gi9oetv0p/

... Completely L0st, I know... ;-)

so thats it? no more attacks?
am i wrong?

 

[grimsin]

Don't go any higher than 1.3v!

you nothing of skylake, as long as your temps atay below 80c, many overclockers go to 1.4-.1.45v

 

[Tasos_3]

Hey philipew first of all i would like to thank you for ur 6600k overclock guide in this post that helped me a lot for my overclocking !!! I would like to ask you why u disable C1E, C6/C7,C8.I dont know what are they for and why u disable them!!I am on 48 core ratio,47 cache in 1.4v in bios and LLC set to high with a Noctua nhd15 for cooling! I pass 3.5 hours in IBT with 85 max core temp but i fail constantly in prime95 28.7, two of my cores crash after 30 or 40 minutes !!! What is your opinion and what do u suggest?