4670K - Getting ~90C @ 4.2GHz Overclock

Akhilcool · Aug 6, 2014

Hi,

I just upgraded from the stock Intel HSF to the CM Hyper 212 EVO with Gelid GC-Extreme thermal paste and finished installing it yesterday. I wanted to get the most out of my i5 4670K knowing that I wouldn't get much of an overclock anyways with the Haswell-Gen CPU.

Anyways, after installing the new cooler, I tried to overclock to 4.2 GHz but had to overvolt my CPU all the way to 1.265V just to remain stable without getting BSOD on Prime95! At those settings, I was getting temperatures all the way up to the low 90's. I was expecting my CPU to get around 4.3-4.5 GHz but I don't think I can push this CPU any further as I will probably hit near 100C just trying to get to 4.3!

I know that some people can get chips from bad batches for overclocking but this seems a little too much. I searched online and saw that people were getting 4.4GHz with <1.25V core voltages. And I'm nearly at 1.27 just trying to get 4.2GHz. I thought it may be a bad job with the HSF installation so I took it off 3 times trying to reapply the thermal paste and installing again. One of those times, I even tried a bit of the stock thermal paste that came with the Hyper 212 EVO and was suprised to notice that the temps were practically exactly the same as they were with the GC-Extreme.

I'm new to overclocking and so I'm not sure if I am doing something wrong or if this is the just the sad reality of owning a bad Haswell chip. Just now, I tried to run the CPU at 3.8 GHz with my voltage set to adaptive mode and I'm getting temperatures of around 82 C using the Small FFTs test in Prime95 (that's the same type if test I used throughout my benchmarking, not sure if that helps). I'm really not sure what is wrong.

These temps seem almost like there ones you would expect from the stock Intel HSF cooler.

If anyone can help me out I would be grateful.

Calnin · Aug 6, 2014

The overclocks vary chip to chip. Some get luckier with one that requires less voltages on the same motherboard.

As for your overclock, what motherboard do you have?

InvalidError · Aug 7, 2014

Akhilcool :

Some people get 4.4GHz below 1.25V but many also don't get above 4.3GHz without going to extremes.

Since Intel forces people to buy their top-binned chips to get the "right" to overclock, most of the headroom has already been skimmed off by Intel. Anything beyond whatever Intel uses as a safety margin is luck-of-the-draw.

This is why overclocking is called the chip lottery.

zAustin · Aug 7, 2014

A push-pull configuration should drop your temperatures down considerably if you're not doing so already. I noticed about a 5-12C difference adding a 2nd fan to my T4 (newer version of the 212). Then again, my fans spin at 1,800RPM all the time. I'm upgrading to the H90 tomorrow for quieter and even cooler performance. But I have no complaints from the T4 heatsink, it is a excellent $30 cooler, and the 212/+ has a good reputation itself. But like Calnin said, chips vary, some get lucky and have excellent, low voltage/temp overclocks, others, very high using similar equipment. I have my voltage set to adaptive at 4.2GHz at about 1.27v and getting about 60C under load. I have overclocked my i5-3570k to 4.6GHz, but I had to revert it due to some instability issues between my CPU and GPU.

Akhilcool · Aug 7, 2014

Calnin :

The Z97-A from Asus. I'm a first timer as far as OC'ing goes so everything is on auto except for clock speed and core voltage.

CompuTronix · Aug 7, 2014

Akhilcool,

Here's the problem:

" ... Core i 2nd, 3rd and 4th Generation CPU's have AVX (Advanced Vector Extension) instruction sets. Recent versions of Prime95 run AVX code on the Floating Point Unit (FPU) math coprocessor, which produces unrealistically extreme temperatures. The FPU test in the software utility AIDA64 shows the same results.

It's not necessary to run AVX code for thermal testing. Prime95 v26.6 produces temperatures on 3rd and 4th Generation processors more consistent with 2nd Generation, which also have AVX instructions, but do not suffer from thermal extremes due to having a soldered Integrated Heat Spreader and a 35% larger Die. ... "

Use Prime 95 Version 26.6 Small FFT's. Here's the link: - http://windows-downloads-center.blogspot.com/2011/04/prime95-266.htm

You'll be happy with the temperatures that 26.6 will yield, which will be 10 to 20C cooler than the current 28.5 version.

