I am trying to overclock my i5 4690k at a frequency of 4.2ghz

File: 1478293063453.jpg (86KB, 858x638px) Image search: [Google]

86KB, 858x638px

>>58290246
What about the motherboard?

>>58290246
evga

File: speccy.png (40KB, 661x526px) Image search: [Google]

40KB, 661x526px

>>58290326
here is my speccy

I have a CX600M and I got my 4690k to 4.6Ghz and my GTX 970 to 1531Mhz core and my build is almost two years old.

>>58290409
what do you think my problem is then?
Nice overclock btw I am eternally jealous

File: self intoxicated, twisted detonator.png (314KB, 1131x1006px) Image search: [Google]

314KB, 1131x1006px

Have you run stress testing programs such as Intel Burn Test or Prime95? I bet you good money 1.14v isn't enough and you don't have an aggressive LLC to pick up the slack.

Vidya is not going to push a cpu hard enough to stress test on.

Pic only semi-related. Its an AMD cpu asbout to cause mobo vrms to explode.

dont overclock retard
you are ruining your cpu for pretty much nothing

>>58290446
I have done prime95 and it held up there, but when playing games my pc restarts at random. on IntelBurnTest should I do standard or maximum?

File: Skylake 4.2.png (168KB, 603x602px)

168KB, 603x602px

>>58290246
Nothing to do with your PSU if your system runs stable at stock clocks.

Increase the voltage in the smallest possible steps until your CPU is stable at 4.2 GHz.

Pic somewhat related: My 6600k at 4.2 GHz.

>>58290471
OP here, I was also wondering if overclocking was even worth it...is it really worth it? (I am asking everyone in the thread)

>>58290446
Yea up the voltage. I got pretty lucky with mine and didn't have to raise it much.

>>58290471
I have a couple 2500k and 4690k systems that are still running and were overclocked since day one. Extreme overclocks pushing 1.4v or higher would be a concern though.

>>58290475

How long did you run prime95 for?

>>58290482
Damn, 1.251 VCORE, isn't that quite a bit for just 4.2 ghz?

>>58290493
what I mean by this is will I notice substantial gaming performance improvements with games such as Arma?

>>58290432
You fucked up somewhere, gonna restart my PC and take a look at my settings and report back.

>>58290515
like 15 minutes..I didn't notice any fluctuations in temps so I assumed it was good. That's probably retarded, right?

>>58290493
Depends on what you do with it. One of my old 2500k systems runs Plex Media Server and acts as a NAS/Cloud server for mobile devices in the home. Clock speed matters with transcoding etc. Gaming wise if you need tip top performance clock speed is still better than more cores.

>>58290529
Ok thanks

>>58290493
You want to overclock just to play games right? So, think of it like this, is 5 fps more really worth all this trouble and the risk of frying my cpu?

>>58290517
It's 1.251 because the CPU is running at 100% load encoding a video in Handbrake. Temps don't even exceed 50° so it's no problem. While gaming it's more like 1.15 or less anyways.

Don't be afraid of a bit more VCORE as long as your CPU keeps cool.

>>58290544

You need to run it for far, far longer than that to ensure stability. In theory to ensure 100% rock solid will never fail overclocks prime95 or IBT should be run for 24hrs constantly. In practical terms you can get away with less time (an hour or so is enough for vidya uses) but never say never - there is still a chance its unstable.

Point is if your chip survives an hour prime95 you can be reasonably confident its stable.

>>58290517
>Damn, 1.251 VCORE, isn't that quite a bit for just 4.2 ghz?
Vcore under 1.4v is fine. 1.2v range is fine. 1.2v is fine too but I wouldn't go beyond 1.4. 1.2-1.3 is a safe range IMO. I've not gone over 1.35 on any of my intel K cpu's and they all overclocked to 4.2 or 4.6 stable.

>>58290582
It depends on what games you play and the settings. If you are GPU- or VRAM-limited it won't matter whether your CPU runs at 3.5 or 5 GHz.

