First overclocking experience - do my results look reasonable?

deroberg

Honorable
Dec 27, 2013
6
0
10,510
I have a 4690k with a hyper 212 plus cooler, and I followed a piece of advise saying to set the multipilier for 4.6ghz with a vcore of 1.2v. This is supposed to be a quick test of its potential. It booted fine, telling me I have a decent chip.

Since everything seemed to be going smoothly, I proceeded to stress test it in that configuration, while playing minecraft :). After around 20 mins I stopped it. My highest core temp was 67c, I dropped it down to 4.3 ghz. I left vcore at 1.2v. I don't have a good reason for going faster than this now. After I get my video card I will see if I should take advantage of any more CPU overclock.

Screen capture

This is what I am seeing on my currently running test. I will probably let it run overnight.

I see that the temps haven't changed since going to 4.3ghz, but I would only expect a significant change if I adjusted voltage. Am I right about this?

My main concern is that these temps are lower than I expected. 40 mins in now, and peak temp is 66c. I keep hearing about how hot haswell runs. Are these temps reasonable for the voltage I am at? Could I be getting inaccurate readings? Do I just need to stop worrying and be happy I got a good chip?

Any other tips would be appreciated. I know it could probably use some fine tuning. Beyond where I am at, I primarily just care about stability/longevity. Although if this thing is really not breaking a sweat, I'll take any low hanging fruit when it comes to upping the multiplier again.
 
Solution
The overclockers here at Tom's and those on staff who write most of our artices, as well as those on several other well known computer enthusiast websites will tell you that 1.300 is the limit for 22 nanometer processors.

My personal experiences also agree with their findings. As I described in the Temp Guide, pushing Vcore too high for too long will cause BSOD's.

To regain stability, the short term fix is to further increase Vcore, which will again result in BSOD's in even shorter time intervals ... or ... a better alternative is to down clock to regain stability at the same Vcore.

Regardless, BSOD's will eventually recur. At this point both clock speed and Vcore need to be reduced to once again regain stability, however, this is a...


You could compare readings between CPU-Z and core-temp to get a more accurate interpretation.
Don't use HWmonitor.

Decrease Vcore voltage until instability occurs, then slightly bump the vcore back up to a stable level.
 
There is a balancing act between voltage and frequency. BTW 67C is not hot at all for DC(Devil's Canyon) If your cpu is stable at 4.6GHz and 1.2v you are lucky. Enable speed step so that your cpu will only go to 4.6 when needed.Also did you set vcore to be adaptive or always set to 1.2. If you can get adaptive working, it will raise & lower the vcore dependent on frequency. While writing this CPU-Z reports less than 0.8V and 800MHz.;-)
-Bruce
 

deroberg

Honorable
Dec 27, 2013
6
0
10,510
I was look for the stepping and adaptive voltage options in my Gigabyte BIOS, but couldn't find them. I see my multiplier dropping to 8 and voltage dropping to 0.8v at idle, so I guess I am good to go on that front.

Is longevity of the processor mostly dependent on voltage, or is frequency important as well? I don't really have a frequency target, so my thought is to pick a voltage that is going to be safe to run at indefinitely (keeping in mind adaptive voltage is on, and it won't be running at full tilt all the time) and then raise that multiplier as far as the given voltage will keep stable.

Obviously temperature plays into this as well. I will err on the safe side if temperature becomes a significant factor.
 

CompuTronix

Intel Master
Moderator
Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

Section 1 - Introduction

Intel desktop processors have thermal sensors for each Core, plus a sensor for the entire processor, so a Dual Core has 3 sensors. Heat originates within the Cores where Digital sensors measure Core temperatures. A single Analog sensor under the Cores measures overall CPU temperature.

Core temperature is 5C higher than CPU temperature due to sensor location. Intel's Thermal Specification is "Tcase", which is CPU temperature, not Core Temperature. Tcase for the i5 4690K is 72C. Tcase + 5 makes the Core temperature 77C.

The relationship between Core temperature and CPU temperature is not in the Thermal Specifications; it's only found in a few engineering documents. In order to get a clear perspective of processor temperatures, it's important to understand the terminology and specifications, ..."

You should give it a read.

CT :sol:
 

deroberg

Honorable
Dec 27, 2013
6
0
10,510
Awesome information. That is what I needed to see. My ambient temp is closer to 19/20C, so I will factor that in to my adjustments, as well as higher ambient temps for worst case scenarios.

Even though this won't be running all out at 100% duty cycle, would I be able to indefinitely maintain a running processor at ~77C tcore and 1.3v mentioned in the guide?

My goal is to achieve similar longevity Intel gets at stock speed on chips that may not be as good as what I have. If my particular processor can be just as robust at higher speeds, that's how I'm going to run it.

Forgive me, I tend to over think these things.
 
If you run "maxed out" you will degrade the cpu over time. Did you read the part about Electromigration? Fortunately if you get your cpu stable, you are only running this at those levels for relatively short periods compared to the life of your cpu. If you plan on folding 24/7 then dialing back the OC would be a good idea.
-bruce
 

CompuTronix

Intel Master
Moderator
deroberg,

If you keep the core voltage under 1.300 and keep it cool enough to limit the core temperatures to the mid-70's, then your processor should become quite obsolete before it begins to show any indications of stability or performance degradation.

CT :sol:
 

deroberg

Honorable
Dec 27, 2013
6
0
10,510
I read all of it. That part says the max is 1.3v. It just doesn't say how long. Maybe 1.3v will let me get 4 years out of this processor. Maybe 1.25v will give me 10 years.

I'm sure its not that simple though. How does Intel bin processors? Just apply a standard voltage and check temperature? What inputs/outputs do they use to determine if they want to put their name on it and ship it out?

I suppose if I want 99% confidence, I should just cap it at 1.2v and move on. Heh
 

CompuTronix

Intel Master
Moderator
The overclockers here at Tom's and those on staff who write most of our artices, as well as those on several other well known computer enthusiast websites will tell you that 1.300 is the limit for 22 nanometer processors.

My personal experiences also agree with their findings. As I described in the Temp Guide, pushing Vcore too high for too long will cause BSOD's.

To regain stability, the short term fix is to further increase Vcore, which will again result in BSOD's in even shorter time intervals ... or ... a better alternative is to down clock to regain stability at the same Vcore.

Regardless, BSOD's will eventually recur. At this point both clock speed and Vcore need to be reduced to once again regain stability, however, this is a certain indication that the processor has already been permanently compromised.

The most effective long term solution is to respect and observe the voltage and thermal limits that Intel and the most informed and experienced overclockers have made available to us. :sol:
 
Solution