• 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.

Help with unstable 6700k build

A

addyskylake

n00b
Joined
Sep 11, 2015
Messages
4
Hi all,

Quite puzzled after a week of testing. I only aim at 4.4Ghz OC as i don't want too much heat/noise, also don't plan on OC-ing the memory. Given various failures i have set the DDR voltage to be 1.23V and SA 1.1V ; memtest shows the memory is good (8 hours testing). CPU Voltage setting is mostly manual with LLC 4 or 5. Reported vcore are from Asus utility and/or hwinfo64 (they are consistent with each other).

Positive tests:
Memtest: 2 passes for 8 hours
Realbench 4.4 @ 1.344 Vcore - 8 hours pass
OCCT 4.4 @ 1.344 Vcore - 3 hours pass
Prime95 28.7 14 hours 4.0 @ 1.30 Vcore
Temp: Max temp 70C, load temp mostly 55-65C

Many P95 Failures:
44@1.36
42@1.3
42@1.36

Bigger problem with the P95 failures is that occasionally it would freeze the machine (other stress tests failure would only report instability). The first few times the Asus MB reported "anti-surge protection"; since then i have disabled it as i read online that the anti-surge protection might not be reliable. My PC is hooked to a sinewave UPS anyway. But after disabling the anti-surge protection the freezes continue, just without the error message.

So the main question is: if i ignore P95, my machine seems to be stable @ 4.4 @ 1.344V Vcore . But the many P95 failures got me worried -- what would you do? Should i suspect the PSU (it's a high-end one and has been stable with my previous IB system)? But i would think PSU load is actually higher in realbench than P95. I've also tried P95 27.9 instead of 28.7 but it doesn't change the situation. And given my temps i don't think P95 is overheating the cpu.


My Build:
6700k + Scythe Kotetsu
Asus z170-A
Crucial DDR4 32GB 2133
Asus GTX 970
Corsair AX750 PSU
 
you are making a lot of mistakes there, you don't start overclocking a machine undervolting the RAM as first choice. are you coming from a AMD platform or what? memory overclock affect nothing the CPU overclocking.. you are also applying too much voltage for nothing, keep using Prime95 and you're gona fry that chip..

look do this, go in BIOS and restore everything at default. once everything stay default just increase the CPU multiplier to 44-45 whatever you want and done.. enjoy and forget it. the Adaptive voltage I think it's default for the Z170-A and it will make a good job.. you don't need to mess with anything else.
 
Hm -- i am starting with a little RAM overvolting (1.23V vs 1.2V stock), going with the thought that sometimes 4 sticks require a little more voltage. Memtest pass so i am not touching that part much.

As for 6700k vcore 1.35V -ish voltage seems very normal -- that's what the MB would gives under auto-voltage (indeed it easily shoot past 1.4V under auto).

I did start with adaptive @ 1.36 -- but voltage jumps up and down so it's harder to reason failures - and with adaptive and prime95 AVX would push the voltage quite a bit higher to exceed 1.4V (the 0.1 V AVX bonus). So by going manual @ 1.36V I thought I am more conservative and shouldn't fry the chip even running prime. Temp is always below 70 and mostly in the 60 range.

Online reviews also mostly suggest 1.3X V (load vcore) at various OC frequencies (e.g. http://www.anandtech.com/show/9533/intel-i7-6700k-overclocking-4-8-ghz)
 
If you are running the memory at stock 2133 you should leave the voltage at auto, no reason to mess with it. Just set it to 2133 and that it.

Put SA back to auto

SO default everything
change to ddr4 2133

set manual vcore to 1.35 for now, multiplier to 44 or 45

if you get stable set voltage in adaptive, do not use auto vcore, at least on my board the voltages were going way to high even at stock speed.

if you cant even get stable at 4.2 then something is wonky, what version of BIOS are you using?
 
Thanks! I am using BIOS 0801 -- there's a 0901 beta BIOS available, but i heard it crash the Asus utilities so i am not using that.

Short of the memory differences (i thought a little bit extra won't hurt? will try your suggestion) i've been running tests at 44 @ 1.36V -- realbench 8 hours stable, just the prime95 problem. And yes with Prime95 i've seen failure even at 4.2 . So i am suspecting something wrong that's unrelated to OC. (Though technically 4.2 4 cores is still an OC)
 
For sure set your memory back to stock and verify your CPU overclock again.
Memory can be a pig on Skylake, there are many reports of people not able to get rated speeds.

My current DDR4 worked fine at 3000MHz on my last motherboard but is only stable at just under 2900MHz with my new one.
Both Skylake.
I had terrible problems testing with Prime95 until I set the memory back to 2100MHz and worked from there.

It might be wise to try the new mobo BIOS, my first mobo was easier to overclock once I updated it.
 
Thanks Nenu. I have 2133MHz memory and i never try to OC it (Just over-volt a little bit in the hope of gaining stability). So maybe i should try even lower it to like 2000 Mhz? (Also at 2133MHz, i am memtest 8 hours stable, if it means much)..
 
All ram is 2100MHz, manufacturers see how far it will overclock and then sell it at that speed with SPD BIOS settings which "should" be stable.
2100MHz at standard volts should be fine. Its when you try to run it at the manufacturers overclock speed that issues occur.

I doubt very much that it is the problem at 2100MHz.
 
You must log in or register to reply here.
Back
Top