Ok for my 4.6oc my settings are the next, sync all cores, 46 core ratio, min cpu cache ratio 8 max cpu cache ratio 42, cpu loadline extreme, adaptive voltage, cpu core voltage offset 0.007, additional turbo voltage 1.261, cpu cache voltage adaptive, offset 0.005, turbo cache voltage 1.209, cpu input voltage vccin 1.930 everything else auto default settings

OP here, why are my temps so awful compared to your guys'? I have a hyper 212 evo and thought I applied the thermal paste appropriately, but I get idle temps of 30-40C and max temps of high 70s

and this is just on stock speeds

>>58290626
If he has an AMD card it helps more than a Nvidia that's for sure.

File: clip+(2016-06-25+at+05.30.43).jpg (25KB, 427x407px) Image search: [Google]

25KB, 427x407px

Set your ram to stock speeds

Overclock the cpu alone and just do cpu voltages. Don't touch cache.

You should be able to do 4.4ghz on this chip at 1.45v, 4.5ghz if lucky.

4.2Ghz should take 1.3v to be 100% stable.

Raise your voltage faggot.

>>58290641
Can I match these settings on my MSI motherboard?

>>58290646
I have nvidia, my speccy is at the top of the thread

File: fisheyeshit.jpg (4MB, 3840x2160px) Image search: [Google]

4MB, 3840x2160px

>>58290642
You probably need to delid your CPU too. It's not hard with a vice grip.

A 212 isn't the greatest though, you should look into a 120MM radiator AIO or some shit. Bigger the better.

>>58290642
How is the ventilation in your case? Fans blowing in the right direction?

>>58290666
>I have nvidia, my speccy is at the top of the thread
Yea I saw that. I was commenting that it matters if you have AMD card typically. You'll see more gains if you overclock both in my experience.

>he fell for the Corsair CX meme

>>58290651
My RAM is at stock speeds, I only touch the CPU voltage

I can try again, but I'm starting to think that overclocking isn't worth it because of the high risk and my shitty temps

>>58290685
Yeah ventilation is great imo, 2 intakes (one on front, one on bottom), 3 exhause (1 on the cpu blowing upwards, one on the top, one on the back)

>>58290699
oh ok thanks, do you think it is worth the risk to overclock for the increase in performance?

>>58290728
>oh ok thanks, do you think it is worth the risk to overclock for the increase in performance?
The whole point of buying an Intel K CPU is to overclock it with ease. Seems pointless to buy one otherwise. I own multiple 2500k and 4690k cpu's and all have been overclocked for years with no troubles. Granted I didn't push the voltage up much to achieve stable clocks. It's worth it if you use it which in my case I do.

OP here, I'm gonna wipe my bios and start with an overlock. What test frequency and voltage do you suppose I start with?

>>58290675

>Delidding

That still a dangerous game for retards and even after you physically delid you still need to 1) not be a retard 2) know what you are doing and 3) put it back together.

I'd rather sand the heatspreader down first.

>>58290666
Most likely yes, just set everything to default settings, increase ALL cores to 42 ratio, then increase the voltage using static voltage to 1.250, change the loadline to Extreme or equivalent setting and increase the vvcin to 1.9.

>>58290800
oh and disable all intel power saving features like C states and inten speedstep and set window power management mode to high performance.

>>58290728
>2 intakes (one on front, one on bottom), 3 exhause (1 on the cpu blowing upwards, one on the top, one on the back)

Sounds like a clusterfuck but it'll do the trick.

>>58290790
>>58290642
Fix your cooling first.
At 70° at stock you shouldn't even think about overclocking unless you want to fry eggs on your CPU.

>>58290800
increase all cores instead of just VCORE?

>>58290838
sync all cores and yes vcore is the value you want to increase, leave cpu cache to default, sometimes cpu cache is called vring etc

>>58290832
fuck okay, what budget cooler do you think will work best for my situation?

>>58290860
Your Hyper 212 should be fine if installed correctly.

