~1.5V safe for 2600k or stop at ~1.4V for 24/7?

[tackle70]

Hi all,
So I've had a 2600k for a couple years now with an ASRock z68 extreme7 gen3 board and an NH-D14 cooler, and it's been clocked at 4.8 GHz on 1.4V during that time.

I've been thinking of bumping up to 5-5.1 GHz, but based on some initial testing, it looks like that's going to require a voltage jump up to the ~1.5V range, and I don't know if that's a good idea or not.

The only CPU-intensive task I do with my rig is gaming, which means temperatures aren't too bad... my max temps at 1.41V/4.8 GHz are about 74-75C.

Related question - would a Corsair H110 be worth an upgrade over my NH-D14?

Thanks!

 

[Remmy]

It does sound very high voltage. Increasing the voltage is primarily what makes the temperature increase, but past a certain voltage the temperature increase stops being moderate, and suddenly starts going up really quickly, so by the time you're getting up to 1.5V from 1.4V @ 75°C, it's probably going to be something stupid like 90°C+

 

[tackle70]

Right... I'll do some temp testing and see what it does.

If memory serves my temps jumped up to the low 80s while playing BF3.

I don't do the standard stability testing at these volts because I know my NH-D14 can't cope with prime95/OCCT/etc above about 1.4V.

 

[Remmy]

Yeah, and even ignoring the temps, with a proper stress test it may just fail anyway. I've got no personal experience with Corsair AIOs, but I hear very good things about the H110, and lots of people get really good overclocks with them.

 

[tinytomlogan]

TBH whats the point in making it faster if you just game.......?

 

[Remmy]

I agree with that to an extent Tom, but there's something about the 5GHz club that is appealing, if you can reach it safely. I had mine at 4.9GHz stable but hot, and couldn't ever get 5GHz as it just BSOD'd straight away :/ In the end I just had it around 4.5-4.7GHz

 

[WYP]

i don't think your d14 has the thermal headroom.
from a quick google 1.5v seems to be outside most peoples comfort zones. i don't know personally if it would degrade your cpu or not, i'd feel uncomfortable going that high though.

if your only gaming the jump to 5Ghz will have little performance benefit TBH.
4% clock bump over 4.8Ghz, isn't noticeable really

 

[tackle70]

TBH whats the point in making it faster if you just game.......?

I've just got the itch to play around with my hardware and this is a way to do it without spending any money

I know it won't affect anything I do with regards to performance.

-edit- I also think my D14 just isn't going to have the headroom... 80C max temp after a couple 3dmark physics runs. Gonna try to play some BF3 real quick and see if it melts or not.

 

[tinytomlogan]

Also asking about coolers over the D14........ This is a review site mate.... durrrrr

 

[tackle70]

I know, I've read the reviews. I know the H110 is better, but I don't know if it's worth the money over the NH-D14. Obviously if I were buying today instead of 2 years ago I'd have gone for the H110, but I'm not sure if a few degrees C is worth ~$100.

20 minutes of BF3 yielded a max temp of 83C on the hottest core, which is a bit high but not awful... better than I was expecting, actually. I expected more heat problems than just +8C from an extra 100 mV through the chip.

 

[Remmy]

It may well be higher on proper stress testing bear in mind though

 

[tackle70]

Oh yeah... on proper stress testing, 1.4V/4.8 GHz is the absolute max I am comfortable with, which is why I have been running that clock as my 24/7 clock for two years.

A 12 hour run of small FFTs will result in max temps of 85-88C on the hottest core at that setting, so I am fairly certain a 1.5V test would eventually fail unless I went with full out water cooling (which has never really interested me).

Just passed a half hour of the firestrike physics test looped at 1.47V/5 GHz/max temp 82C.

I'm getting the sense from you guys that this voltage is a bit disconcerting for air cooling, though... maybe I'll just save the BIOS settings after I get it (game) stable and then revert back to my normal 4.8

Thanks for the thoughts guys!

 

[tinytomlogan]

OCCT - Linpack mode - AVX enabled - 90% free memory.


THATS how to stress test and look at temps

 

[tackle70]

OCCT - Linpack mode - AVX enabled - 90% free memory.


THATS how to stress test and look at temps

hahahahaha wow, my D14 didn't like that at all... hit 90C on the hottest core almost instantly even at 1.4V/4.8 GHz

Only one out of the 4 cores survived under 90C within the first two minutes (and it was at 75C, so I dunno if I need to re-mount my cooler or not)

And here all along I had thought that prime95 was the worst case scenario for temperatures... guess I was wrong about that :lol:

 

[tinytomlogan]

thats how the pro's do it

 

[tackle70]

thats how the pro's do it

Well, doing it that way convinced me to upgrade my cooling, so thanks TTL for the tip!

Swapped out my NH-D14 for a Corsair H110 and my temps just dropped through the floor. At 1.41V (4.8 GHz), I've had AVX capable linpack running for a few hours now, and the maximum temp on any core was 81C.

Not sure how to account for such a huge (10C+) difference in temps between my NH-D14 and the H110... either my graphics card waste heat was affecting my NH-D14's cooling capability more than I thought, or I must have had the cooler mounted incorrectly (though I had mounted/re-mounted it a bunch of times to ensure correct operation).

Anyways, I'm now sold on this H110... best $100 I've spent in a while. My gaming temps at 4.8 GHz playing BF3/BF4 now sit in the low-mid 60s, AND I was able to turn fans down so that the whole rig is whisper quiet.

I can't get 5 GHz stable at volts/temps I'm comfortable with, sadly. I can get it stable for games but not for prime95/OCCT, so I'm just going to leave it at 4.8 GHz and keep the cool temps and quiet fans.