For the complete explanation, please read: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

Also, nowhere in this thread is your ambient temperature mentioned. All processor temperatures are based on ambient, so it's a variable which is a critical piece of information. 22C is "Standard Ambient", which is the industry standard for a proper computer environment.

" ... Here's the temperature conversions and a short scale:

Cx9/5+32=F ... or ... F-32/9x5=C ... or more simply ... 1C = 1.8F

30.0C = 86.0F Hot
29.0C = 84.2F
28.0C = 82.4F
27.0C = 80.6F
26.0C = 78.8F Warm
25.0C = 77.0F
24.0C = 75.2F
23.0C = 73.4F
22.0C = 71.6F Standard ... or ... 22.2C = 72.0F
21.0C = 69.8F
20.0C = 68.0F Cool

As Ambient temperature increases, thermal headroom and overclocking potential decreases. ... "

CT :sol:

Akhilcool · Aug 7, 2014

CompuTronix :

Akhilcool,

Here's the problem:

" ... Core i 2nd, 3rd and 4th Generation CPU's have AVX (Advanced Vector Extension) instruction sets. Recent versions of Prime95 run AVX code on the Floating Point Unit (FPU) math coprocessor, which produces unrealistically extreme temperatures. The FPU test in the software utility AIDA64 shows the same results.

It's not necessary to run AVX code for thermal testing. Prime95 v26.6 produces temperatures on 3rd and 4th Generation processors more consistent with 2nd Generation, which also have AVX instructions, but do not suffer from thermal extremes due to having a soldered Integrated Heat Spreader and a 35% larger Die. ... "

Use Prime 95 Version 26.6 Small FFT's. Here's the link: - http://windows-downloads-center.blogspot.com/2011/04/prime95-266.htm

You'll be happy with the temperatures that 26.6 will yield, which will be 10 to 20C cooler than the current 28.5 version.

For the complete explanation, please read: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

Also, nowhere in this thread is your ambient temperature mentioned. All processor temperatures are based on ambient, so it's a variable which is a critical piece of information. 22C is "Standard Ambient", which is the industry standard for a proper computer environment.

" ... Here's the temperature conversions and a short scale:

Cx9/5+32=F ... or ... F-32/9x5=C ... or more simply ... 1C = 1.8F

30.0C = 86.0F Hot
29.0C = 84.2F
28.0C = 82.4F
27.0C = 80.6F
26.0C = 78.8F Warm
25.0C = 77.0F
24.0C = 75.2F
23.0C = 73.4F
22.0C = 71.6F Standard ... or ... 22.2C = 72.0F
21.0C = 69.8F
20.0C = 68.0F Cool

As Ambient temperature increases, thermal headroom and overclocking potential decreases. ... "

CT :sol:

So if I get BSOD from using the latest version of Prime95, should I disregard it and only consider if I get a BSOD from version 26.6 of Prime95?

[Edit: BTW my ambient temps are around 24 C.]
[Edit: Ok I just tried v26.6 and I am getting around 20 C cooler temps from its small FFT test.]

Akhilcool · Aug 7, 2014

IntelBurnTest v2.54 on high gives me temps around 10C warmer (still 10C cooler than Prime95 v28.5) than those from the Prime95 v26.6 small FFT test as well. Should I disregard these temps as well?

Akhilcool · Aug 7, 2014

I have a HAF 912 case with 2 front 120s, 1 rear 120 and a top 200mm+ fan (not sure of the exact size). I thought the top fan was disturbing the air flow so I disconnected it but saw no change in temps. Plus, have you tried using Prime95 v28.5 on small FFT?

Akhilcool · Aug 7, 2014

The fact that your 4690k remains under 50 at 4.7 GHz seems odd. Either you haven't put a full load on it before or your temp probes are faulty because there are many people on water cooling that wish they had those temps at 4.7 GHz. Try using Small FFT on prime95 v28.5 as I said before and see if you can still remain under 50.

CompuTronix · Aug 7, 2014

Akhilcool,

For nearly 8 years and over 2,500 hours of research, testing and analysis of Intel processor temperatures, I've run every utility available.

For thermal benchmarking, I use P95 v26.6. For stability testing, I run every utility, app and game on my rig. The bottom line is that I rely on P95 v26.6 for testing my temperatures.

