Jump to content

OC'ing 4770K and 'stability' definition


Rodrigodrt

Recommended Posts

New user here, hopefully I've read all the rules correctly :rolleyes:  (non english speaking also)

 

So, recently I've purchased a new rig, and it's specs are:

 

I7 4770K Haswell

MSI Z87 G45 GAMING

MSI TWIN FROZR 770GTX

2x4GB Corsair Dominators 1600/9/9/9/24

SSD Samsung 840 PRO

PSU Corsair 650TX

*Coolermaster X6 Elite (The tricky part, the whole sink and fan has an angle, so they wont fit with high profile memory sticks, so i mounted the FAN in the back of the sink, and instead of blowing air towards the sink, it now sucks the air in front of it through the heatsink, and dump it towards the 120mm rear fan)

 

And hearing so much about the 4770k OC capabilities, I've did my own researche and as I installed the SO and everything else, right off the bat I just tweaked a few settings and OC'ed:

 

Changed the FSB Multiplier to 41 from 35 right off the bat for 4.1Ghz for starters

Manually set VCore to 1.150, everything else on AUTO

 

Set RAM to XMP 1600 profile and manually set timings to 8-9-8-22 1T from 9-9-9-24 2T

 

Everything went out well, systems boots rapidly, no bsods or anything, ran prime95 for about 40 minutes (I know, should run for several hours) anyways, using blend tests it remains at 60~65C high temperatures, after about 15 minutes, then it (probably) start running tests with small fft's and the temps would go up to as high as 80~85C wich i believe is pretty normal for such kind of test, i didnt to run too much time like this though, and stopped at 40 mins mark.

 

 

The thing is, i wanna go for some for a 24/7 overclock , and yet remain in the 'safe zone', I decided to not go over 1.25Vcore, so I changed the settings again, and now to:

 

4.3Ghz @ 1.230VCore

Ring frequency @ 4000 (for some reason, it simply would even start prime with a 3900)

Ring voltage 1.200 (Auto, its always around 1.216v on auto no matter what i do, so i leave it on auto)

System Agent V +0.050,

Rest on auto and kept RAM with my "custom" timings

 

Will boot normally, now running the same tests I did before on prime, the first 15 minutes, the temp would stay @ 68~75C (for obvious reasons) and the problem is, when it starts with small fft's again, this time the temps sky rocket, it will reach 95~100C in a blink, almost throttling. I didnt noticed untill a couple of minutes and when i realized, i just stopped the test, so i really can't tell if my system is rock solid, I ran Large FFT tests only for about 15mins just to see if the temps would rise again but it remained at an average of 75C, however i kinda know this test isnt very CPU stressing.

 

 

I kinda ignored the fact and went to play some stuff such as PlanetSide2/Grid2/Black Flag and everything worked nicely, i kept monitoring my temps and by simply gaming, and browing and doing my everyday stuff, temps would never get past ~60C, wouldnt BSOD or anything.

 

So the question here is, I wont be able to run prime for like 12 hours due to heating, and since the temps are still low for my gaming purpose, i feel like i have some more room for pushing the frequencies thru 4.4/4.5 god knows 4.6 and remain within 1.25V since the temps should remain at low 60's Celsius, now how will i be able to tell if my system stable? can I simply run Large FFT's for a couple of hours if my temps will stay at 75C and consider it stable for my everyday use? Ignore prime if I can simply play my games with no problems? or even, insist in a Liquid Cooling system and try by all means be Small FFT stable?

 

 

Any insights are welcome, I apologize for the long text and for my faulty english :thx:

Edited by Rodrigodrt

Share this post


Link to post
Share on other sites

It's really up to you.  If you want to be able to say it's completely stable doing anything that anyone might possibly through at it, then passing all the tests for hours is the only way to know for sure.  But if it does everything you need it to without any problems, then you can keep pushing it.  But be warned, either way, if it runs at high temps consistently then it will hurt the life

Share this post


Link to post
Share on other sites

It's really up to you.  If you want to be able to say it's completely stable doing anything that anyone might possibly through at it, then passing all the tests for hours is the only way to know for sure.  But if it does everything you need it to without any problems, then you can keep pushing it.  But be warned, either way, if it runs at high temps consistently then it will hurt the life

 

Oh thats for sure, but then again, I keep monitoring the temps and right now at 4.3Ghz, with the same Vcore, it still wont get past 60-62C on any game I've tried, and I'm quite sure those are safe temps, I just can't run prime for too long, cuz at some point, at some kind of test it will really make the temps hit the ceiling, but it wont BSOD or halt. I wonder if I theres  way to set up prime in a way I could test stability with custom settings for hours and not have it come to that particular test in wich the temps will go nuts... Either way, I'm considering the Corsair H80 in hopes that I can pass the Small FFT tests... I'm quite sure my current fan isnt the worst, but isnt nowhere near the best, I just cant replace it right away...

Share this post


Link to post
Share on other sites

Seems like pretty high temps.  Did you change the bios setting for the cpu fan to turbo?  What tim/paste are you using?  You might even try re-seating the heatsink.  If the cpu fan is too close to the case fan it can cause an airflow issue, removing the case fan can help.  You might even try turning the fan around to blow toward the ram (not ideal for long term but might at least show you if it's an airflow issue or a cpu issue)  You can also run Unigine Heaven on loop, lower cpu stress than prime but much more 'real world'.

