Ivy Bridge 3570K and EVO 212 CPU COOLER temps

[mohammed441]

Recently about 2hours ago, I did an overclock with my fairly new CPU. About 2 weeks old, I decided to overclock to get the most performance out of my PC, so I did some researching, followed a tutorial online on how to manage to change clock speeds etc.

So that's what I did, I thought jumping from 3.4Ghz to 4.5Ghz (3570k) was okay as I saw many people with similar spec PC's get the overclock without any problems. So after I ran the overclock, I restarted my PC, while browsing for more info, my comp froze for like 2 seconds and then went back to normal (got a bit worried). So I decided to download Prime95 and start the torture test. After running the torture test I was exceeding temperatures on most of my cores at over 90c!! (after 1hr20mins ish length) I was really worried, so I stopped the overclock and looked around if it that was normal because a lot of people were saying that ivy bridge chips get extremely hot.

Right now I'm running idle with 4.5Ghz, at ~37c - is that too much? Were my prime95 temps too high? - by the way during the torture test, I got 0 errors.

Do I need to re apply thermal paste? Should I lower the clockspeed to 4.2 or lower?

Please help..

thanks!

 

[bpmcleod]

Are you using a custom OC? Or just a preset overclock in your UEFI ? 90c at 4.5 is pretty high. I know all chips are different, but the 3570k should be able to hit 4.5 with temps around 75 or so on air. WHat fans do you have on you 212 evo ?

 

[mohammed441]

Are you using a custom OC? Or just a preset overclock in your UEFI ? 90c at 4.5 is pretty high. I know all chips are different, but the 3570k should be able to hit 4.5 with temps around 75 or so on air. WHat fans do you have on you 212 evo ?

hello, thank you for responding. I am using the bios UEFI that came with my motherboard (Z77 UD3H)

So you're saying I shouldn't be hitting over 75 on my torture test? - btw, my temps dropped from 103/104/105 down to 80c at times, it didnt stay at over 90 for the majority

My 212 EVO heatskin came with the fan, thats the one im using

 

[kiddagoat]

If you could please, fill out your full system specs for us. That would help a bit.

Also, I see this issue a lot with overclocking threads and the usual questions that are asked. This is something I see a lot on tech forum communities where people just attempt to copy and paste settings from another machine and assume it is going to work, and then when it doesn't work the person believes something has to be wrong with their hardware. I get so many people attempting to return or exchange product at work because it doesn't hit the same OC they seen on a forum.

Just because you see a similar system, same CPU, same cooler, same motherboard, etc.... Does not mean you will get similar overclocking potential. Every chip is different, every motherboard, every stick of RAM.

Anyone who has been overclocking for awhile and has the experience will tell you, on a CPU that is only 2 hours old, you should not be trying to OC it that high. You have to do it in stages which the guide you read probably told you to do.

Temperatures are going to vary quite a bit. I would say assuming proper mounting of the heatsink, good thermal compound application, and proper case airflow, you should be keeping that CPU below 80C.

Also, I prefer to use IntelBurnTest for OC burn-in and stability testing. Try that instead of Prime95 and see how that turns out for you.

So as noted by bpmcleod, are you using a predetermined OC profile or are you manually setting the UEFI??

Give us the settings currently being used in the UEFI. I believe on Gigabyte boards all those settings are listed under M.I.T.

 

[mohammed441]

If you could please, fill out your full system specs for us. That would help a bit.

Also, I see this issue a lot with overclocking threads and the usual questions that are asked. This is something I see a lot on tech forum communities where people just attempt to copy and paste settings from another machine and assume it is going to work, and then when it doesn't work the person believes something has to be wrong with their hardware. I get so many people attempting to return or exchange product at work because it doesn't hit the same OC they seen on a forum.

Just because you see a similar system, same CPU, same cooler, same motherboard, etc.... Does not mean you will get similar overclocking potential. Every chip is different, every motherboard, every stick of RAM.

Anyone who has been overclocking for awhile and has the experience will tell you, on a CPU that is only 2 hours old, you should not be trying to OC it that high. You have to do it in stages which the guide you read probably told you to do.

Temperatures are going to vary quite a bit. I would say assuming proper mounting of the heatsink, good thermal compound application, and proper case airflow, you should be keeping that CPU below 80C.

Also, I prefer to use IntelBurnTest for OC burn-in and stability testing. Try that instead of Prime95 and see how that turns out for you.

So as noted by bpmcleod, are you using a predetermined OC profile or are you manually setting the UEFI??

Give us the settings currently being used in the UEFI. I believe on Gigabyte boards all those settings are listed under M.I.T.

my processor is not 2 hours old, I overclocked two hours ago!! I have built and used this comp for about 2 weeks now if not more.. Like stated before I used the UEFI and changed the core clock speed from 34 to 45

 

[bpmcleod]

SO you manually did your overclock? What did you set the vcore to for the CPU then? This could be another reason why its getting so hot. Maybe you set it to a number that is too high. Also as he stated in his previous post, overclocking should be done gradually. Going straight from 34 to 4.5 is an enormous jump. 3.4 to 4 is ok because most if not all 3570ks should be able to be stable at 4.0 at stock voltages as intel over volts the CPUs from stock anyways. Past 4.0 and maybe even past 3.8-3.9 if you want to be safe, should be done in increments and slowly at that. Going up to 4.1 and finding the lowest stable voltage for it, then 4.2 etc. Plus you should probably find some higher CFM/RPM fans to use on the 212. Im not sure how good the stock fans are as I watercool myself, but stock fans are usually mid grade at best.

 

[mohammed441]

SO you manually did your overclock? What did you set the vcore to for the CPU then? This could be another reason why its getting so hot. Maybe you set it to a number that is too high. Also as he stated in his previous post, overclocking should be done gradually. Going straight from 34 to 4.5 is an enormous jump. 3.4 to 4 is ok because most if not all 3570ks should be able to be stable at 4.0 at stock voltages as intel over volts the CPUs from stock anyways. Past 4.0 and maybe even past 3.8-3.9 if you want to be safe, should be done in increments and slowly at that. Going up to 4.1 and finding the lowest stable voltage for it, then 4.2 etc. Plus you should probably find some higher CFM/RPM fans to use on the 212. Im not sure how good the stock fans are as I watercool myself, but stock fans are usually mid grade at best.

I went onto CPU frequency advanced settings, then changed the core clock speed from 34 to 45.

I took your advice by the way, and went down to 42 (4.2Ghz) and I used that software IntelBurnTest and made it so it would run 5 times, (it said 5 or more is recommended) it passed with no errors and a good thing that my temperatures didn't even exceed ~75. The thing I am noticing though, (I dont know if i am being paranoid) but my fans are louder than usual when i play games.

Guys, thanks for all the help, just to be sure i will probably purchase more 120mm fans because my mobo and case fans support more.

and sorry, I forgot to share my computer spec if this helps:
600W Corsair Builder Series CX CP-9
2GB Asus GTX 660 Ti DirectCU II OC
Gigabyte GA-Z77X-UD3H, Intel Z77
1TB Seagate ST1000DM003 Barracuda
Intel Core i5 3570K,1155, Ivy Bridge (OC'D from 3.4Ghz to 4.2Ghz)
Cooler Master Hyper 212 EVO
GIGABYTE Z77-UD3H
8GB (2x4GB) Corsair DDR3 Vengeance
Samsung SH-118AB/BEBE Black 18x SAT

 

[bpmcleod]

If you are manually chaning the CPU frequency, are you letting the vcore change automatically? Im slightly confused in this. If you are manually changing your multiplier, you should be manually setting the voltage as well.

 

[mohammed441]

If you are manually chaning the CPU frequency, are you letting the vcore change automatically? Im slightly confused in this. If you are manually changing your multiplier, you should be manually setting the voltage as well.

Hey. Sorry for the misunderstanding.

I changed my Clock Speed Ratio and kept everything else as auto (default, didnt touch anything else):

If you want to know my voltages..
My CPU Vcore 1.80V
My CPU Vtt: 1.050V
My CPU VLL: 1.800V

 

[Pehla]

guys... i belive hyper evo is good coler but it cant handle 4.5ghz OC..i just think u need to lower ur clocks to 4,2 ghz and u will be fine!!

 

[bpmcleod]

Ive seen plenty run that cooler at 4.5 or so. They just had better fans then the stock ones

 

[buildzoid]

Dude 1.80 vcore is way too much voltage that's not too far from what was used to set the i7 3770k freq. WR
Auto voltage control on even on really expensive motherboards is pretty bad so don't ever use it for anything unless you want to cook a chip.

How to set voltage safelyish(this is never 100% safe):
1. take voltage to 1.20v and start adding .025v till you manage to run 3 consecutive runs of cinebench without crashing
2. once you do that start taking away by .005v until you can't pass the 3 cinebench runs
3. now add .01v. that should be you minimum stable voltage now you can try a burn in test if you like, but I just run a bunch more cinebench runs in the background of internet surfing.
4. If you crash go to step 3.

 

[adulaamin]

Yeah your vcore is too high... my 3770k does 4.5 at 1.2v also with the same cooler albeit using 2 Scythe AP-29s in push/pull... never goes above 75C in intelburntest 10 passes max mem... you might already be killing your chip with that 1.8 vcore...

 

[bpmcleod]

I think he was stating his pll voltage. All 1155 boards are stock 1.8 pll voltage due to sb chips. I doubt he has 1.8 going through his chip or it would already be fried since ib chips pll is max 1.7 or so. I think he said earlier he has it on auto.

 

[cadaveca]

since ib chips pll is max 1.7 or so

I've seen this reported a few times now, and 1.6V as well. It's not true. Stop spreading BS.

vPLL for IVB chips is 1.8 V.



Not 1.6 V or 1.7 V. Stope believing everything you read on the internet.


The real voltage can be found here(page 86):

http://www.intel.com/content/www/us/en/processors/core/3rd-gen-core-desktop-vol-1-datasheet.html


And just FYI, Intel recommends the MINIMUM is 1.71V, and the MAX 1.89V.

I have found MANY MANY voltage settings are reported incorrectly on various forums, and they are all up above in that diagram.

 

[bpmcleod]

Apologize for having the voltage off. Regardless, his chip would still be fried long before it reached 1.8 on air.

 

[cadaveca]

Apologize for having the voltage off. Regardless, his chip would still be fried long before it reached 1.8 on air.

Not with current CPUs. Simply, the CPU would run very hot, throttle, and run at a lowered speed, if not shut down. IF no load is placed on the CPU for very long, the voltage is harmless. 1.8V-2.0V is the range used by the guys running LN2. Chances of killing and Intel chip with voltage is near zero with 1155 and 2011 sockets, as evidenced by Intel offering a warranty that will replace a CPU, with no questions asked, for all chips on those sockets, and each warranty is less than $50($25 for 3770K).

http://click.intel.com/tuningplan/

The Performance Tuning Protection Plan being offered by Intel is a chance for you to experiment with the overclocking features of your processor without the worries of what will happen if you push the processor too far. The Plan allows you a single processor replacement, hassle-free, from our customer support. This is in addition to your standard 3 year warranty. In other words, if it fails under normal usage, we will replace it under the standard warranty; if it fails while running outside of Intel's specifications, we will replace it under the Performance Tuning Protection Plan.

Some chips will die with voltage, sure, but if it does, you can even buy the warranty, wait 30 days, and they'll STILL replace the chip.

 

[Widjaja]

I've seen this reported a few times now, and 1.6V as well. It's not true. Stop spreading BS.

vPLL for IVB chips is 1.8 V.

Not 1.6 V or 1.7 V. Stope believing everything you read on the internet.

The real voltage can be found here(page 86):

http://www.intel.com/content/www/us/en/processors/core/3rd-gen-core-desktop-vol-1-datasheet.html

And just FYI, Intel recommends the MINIMUM is 1.71V, and the MAX 1.89V.

http://www.techpowerup.com/forums/attachment.php?attachmentid=50547&stc=1&d=1363544787

I have found MANY MANY voltage settings are reported incorrectly on various forums, and they are all up above in that diagram.

I often wonder if some of these alleged voltage and over clock results coinciding with low temperatures and 100% stability are nerd trolls attempting to make other nerds feel bad about the meager overclock they are getting and much higher temperature results.

I know all hard ware is capable of different overclocking capabilities there's a lot out there which makes me think BS.

I remember my Brother was a hard core over clocker and this admin on the forum he went to listed his CPU overclock higher than his in the system specs.
My Brother then asked the guy who he had replies from in the past, how he achieved the overclock and did not respond.

 

[cadaveca]

I often wonder if some of these alleged voltage and over clock results coinciding with low temperatures and 100% stability are nerd trolls attempting to make other nerds feel bad about the meager overclock they are getting and much higher temperature results.

I know all hard ware is capable of different overclocking capabilities there's a lot out there which makes me think BS.

I remember my Brother was a hard core over clocker and this admin on the forum he went to listed his CPU overclock higher than his in the system specs.
My Brother then asked the guy who he had replies from in the past, how he achieved the overclock and did not respond.

Right. Honestly, I think this happens less today than it did say 5 years ago, but yeah, I am sure it's still common practice for many. Simply put, I don't care if anyone has a faster rig than I do, I don't compete is the "pro" overclocking circles, and only want to ensure that everyone enjoys their PCs.

I think, part of the problem is that extreme users, who only benchmark, have very different requirements for voltages and such for benchmarking, than is required for 24/7 operation. Many people will read these extreme guys, looking for tips to get more out of their systems, and then try what those guys do, not realizing that thing are very different sub-zero vs ambient. There is very much some art left to running a system at sub-zero, let me tell you.

Over the years, I've learned which guys in the extreme scene to listen to, seen who influences who, and learned who works for what companies. In nearly every instance, those guys running extreme clocks work for one company or another, or their close friends do. I ahte to say it, but it's only these guys I ever pay any attention to, and even then, everything is taken with a grain of salt, or two.

That said, this then also rolls into this situation in this thread.


The OP has a cooler recommended by many, but when he runs heavy loads, it appears to be inadequate. For stress-testing, sure, you bet it is, but normal loads...like gaming, it should be fine. If not, it's due to either a high-leakage chip, or a bad TIM job under the IHS.

Not many understand that 3770k @ 4.5 GHz, usually pulls around 100W only, so cooling such is actually far easier than most imagine. Watercooling is very much a waste on this platform.


90C is fine for 3570k, at prime, with that cooler. There's no need for any alarm here, at all.

 

[Sasqui]

If not, it's due to either a high-leakage chip, or a bad TIM job under the IHS.

Suspect the latter