CT :sol:

Akhilcool · Aug 9, 2014

Ok so I tried overclocking my CPU to 4.3 GHz and for some reason, I can't get it stable until I push the core voltage to 1.28V! This is not only to get my stability test to run without receiving a BSOD, but until I pushed the voltage to 1.265V, I couldn't even boot to Windows before seeing a BSOD. When I pushed it past 1.265, I was able to boot and run a stability test but it would give me a BSOD soon afterwards until I went all the way to 1.28V as I said before. On the other hand, before I overclocked to 4.3, I had it set to 4.0 and it only required 1.15V to pass my stability testing. Why does it require so much more of a voltage (0.13V specifically) just to get an additional 0.3 GHz overclock?

P.S. I used P95 v26.6 and x264 to benchmark and my temps were fine, but I can't get to 4.3 GHz without running at 1.28V.

CompuTronix · Aug 9, 2014

It takes about .040 to .050 volts for each increase of 100 MHz. Accordingly, to get from 4.0 @ 1.15 to 4.3, you should need at least 1.27, so 1.28 is about right.

Regardless, not all 4670K's are created equal. If you've overclocked correctly and followed the guides that are available, then you simply have a CPU that's not exactly a stellar overclocker. Some call it the silicon lottery, others call it the luck of the draw.

I had to go through four different 4770K's before I found the one I have now, which runs at 4.7 on 1.28 Vcore. Not the best, but better than most.

Sorry you didn't get a better chip.

CT :sol:

Akhilcool · Aug 9, 2014

Is it ok then to run at that high of a voltage (1.28)? I'm getting temps around the high 70's to low 80's with P95 v26.6. I would probably get BSOD with P95 v28.5 or IBT so I haven't even tried those. I also use x264 and my temps are in the mid 80's using it so I'll let that run overnight. I'm just worried that I'll damage the CPU over time at that voltage.

EDIT: My temps actually are around 70 C with x264, not high 70's to low 80's. Also, I actually need to overvolt to 1.295V to keep the 4.3 GHz clock (I receive a BSOD with anything lower using x264). -.-

Pr3di · Aug 9, 2014

@Akhilcool - if it makes you feel better, I tried to OC my i5 4670k, and I kept in under 1.27.
However, if I try anything over 4.0 Ghz, I have to manually reset CMOS, because my PC crashes and doesn`t boot.

So be happy with your 4.2, and remember it could be worse.

CompuTronix · Aug 11, 2014

Regarding Vcore:

" ... Section 8 - Overclocking and Vcore

Overclocked processors can reach up to 150% of their Thermal Design Power (TDP) when using manual Core voltage (Vcore) settings, so high-end air or liquid cooling is critical. Every processor is unique in it's overclocking potential, voltage tolerance and thermal behavior.

Regardless, excessive Vcore and temperatures will result in accelerated "Electromigration" - https://www.google.com/?gws_rd=ssl#q=Electromigration - which prematurely erodes the traces and junctions within the processor's layers and nano-circuits. This will eventually result in blue-screen crashes, which will become increasingly frequent over time.

CPU's become more susceptible to Electromigration with each Die-shrink, so 22 Nanometer architecture is less tolerant of over-volting. Nevertheless, Vcore settings should not exceed the following:

-> Core 2

1st. Generation 65 Nanometer ... 1.50 Vcore
2nd Generation 45 Nanometer ... 1.40 Vcore

-> Core i

1st. Generation 45 Nanometer ... 1.40 Vcore
2nd Generation 32 Nanometer ... 1.35 Vcore
3rd Generation 22 Nanometer ... 1.30 Vcore
4th Generation 22 Nanometer ... 1.30 Vcore <---

When tweaking your processor near it's highest overclock, keep in mind that for an increase of 100 MHz, a corresponding increase of approximately 40 to 50 millivolts (0.040 to 0.050) is required. ... "

CompuTronix · Aug 12, 2014

Sean76 :

That's not good advice.

4670K - Getting ~90C @ 4.2GHz Overclock

Distinguished

Dignified

Titan

Honorable

Distinguished

Intel Master

Distinguished

Distinguished

Distinguished

Distinguished

Intel Master

Distinguished

Intel Master

Distinguished

Honorable

Intel Master

Intel Master

Share this page