 

Share this post


Link to post
Share on other sites

Those temps seem normal. At 1.2v you are going to be in the 80c range just because the way Haswell chips are. The problem is the chip going much higher with "Small FFT" and thats not good. I did a mini i7 4770 oc guide here at OCC which may or may not help considering you already know what you are doing. When making the guide I used a Corsair H100 to keep it cool. I also did "Small FTT" in prime as its what stresses the cpu more. Using a Thermaltake C4 CPU cooler I get around 83c~ with 1.2v so you aren't very far off.

 

I would play around with the voltage and settings more at 4.2ghz mak that stable before going higher. My chip was a pain in the butt after 4.2ghz and had a hard limit of 4.6ghz with insane voltage. If you pass 4.5ghz, you are very lucky as at lot can't even hit 4.2ghz. without 1.2 vcore.

 

For the extra heat using during the "Small FFT" test It could be a few reasons for this. 1) the Ring Voltage is set to autio giving higher voltage 2) CPU cooler can't handle the haswell (most likely) or 3) your chip has a defective sensor or IHS issues.

 

 

Basically if it' crashes with "Small FFT" than your oc is not stable. If you continue to run the cpu in games and other apps than may come unstable when stressed and crash your computer. Most games don't even come close to stressing the cpu 100%. Your gaming scores may suffer from a faulty occ as the chip could be given a bunch of errors. Back  the cpu down and start tweaking it until it's stable, than raise the speed again and see what happens.

Share this post


Link to post
Share on other sites

Those temps seem normal. At 1.2v you are going to be in the 80c range just because the way Haswell chips are. The problem is the chip going much higher with "Small FFT" and thats not good. I did a mini i7 4770 oc guide here at OCC which may or may not help considering you already know what you are doing. When making the guide I used a Corsair H100 to keep it cool. I also did "Small FTT" in prime as its what stresses the cpu more. Using a Thermaltake C4 CPU cooler I get around 83c~ with 1.2v so you aren't very far off.

 

I would play around with the voltage and settings more at 4.2ghz mak that stable before going higher. My chip was a pain in the butt after 4.2ghz and had a hard limit of 4.6ghz with insane voltage. If you pass 4.5ghz, you are very lucky as at lot can't even hit 4.2ghz. without 1.2 vcore.

 

For the extra heat using during the "Small FFT" test It could be a few reasons for this. 1) the Ring Voltage is set to autio giving higher voltage 2) CPU cooler can't handle the haswell (most likely) or 3) your chip has a defective sensor or IHS issues.

 

 

Basically if it' crashes with "Small FFT" than your oc is not stable. If you continue to run the cpu in games and other apps than may come unstable when stressed and crash your computer. Most games don't even come close to stressing the cpu 100%. Your gaming scores may suffer from a faulty occ as the chip could be given a bunch of errors. Back  the cpu down and start tweaking it until it's stable, than raise the speed again and see what happens.

 

 

It doesn't really crash for, say, 5 minutes on small FFT tests, but the temps are too high, 95C~ all the time, It may be actually stable, but I wouldnt dare run for too long at these temps, and then again, I was able to keep using it the whole day long for gaming and surfing with no problems , peak temps were 69C, but an all time average of 60C.

 

And, it may all come down to the cooling, after reading some replies, I guess I'll get a WC Corsair H80I, and try lower those temps and keep it on small FFT testing for a couple of hours.

 

Currently, as I said, my X6 Elite is not mounted as it supposed to, the front fan had to be mounted in the back, and instead of blowing air through the sink, it draws air from it, so I think this may be the culprit, heres a pic of how it is:

 

dia.jpg

 

 

 

couldnt put it the right way or the fan would not fit with the RAM heatsink being too tall, now not changing the subject of the topic much, in case i swap this for a H80I, would it be best to place it on top, or at the rear of the case? blowing/drawing or air inside/outside?

 

 

thanks

Share this post


Link to post
Share on other sites

I'm going to agree with you and say that cpu cooler is at lest one reason why you can getting high tempatures. It's very possible that's all the problem is, though when you get a new cooler don't be surpised if you can't get a stable 4.5ghz overclock under 1.2v or at all.

 

Also if you can get the Thermaltake C4 cooler, it's extremely cheap compared to a H80 /100 and works very well.

Share this post


Link to post
Share on other sites

I'm going to agree with you and say that cpu cooler is at lest one reason why you can getting high tempatures. It's very possible that's all the problem is, though when you get a new cooler don't be surpised if you can't get a stable 4.5ghz overclock under 1.2v or at all.

 

Also if you can get the Thermaltake C4 cooler, it's extremely cheap compared to a H80 /100 and works very well.

Thanks, while my current goal would be to stay at 4.4/4.3 , respecting my 1.25v rule, (maybe i'll even consider a little bit more after I read a lot of ppl saying its safe until 1.3~1.35) I dont want to heave trouble with an upgrade to stay at 4.2 and gain only 100mhz. I took a look @ the C4, its a possiblity, and cheaper too, yet theres always this problem with big coolers not fitting with high profile RAM. As for the H80, I saw some nice reviews about it, and since I have a spare Logitech G600 Mouse, I was proposing a 1:1 trade with the guy thats selling it. Now would the C4 have the same performance, do you know? what i know is that Coolers such as a Noctua NHD14 would mach a WC performance, if not, outcome a little bit better, however they would never fit in my case cause they're monsters.

Edited by Rodrigodrt

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...