>>58290860

Chances are you jsut need to clean your shit and turn up case fan speed before spending any money. 70c at stock clocks on an i5 using a hyper evo tells me the cooler isn't being fed enough air to decently push through the heatsink and/or not enough exhaust.

>>58290860
cryorig h7

>>58290651
im using 4690k and im running 4.4GHz at 1.275v with 212 cooler
cpu never goes above 80

>>58290918
>>58290912
this. clean and reseat the heatsink and use good paste.

>>58290246
Stress test you stupid fuck. You're probably dropping cores every which way.

>mfw idling at 32C with a NH-D14 and just one 120mm fan leaning on one side (no stock fans)

how the fuck does this nigga hit 70 idle with a 212?

>>58290246
It most likely needs higher voltages. My 3570k can hit only 4.2GHz on stock voltages too, anything above that requires absurd voltage increments (something like 1.35V to get stable 4.4GHz).

>>58291015
No no no no not idle of 70s, max temp under 100% load of 77-83C

File: idle temps.png (120KB, 1075x799px) Image search: [Google]

120KB, 1075x799px

OP here, here are my idle temps at a freq of 4.2ghz @ 1.25 VCORE....what do you think?

File: cf22a35b38f15eec030225b2a82d98c1_original.png (358KB, 680x461px) Image search: [Google]

358KB, 680x461px

>>58290675
>>58290794
I just bought me one of these badboys to simplify the process.

Anyone else use coollaboratory liquid pro on their cpu? Just got my shipment today and plan on using on my thinkpad then delid my cpu to improve the temps.

>>58291215
Looks good, I have my 4690k at 4.2 with similar voltage and temps.

>>58291215
I just noticed that the Core Voltage in the CPU-Z pic is not 1.25v, why is it not?

>>58291244

What worries me about delidding is not securing the heatsink back down (I assume you'd use some new adhesive) and thus it sliding as you secure the cooler back down.

File: prime95 temps.png (2KB, 85x95px) Image search: [Google]

2KB, 85x95px

I ran prime95 for 5 seconds and immediately got these temps.....seems terrible

>>58291252
you forgot to disable cpu intel speed step or cstates

>>58291286
is that a big deal? will not disabling it make my overclock unstable?

>>58291281
yes thats shit tier. see >>58290961

>>58291279
I figure I'll just use some super glue since I also bought the separate attachment to hold it in place.

>>58291279
>>58291318
the whole point of delidding is not to use the fucking top again, so you have direct contact between the cores and the heatsink, you mongoloids

>>58291335

Are you fucking high?

>>58291335
Couldn't you shatter the cpu that way? Also I plan on using liquid pro and my cpu cooler is the hyper 212 which has some aluminum on the bottom.

>>58291303
>>58290961
I used Artic Silver and thought i seated it properly...ugh putting thermal paste on is the most nerve-wracking part of building :0(

>>58291301
yes

>>58291384
yes you can, you have to use standoffs

>>58291394
putting on thermal paste is the easiest part you fucking retard

>>58291413
What's the best thermal paste application method?

EVGA, and more specifically, EVGA supernova g3 series.

>>58291511
https://www.youtube.com/results?search_query=thermal+paste+application+methods

>>58291519
cool story bro, but i think some context would help get your point across

>>58291511
Just use the pea method.

https://www.youtube.com/watch?v=-hNgFNH7zhQ&t=132s

Realistically, how much will the temps improve after i reseat the cooler? Would it just be more worth it to get a better cooler?

>>58291593
1. reseat the fucking cooler with new thermal paste (and use the correct amount)
2. set the fucking fan control shit in BIOS/UEFI to performance.
3. check temps during load

THEN you can start to wonder about getting a new cooler

>>58291646
Ughhh i just want someone to reseat it for me, i feel like I'm gonna fuck this up again

also i hope you removed the plastic cover under the heatsink

>>58291687
go to a pc repair shop then, scrublord

>>58290832
>At 70° at stock you shouldn't even think about overclocking unless you want to fry eggs on your CPU.

