do these temps make sense?

kty1987

Reputable
Dec 6, 2014
4
0
4,510
hi guys, fairly new to overclocking here

so i have an asrock extreme 6 z97 board with a 4690k that i just bought
for cooling i got a noctua nh-d14

in an old antec 900 case with a radeon r9 290 gpu

so i thought i'd give overclocking a try

i looked up the results other people have gotten thus far as a rough guideline

right now i am at 88degrees (read off realtemp 3.70) on pretty much all 4 cores
running at 44x (4398.78MHz) with a vcore of 1.175

does the temperature make any sense? i expected the cooler to cool much better than that.. especially since my voltage isn't anywhere near that high

thanks for replies guys,

yes i am using small FFTs for prime95 to test

i was using the latest version, however.

once i switched to 26.6, my temps dropped to very acceptable ranges (60s-70s)
i don't understand though, does that mean there's something inherently wrong with the latest version of prime95? why does it make the temps shoot up, and why is it not considered a good measurement of how hot your chip can get?
 
Solution
What version of Prime95 are you using? If you're using the most recent version (or any version past 26.6 for that matter), there's an AVX bug in the newest versions that causes Haswell CPUs to get insanely hot (as much as 20 C hotter than they should be under 100% load). Try an older version of Prime 95 or use a different program like Aida64 or OCCT.

barto

Expert
Ambassador
That's definitely very warm. How much did you raise the vcore?

Overclocking is a trial by error process. You work your way up to 4.4 and not jump right to it. To overclocking, start increasing the frequency until you find an unstable clock. Then increase the vcore a little to make the unstable clock stable. Continue increasing clock again and repeat the process. Because of the ease with overclocking with the multiplier these days, just find the unstable clock first.

Also, make sure that your CPU cooler fans are facing the correct way (up or out the back) and did you apply thermal paste?
 

kty1987

Reputable
Dec 6, 2014
4
0
4,510
i read a thread on the forums that said to start by going to 4.6 and 1.2 vcore and see if it will boot, as it is about the 50 percentile ish range

if it does then tweak it from there

im not too concerned with the multiplier atm though, was just curious since i thought temperature is decided by voltage, and i am not running that much juice through the chip yet...
 

kty1987

Reputable
Dec 6, 2014
4
0
4,510
hi thanks for the responses,

my ambient is approx 12-14 degrees Celsius. cold winder night with no heater on
i am using prime95, the option with maximum temp stress test

i'll check that thread out too, thank you ^^
 

Ellis_D

Reputable
Jul 20, 2014
209
0
4,710
What version of Prime95 are you using? If you're using the most recent version (or any version past 26.6 for that matter), there's an AVX bug in the newest versions that causes Haswell CPUs to get insanely hot (as much as 20 C hotter than they should be under 100% load). Try an older version of Prime 95 or use a different program like Aida64 or OCCT.
 
Solution

CompuTronix

Intel Master
Moderator
To be thorough, here it is from the Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

" ... Section 12 - Thermal Testing @ 100% Workload

Prime95 Small FFT's is the standard for CPU thermal testing, because it's a steady-state 100% workload. This is the test that Real Temp uses to test sensors. The link above is to version 26.6, which is well suited to all Core 2 and Core i variants.

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 high temperatures. The FPU test in the software utility AIDA64 shows the same results. ...

... 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. ... "

Prime95 version 26.6 - http://windows-downloads-center.blogspot.com/2011/04/prime95-266.html

Also, while Prime95 Large FFT's is described as "maximum heat, power consumption" this statement is somewhat misleading. It actually refers to overall heat created by the processor and memory modules, not just heat created by the processor itself.

Large FFT's is a fluctuating workload with fluctuating temperatures.
Small FFT's is a steady-state workload with steady temperatures.

The proper test to use for testing your processor temperatures is Small FFT's.

CT :sol:
 

kty1987

Reputable
Dec 6, 2014
4
0
4,510


thanks to everyone who responded! i think this is the easiest to understand solution.

I was running latest version of prime95 and that damn near brought my chip to 100 degrees in a split second
after swapping to 26.6, my temps are at 70 degrees stable at 1.25v after OC

in short, use older version of prime95 as a synthetic test since the new versions use a new instructions set that could overheat the chips to unrealistic levels that you would never see in real world 100% load, therefore not necessary as a stability test :D
 

CompuTronix

Intel Master
Moderator
kty1987 and Ellis_D,

There is no "bug" in later versions of prime95.

Understand that the author of Prime95 is on a quest to find the next prime number ... so nothing else matters. As Intel and AMD processors have progressed, he has taken advantage of their latest instruction sets (AVX, AVX2, FMA3 since Core i 2nd Generation Sandy Bridge) by writing code into Prime95 that greatly accelerates mathematical calculations, which are performed on the processor's Floating Point Unit (FPU).

This creates exclusive FPU workloads that cause the processor to greatly exceed the steady-state 100% workload thermals which we have used as a standard for many years. If you download and run AIDA64's FPU test, you will get the exact same thermal results.

AVX code is used to accelerate high-level number crunching, which comes in handy for processors running CAD, encoding, encryption, banking servers and scientific calculations. However, the vast majority of us here are PC gamers who may also run apps which don't rely on AVX code.

The bottom line is that if you're an overclocker who runs any software that might use AVX, AVX2 or FMA3, then you may need to downclock and reduce Vcore to accommodate operational temperatures that don't exceed recommended norms.

What does all this mean? You can read about it in this Tom's Sticky: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

CT :sol: