i7 2600k overclock causing lag in DCS?!

[D4n]

I have my 2600k at standard clock. Then, in DCS, I look at fps, then use Intel Extreme Tuning Utility to bring clock up by 200 MHz (3,8 - 4,0 GHz instead of 3,4 -3,8 ) BUT then I get micro stutters/lag every few seconds (few frames just freeze every 2-3 seconds)... Is this due to I have to overclock first, THEN launch DCS game? Or I need to increase CPU voltage too in BIOS?! (Should I add 0.05 volts or even 0.1 volts to the voltage? So instead of standard 1.2 V up it to 1.25 or even 1.3 V ?! )

Edited June 6, 2016 by DanielNL

[Hadwell]

I run my 2600K at 103x45 4.6ghz and 1.275v using a corsair H80 cooler.

 

not sure if this helps

 

also 16gb 1866 ram, and a samsung 840 pro ssd.

[hannibal]

i also OC my 2600K, but i do it manually through BIOS.

pushing my 2600K @ 4.6GHz.

 

these are my settings..

Ai Overclock Tuner = XMP

BCLK = 100

Internal PLL Overvoltage = enable

Mem Freq = DDR3 1600MHz

EPU Power Saving Mode = disabled

DIGI -> CPU Load-line = extreme

CPU Voltage Freq = auto

VRM Freq = auto

VRM Spread = disable

CPU Power Phase = optimized

CPU Power Duty = extreme

CPU Current Capability = 130%

 

CPU voltage = Manual

CPU Manual Voltage 1.35

DRAM Voltage = 1.5

CPU Spread spectrum = disable

CPU Config -> CPU Ratio = 46

Execute Bit = disable

Intel Virtualization = disable

[D4n]

Well I have Intel Stock fan (I've never ever in my life installed a CPU/fan, although I'm 23 :P )

 

I'll see if my stock cooler can keep my 2600k below 80 °C at 4.0 GHz with 1.1-1.2 volts... :) ;)

[Hadwell]

lol, using a stock cooler on a K part should be illegal

 

and the punishment for breaking that law should be 100 hours of community service installing after market coolers on K parts lol....

 

why does intel even give coolers away with K cpus, since those crappy little pieces of aluminum are so inadequite for the whole idea of a K part?

Edited June 7, 2016 by Hadwell

[D4n]

So how much % gain do I get and how many years of life left if I buy good fan and OC to 4.5 GHz ? (cpu bought 2011)

[Hadwell]

4.5ghz is a stable overclock, even for the worst binned 2600ks, your cpu won't die if its properly cooled...

 

you'll probably want a whole new pc, and probably a new pc after that, before it would die....

[MegOhm_SD]

A noctua air cooler will amaze you with the ease of installation. I run a 3770k at 4.6 on one with no issues. It's too easy.

[cichlidfan]

A noctua air cooler will amaze you with the ease of installation. I run a 3770k at 4.6 on one with no issues. It's too easy.

 

I've been using them for a long time and I love them. Just installed a monster NH-D14 onto a i7-6700k this afternoon.

[dooom]

i run a coolermaster haf 930 case and have no problem OC my 2600k to 4.3 with stock fans and case.

 

no lag here.

[cichlidfan]

i run a coolermaster haf 930 case and have no problem OC my 2600k to 4.3 with stock fans and case.

 

no lag here.

 

That's pretty good.

 

I've never tried an overclock with a stock cooler. In fact, aside from machines built for other people, I don't recall ever installing a stock cooler.

[hannibal]

i forget to mention, i am using a water cooler solution, the corsair H100i V2

[OldE24]

depending on your case you should use an after market cooler if you are going to overclock the intel ones are not good.

 

personaly i use a Cheap coolermaster hyper 212. monitor your temps, and make sure you turn the IGPU off in the bios if it is on.

[agrasyuk]

lol, using a stock cooler on a K part should be illegal

:doh:

not sure if this helps

also 16gb 1866 ram, and a samsung 840 pro ssd.

how does that matter?

 

what i do find strange is that no one asked OP for results of burn-in test.

Daniel, how long did you run a stress test on your CPU to verify the overclock stability? Did it pass at all? while you can't go wrong with aftermarket cooler like 212 or noctua (budget permitting) mentioned above, low to low-med overclock is definitely possible on stock cooler. but before adjusting multiplier i would lock the voltage. if left on auto your board might be rasinig the V by itself to try and keep up with whatever numbers you dialed. so first while still at stock freq. find the voltage at which your cooler and case can keep the CPU during max heat stress test below say 70C (that would be max temp for my personal comfort level, my day to day target is below 60). and then slowly raise the multiplier to find which clock (again , under max heat stress test) the voltage can support. when you start failing the tests throttle back. this might take an evening or two. thats just in a nutshell, much better detailed guides available on the topic.

 

another but IMHO less likely reason - hitting 80C temperatures you are not very far from the thermal throttling threshold.

[Hadwell]

:doh:

 

you forgot the context in wich i said that...

 

 

I will elaborate... the reason why non-Ks have locked multipliers and Ks are unlocked... Ks are there for overclocking, to... you know... make use of that unlocked multiplier... aka stock heatsinks shouldn't be on Ks... even if at lower clocks they can still run, you're better off saving your money and getting a non-K, if you're not going to buy an aftermarket cooler and overclock...

 

is why i can't understand why intel includes such a bad heatsink with K cpus...

 

and ram and not having DCS on a decent ssd are two more reasons for screen freezes.

Edited June 7, 2016 by Hadwell

[Malefic Rage]

Does the stuttering only work on an overclock? That is odd, I'm not aware of anything that should cause that. If the CPU is unstable the whole system should become unstable with software and even the OS crashing.

