4670K-What's the hurt with 1.4V (or more) if i can keep the heat under 80C?

rrc2

Reputable
Jun 2, 2014
9
0
4,510
I am running the 4670K with a CNPS11X(air-cooled) in an ASUS Maximus VI Hero . I set the adaptive voltage to total 1.375 with all cores at 46(4.6GHz). In CPU-Z it shows the voltage at 1.39V. HWMonitor shows highest core temp at 77C under Prime95 under high FFT.

If I can keep the cooling down (under 80C is what I am reading that seems acceptable), what's the hurt in going to voltages higher than 1.4V, to 1.45v, to get upwards of 4.7-4.8Ghz out of the CPU?
 
Solution
The term is Electromigration - https://www.google.com/?gws_rd=ssl#q=Electromigration

Excessive Vcore and temperatures will result in accelerated "Electromigration", which prematurely erodes the traces and junctions within the processor's layers and nano-circuits. This will in turn eventually result in Blue-Screen crashes, which will become increasingly frequent over time.

Short-term solution (not recommended): Increase Vcore to regain stability, which will further accelerate Electromigration and shorten the time interval to the reoccurrence of Blue-Screen crashes.

Interim solution: Decrease overclock to temporarily regain stability.

Long-term solution (recommended): Limit your overclock, Vcore and temperatures to stay within...

dandn0ten

Reputable
Feb 19, 2014
754
0
5,360
thats hella high voltage for such a low frequency. higher voltage usually causes heat issue which also reduces cpu life span. btw dont run test on adaptive mode because thats reason why your vcore is so high. try to lower the vcore until you find a stable point with manual mode and when youre done with stress test, switch to adaptive for daily usage
 

rrc2

Reputable
Jun 2, 2014
9
0
4,510
Thanks for the response. The voltage i am giving is after increase GHz and V under manual voltage mode. So, it takes 1.312 (1.296 in BIOS) to run stable @ 4.5GHZ. I can do 4.4GHz @ 1.25 in bios. 4.6Ghz takes the 1.375V (and runs closer to 1.4V in CPU-Z under load) to run stable.

At 4.6Ghz/1.375 in bios (pushing towards 1.4V in CPU-Z) temps are under 80C.

Yes, these are higher voltages than I've seen usually (OMGosh when I read at/about 1.1 @ 4.7). If I keep the heat under control is this okay? Does the voltage hurt or is it the heat of a combination or both and to what extent damaging?
 

dandn0ten

Reputable
Feb 19, 2014
754
0
5,360
from what you mention on the first post, i can tell that is good temp for the high voltage and yes as long as you dont exceed vcore to 1.4 ( dangerous zone) and stay below 80c then youre good to rock. :)
 

rrc2

Reputable
Jun 2, 2014
9
0
4,510
What is the danger in going over 1.4V?

What if water cooling allowed me to get temps under 80C over 1.4V which I'd push to get 4.7-4.8Ghz?
 

rrc2

Reputable
Jun 2, 2014
9
0
4,510
So lets talk about V at 1.45 - 1.5. If I am able to get 4.8Ghz at 1.45 under 80C is that okay? Pretty sure I can with corsair watercooling.
 

dandn0ten

Reputable
Feb 19, 2014
754
0
5,360
thats fine as long you dont exceed over 80c but im pretty sure you cant stay under 80c with that voltage even with AIO cooler. btw what longest time on your stress test? because its usually about few hours or maybe more and run on several stress tests not only prime95
 

rrc2

Reputable
Jun 2, 2014
9
0
4,510
At this point, it seems that 43 cpu and 42 cache @1.265 is where I am at. The consensus I get is that higher voltage = lower life span and I would assume higher temps also adds to this. I couldn't get long term stability (would get the 124 BSOD just outta the blue under no load).

Intel sells a K model advertising unlocked, maybe I don't get the big picture, but you'd think there would be tighter quality control on those type of processors for better consistency. It's kind of like buying a $200+ pack of baseball cards hoping you get that winning card. Just saying...
 

dandn0ten

Reputable
Feb 19, 2014
754
0
5,360
yeah its just like a lottery game but we called it as silicon lottery. thats why every cpu is different and also depends on your luck. im lucky to own a pretty good chip. its currently running at 4.8ghz on 1.285 vcore. i will delid it someday when i can find a safe method to do so.
 

CompuTronix

Intel Master
Moderator
The term is Electromigration - https://www.google.com/?gws_rd=ssl#q=Electromigration

Excessive Vcore and temperatures will result in accelerated "Electromigration", which prematurely erodes the traces and junctions within the processor's layers and nano-circuits. This will in turn eventually result in Blue-Screen crashes, which will become increasingly frequent over time.

Short-term solution (not recommended): Increase Vcore to regain stability, which will further accelerate Electromigration and shorten the time interval to the reoccurrence of Blue-Screen crashes.

Interim solution: Decrease overclock to temporarily regain stability.

Long-term solution (recommended): Limit your overclock, Vcore and temperatures to stay within reasonable guidelines.

You might want to read the following Sticky: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

1.4 Vcore is not recommended. Unless you're into competitive overclocking on extreme cooling, most overclocking guides as well as many experts will insist that 1.3 Vcore is as high as a 22 nanometer processor should be pushed.

Vcore settings should not exceed the following:

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

At 1.3 Vcore, whatever stable overclock a particular processor sample is capable of giving you is what you should consider to be it's limit. I know that's not the answer you want to hear, but that's the most appropriate answer I can give you.

Only the very best of samples will give you 4.8. I spent a huge amount of time testing 5 of these i7 4770K's before I found the one I'm running. The only way she'll give me 4.8 is if I turn off Hyperthreading, but I'm very pleased with it at 4.7.

As dandn0ten has pointed out, when it comes to overclocking, not all processors are created equal; each one is unique.


CT :sol:
 
Solution

rrc2

Reputable
Jun 2, 2014
9
0
4,510
Thank you for your response; as well thank you to dandn0ten.

I've resigned to the fact I didn't get one of the better samples.

Last question, how will graphine change our situation? :)