Help me OC my PC

Erothes

Commendable
Sep 26, 2016
144
0
1,690
Want to get help and learn how to OC to be able to play games and get higher performance (basically the most common reason for OC lol) and get tips. Preferably doing it over skype and so on.
My specs:
i7 6700k
gigabyte 1060 6 gb xtreme gaming
asus z170 pro gaming
8 gb of 2133mhz ram (bought 2x8 but one was dead, planning on buying another 2x8)
EVGA SuperNova 750W G2
Additional:
WD blue 1tb (planning on buying another hdd and a ssd)
Fractal Design define r5

I have MSI afterburner and rivatuner installed, along with other things to monitor GPU and CPU.
I am using the UniGine Valley for benchmarking.
I am playing on 1080p with double monitor (one using the dedicated card and the other integrated graphics, I am using one for gaming and the other for web and other stuff).
 

lodders

Admirable
With a CPU that good, you probably won't see any benefit to gaming performance from overclocking.

But if you really want to, a good basic overclock can be done by...
Set your CPU core voltage to 1.2V, and set your CPU multiplier to 40. Checking that your computer can still play games without crashing, and that the CPU does not go above 75°C. If that works OK, set the multiplier to 41. Still OK, then repeat until it crashes or gets too hot. When it does, turn the multiplier down a bit.
Well done, you have now overclocked your CPU
 

adamscurr

Distinguished
Apr 6, 2013
232
0
18,760
With decent cooling, you 6700k should do 4.6 without going crazy... Lodders had some good advice for how to do it... Just rinse and repeat, raising the multiplier until it doesn't go any higher... You can raise the voltage at that point, but I've found that I don't like to raise it too high because to tend to get less on your return...

For example, my 6700k is running at 4.6 on 1.23v. I can run it stable at 5.0 on 1.43, but my temps are in the 70's and low 80's whereas at 4.6 I'm running a cool 50 something under load and in the mid 60's at the highest. For me, the difference between 4.6 and 5.0 did not justify the additional voltage and temps... :)

Adam
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


Where do you do that? it is annoying since in the BIOS it says I run the cpu at 1.2 to 1.3 voltage, but in the AI suite 3 it says 0.8.. and when I changed it in the AI suite 3 to something like +0.200, it says 1 but in the CPUID it still says 0.8
 

Makentox

Estimable
Sep 24, 2016
1,099
0
2,960

Well wont ask what was the reason to buy such a bad cooler which is not suitable for overclocking but i believe u can overclock your cpu to around 4.5 / 4.6 with decent temperatures. It also depends on how good is your chip and case airflow.
Download prime95 for stress test and use HWmonitor for monitoring temps.
Put manual voltage to 1.300V and LLC to level 6( u will find it in VRM settings), and put multiplier to 46. Stress test your system. If it stable go back to bios and decrease v by 0.010V, repeat it again until system become unstable. Once it happen rise voltage by 0.010v. So this voltage has passed prime95 but doesnt mean it will be stable in games. So from this point u should just play games, if u bsod rise voltage by 0.010v more. It should be stable now.
For GPU is very simple, you rise your clock speed until gpu will crash, then do the same with memory stress test using furmark make sure u see no artifacts.
From my own experience best games to test! CPU Rise of tomb raider, there no game that were more demanding than this... Witcher3, metro, crysis is nothing compared.
For GPU Rise of tomb raider had big impact on memory, and free steam game "unturned" had huge impact on GPU clock speed.


 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


How long should I stress test for?
What is LLC and why should I set it to level 6?
How do I know if it is unstable?
 

Makentox

Estimable
Sep 24, 2016
1,099
0
2,960


Ofc you should monitor your temperature when stress testing and not let it go higher than 80C
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


I am guessing to run the prime95 blend?
and why does the CPUID HWmonitor and CPU-Z say that my core voltage is around 0.8? when I set it in the BIOS to be 1.300 and in the AI suite 3 it shows around 1.3 too, is it because it is only 1.3 under load?
oh and 80+ celcius means that it is unstable?
and btw should I leave my minimum multiplier to be 8?
 

Makentox

Estimable
Sep 24, 2016
1,099
0
2,960
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


it seemed to be a glitch! the fourth core, or rather #3, it was 65 in the first one but in the next one it was the same temps, so around ~90 ish
oh and love linus :)
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


Stress test seems to be going good, multiplier: 44, core voltage: 1.250
Prime95, 2.66, small, duration: 15 min
Took screen of HWMonitor every 5 minutes.
5 minutes:
https://prnt.sc/f9vzk0
10 minutes:
https://prnt.sc/f9vzlc
15 minutes:
https://prnt.sc/f9vzmc
 

Makentox

Estimable
Sep 24, 2016
1,099
0
2,960
temperature is still high.. so weird as u can see from pic u get 1.293 vcore. Try to lower your LLC to level 5, and check if vcore in HWmonitor went down. Also try to play some game to see what temperature you get, you shouldnt go past 80C during gaming. If you do so lower your clocks and voltage... You bought horrible cooler.
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


thing is, when I tried gaming on the llc level 6 and stuff, what happend was that the temperature did not go abouve 45 celcius but the usage did not go above 65% and just jumped between and inbetween 65% and 35%. All the cores were at 800 mhz with once in a while 1 or 2 jumping to 1000 mhz for a second. And I got these annoying random lag spikes that lasted like 1/3 of a second.
I tried something with running prime95 quickly and going in game so that all cores are forced to run 100% and instantly got much more fps and the random lag spikes, while still there, were reasonably better.

Stress test seems to be going good, multiplier: 44, core voltage: 1.250, LLC: Level 5
Prime95, 2.66, small, duration: 20 min
Took screen of HWMonitor every 5 minutes.
5 minutes:
http://prntscr.com/fa1ddy
10 minutes:
https://prnt.sc/fa1dg2
15 minutes:
http://prntscr.com/fa1dj3
20 minutes:
https://prnt.sc/fa1dll
 

Makentox

Estimable
Sep 24, 2016
1,099
0
2,960


Well now temperature is much better, try 45 multi now and rise vcore 0.010 every time until stable, up to 1.280v in hwmonitor, if still unstable then put back to 44 and adjust voltage accordingly.
Weird u have lag tho, what games did u try to play? I believe that u have cs go installed, try it. Its not heavy but still CPU focused game or most other multiplayer games should work as they are more CPU games.
Also to make things simple to monitor while gaming, download Rivatuner that works with MSI afterburner. It will show all info like usage and temp on screen while gaming, no need to alt tab everytime.
Also make sure that u dont touch GPU overclock now, be finished with CPU first.
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


I was playing The elder scrolls online, I dont play cs:go
and I have rivatuner, that is how I saw that the usage was low and the temps, but the core clocks is what is bad.. they are suppose have a bigger multiplier while gaming than 8.
Oh and I have dual screen, so I dont need to alt tab, I am using igpu multimonitor feature of ASUS to run the monitor I dont play on.
while playing:
http://prntscr.com/fa2e82
https://prnt.sc/fa2e71
And i've raised the multiplier to 45 and did not change the voltage, still at 1.250, under stress tests, the voltage seems to keep a stable 1.270-1.280 occasionally going as high as 1.355. The problem is the temperatures are very high, going 84 max after only 2 minutes. I will try to up the voltage.
 

Erothes

Commendable
Sep 26, 2016
144
0
1,690


okay! I did the multiplier 45 and tried adjusting the voltage, unstable, temperatures through the roof.
Went back to the 44, 1.250 and the temperatures are going into the 80's at the beginning of the stress test!
 

TRENDING THREADS