 

Do you know what specs overclocking utility is actually changing? Have you manually overclocked other parts of the system such as the bus speed? If so, there may be a timings issue somewhere.

[D4n]

:doh:

 

how does that matter?

 

what i do find strange is that no one asked OP for results of burn-in test.

Daniel, how long did you run a stress test on your CPU to verify the overclock stability? Did it pass at all? while you can't go wrong with aftermarket cooler like 212 or noctua (budget permitting) mentioned above, low to low-med overclock is definitely possible on stock cooler. but before adjusting multiplier i would lock the voltage. if left on auto your board might be rasinig the V by itself to try and keep up with whatever numbers you dialed. so first while still at stock freq. find the voltage at which your cooler and case can keep the CPU during max heat stress test below say 70C (that would be max temp for my personal comfort level, my day to day target is below 60). and then slowly raise the multiplier to find which clock (again , under max heat stress test) the voltage can support. when you start failing the tests throttle back. this might take an evening or two. thats just in a nutshell, much better detailed guides available on the topic.

 

another but IMHO less likely reason - hitting 80C temperatures you are not very far from the thermal throttling threshold.

 

 

I did the CPU-Z bench test thing for 3-5 minutes, temps didn't go above 70-75 °C. (Thermal throttling didn't activate, I checked with IXTU)

 

I also tried 3.9-4.1 GHz, but then black screen, had to restart PC (not enough volts I guess)

 

 

 

I don't know how IXTU works exactly, Malefic... no idea if it changes bus speed automatically... Maybe it's just caused by the OC without increasing voltage...

 

Noctua are very expensive in Germany... I'd get another one then.

[PureEvil]

Maybe get used noctua cooler? It's not getting any worse over time.

[D4n]

Good idea! :D

 

So would you say yes to this one? Btw. is it silent? ( 17,6 dB(A), is that loud?)

 

http://www.ebay.de/itm/Noctua-NH-U9-fur-Sockel-775-1150-1155-1156-1366-32789-/162081662323?hash=item25bcd1e573:g:VE8AAOSwDV1XRGN7

 

Should I buy new thermal paste too or leave the old one? (Wouldn't part of the paste stick to the intel stock cooler if I remove it? :/ )

[PureEvil]

I was thinking about something like NH-D15. Macho is good too i heard.

 

New thermal paste, yes, always.

[Rangi]

@DanielNL , New paste, it's easy and cheap.

Edited June 7, 2016 by Rangi
Context

[agrasyuk]

Malefic, no. It is entirely possible to have unstable CPU and keep working. The reason is that not all cores, even on same chip, created equal. in our situation where we try to get as close to the tipping point we will start seeing individual cores fail before the entire system locks up. So windows will boot, things will work but when you check prime95 you will see individual cores erroring and shutting off, and system still will be still usable. For some reason my 6600K has one core that insists to run 3C degrees higher then the rest and was consistently shutting of during tests and the system kept going. Hence I'm stressing the importance of long stress tests with per core monitoring.

 

Hadwell, there is no arguing that aftermarket cooler provide much better performance and that performance is vital for good overclock. I also try and run aftermarket when I can , but stock coolers are still fully adequate for stock configs. and proven to sustain mild to mid overclock as well depending on case ventilation. So I think you overreacting there a bit with "criminal" ;) regardless, Intel took note of your opinion and the Skylake K chips ship without coolers altogether.

Situation at hand however is that OP for some reason has to use stock and even if he is going to get a better heatsink later it still going to be a good exercise getting his current setup stable.

 

 

OP, I wasn't aware CPU-z has stress generation . Maybe I'm out of date but instead of using any kind of utilities I preffer to do everything in BIOS. Perhaps we first try to eliminate that software variable? Ditch the utility, lock your volts and start dialing multiplier in bios?

5 minutes test is better then nothing, but when you reach closer to your goal clock I would at least test overnight.

 

ADD:

I would strongly advise on not reusing TIM. Given the fact both surfaces have to be squeeky clean I actually don' see how it can be done altogether. Get a small tube of something good , like the time tested AS, will go long way

Edited June 7, 2016 by agrasyuk

[Malefic Rage]

Malefic, no. It is entirely possible to have unstable CPU and keep working. The reason is that not all cores, even on same chip, created equal. in our situation where we try to get as close to the tipping point we will start seeing individual cores fail before the entire system locks up. So windows will boot, things will work but when you check prime95 you will see individual cores erroring and shutting off, and system still will be still usable. For some reason my 6600K has one core that insists to run 3C degrees higher then the rest and was consistently shutting of during tests and the system kept going. Hence I'm stressing the importance of long stress tests with per core monitoring.

 

Interesting. I didn't know that. I'm used to my old AMD CPU that just scales down every core at once. Thanks.

[D4n]

Ok now I have to decide...

 

either that USED Noctua one for 35 €

 

http://www.ebay.de/itm/Noctua-NH-U9-fur-Sockel-775-1150-1155-1156-1366-32789-/162081662323?hash=item25bcd1e573:g:VE8AAOSwDV1XRGN7

 

Or, as I thought about water cooling for even lower temps and noise, for 50 € on Amazon, NEW:

 

http://www.thermaltake.com/products-model.aspx?id=C_00002063

 

http://www.thermaltake.com/products-model_Specification.aspx?id=C_00002063

 

Does anyone of you use water cooling? Is it harder to install than air cooling/what does my PC case need?

[OldE24]

https://www.amazon.com/s/ref=nb_sb_noss_2?url=search-alias%3Daps&field-keywords=hyper+212+evo

 

this is what i use on my 2600K cheap and works well for the price. clocked at 4.4 and doesnt break 60C on a hot day.