He is using a fucking 4790k. 70C stock is normal because Intel purposefully gimps that cpu by using thermal grease under the IHS.

It won't matter even if he puts liquid cooling on it, the CPU will still have high temps.

The only way you can get it to proper temps is by delidding and either doing direct die cooling (DANGEROUS, needs special shim, only really recommended for water cooling), or applying liquid metal thermal interface material between the die and the IHS.

You'd get 10-20C reduction in temps if you delid and use liquid metal. Of course you can always be a good goy and buy the Intel Extreme platform for only $1000.

>>58291759
I actually have a 4690k, don't know if that negates anything you said tho

>>58291797
same 4th generation, same design

>>58291759

>Of course you can always be a good goy and buy the Intel Extreme platform for only $1000.

The ridiculous glue and thermal cum Intel use on their consumer parts hinders cooling so much. I hope Zen is at the very least soldered and the heatsink isn't glued down with enough adhesive to seal a door shut (as thats what the mnassive temp drop on a kabylake chip drama was all about).

File: DSCF2579 re.jpg (2MB, 2560x1920px) Image search: [Google]

2MB, 2560x1920px

>>58290794
>That still a dangerous game for retards and even after you physically delid you still need to 1) not be a retard 2) know what you are doing and 3) put it back together.

If you have 2 steady hands and you aren't a mouthbreathing fat moron who builds computers out of his allowance, then you can easily delid the cpu. I did it ages ago.

Just know that if you fuck up, then it's your fucking fault and you have to buy a new CPU from your own money. But that part comes with not being a moron.

>>58291854
Or buy this >>58291244 if you have the money.

I'm like 1.32 volts 4.5ghz 3570k

all good as long as your cooler is decent

Had this overclock for like 5 years now

>>58291279
>What worries me about delidding is not securing the heatsink back down (I assume you'd use some new adhesive) and thus it sliding as you secure the cooler back down.

the little cage on the motherboard will do that fine. The IHS will fidget around a little, but the surface tension between the new TIM you apply (preferably liquid metal) will prevent it from falling off unless you pry it open.

>>58291335
direct die cooling is possible, but not recommended for multiple reasons. One is that most coolers will sit too high to touch the die, the other is that the weight of the cooler may chip the corner of the silicon die, which is instant game over.

Some company made custom shims to support the weight of the coolers, so you could do direct die cooling easier.

It's mostly recommended with water cooling, where you only have to house a small connector on the CPU instead of a kilo of aluminium fins.

>>58290246
>4.2ghz @ 1.14 volts
That's barely over stock voltage. You are retarded. You need to be around 1.2, probably more

I have a CX650 with my 4690k at 4.7GHz 1.295v just fine

>>58292144
What are your temps on that?

>>58292185
A couple hours of prime95 will put it at 82* with a 212 EVO

>>58292229
Prime95 is bloody useless. Try IntelBurnTest, it can find in 2 minutes what prime95 cannot in 20 hours.

>>58291279
Back in the day AMD didn't use IHS (Heatspreaders). Core was bare and you could connect the bridges via a graphite pencil to overclock. The downside with a bare core is a large heatsink or liquid cooler could crack the core on a corner if you tightened it too much.

>>58291335
I hope they aren't putting it back on. It's like lapping your heatsink and using thermal pads instead of a good paste.

>>58293049
>Back in the day AMD didn't use IHS (Heatspreaders).

They did that for the Athlon and its derivatives (Athlon XP, Sempron, Duron).

K5, K6, K6-2, K6-3, Athlon 64, and everything afterwards all used IHS.

>I hope they aren't putting it back on. It's like lapping your heatsink and using thermal pads instead of a good paste.

You can replace the TIM between the die and IHS with liquid metal easily, and get much better temps while still keeping the IHS on. There's no downside.

Not using the IHS and cooling the bare die is only worth a damn if you use liquid cooling. With a normal heatsink, the weight alone may chip the die.