cancel
Showing results for 
Search instead for 
Did you mean: 

7700k degradation?

A_radek
Level 7
Short story:

2 years ago I oc’d a 7700k to 4.8ghz @1.25v. It passed all tests I could throw at it. Today - two years later - it needs 1.33v to pass the same tests. Cpu Degradation?

Wall of text:

Roughly two years ago I bought a 7700k and a prime z270-a motherboard. The sole intention for this was to run a flightsim that is mostly singlethreaded, requires ~4.9ghz cpu for a pleasant VR experience and on top of this uses avx instructions. So the plan was to OC that cpu right from the start.

A Noctua d15 (twin fan&tower) together with 5x 120mm noctua chassis fans were installed. Not as powerfull as a 360 water radiator but good enough. I then followed this article:

https://edgeup.asus.com/2017/kaby-lake-overclocking-guide/ *

Settled for the recommended 4.8ghz oc @1.25v and it all worked like a charm. Passed every test I could throw at it including a few hours of avx2 enabled prime 95. Rig has been working flawlessly ever since with temps maxing out in the low 60’s during gaming.

Recently figured I’d try and see how far this cpu will go as I never really did that. But first stress test and monitor current temps.. to my surprise two cores instantly failed in prime 95 and OCCT avx2 instructions start spawning errors immediately. These are the tests my oc passed with flying colors two years ago!

So after checking fans, reseating the Noctua chiller and meticulously dusting everything to no avail, I increased voltage and had to go all the way up to 1.33 for that same prime 95 stable 4.8ghz. Of course temps go through the roof at that voltage peaking at 85c on the hottest cores, which is a tad much for my taste. So down to 4.7ghz I go @1.25v.

*Really? This is the fabled cpu degradation? The dark side of overclocking people eventually end up at? What about those 5ghz 1.4v+ overclocks? Are those valid for 5-6 months only? Or did I just draw a really short straw?

*Not like 4.8 to 4.7 is much of a difference in performance but disappointed nonetheless as I thought I was playing it safe with perhaps a bit more on tap. I realize none can give me a solution but are others experiencing similar performance degradation?
2,727 Views
2 REPLIES 2

Falkentyne
Level 12
That isn't degradation. That's simply the CPU wearing down with time as all chips do. Each and every chip will lose its max overclock over time. You are not going to have the same clock 2 years later that you had when the chip was new. Think for a minute. If your chip lasted exactly the same now as it was new, why doesn't Intel have a 10 year warranty like Seasonic does with their Prime Platinum Power Supplies? There's a reason the warranty is only 3 years. The chip must be guaranteed to still run at *STOCK SPEEDS* in 3 years. All chips will lose their max overclocks over time 🙂
If your chip lasted the same overclock 2 years from when it was new, then it should also last 10 years at that overclock. And they won't and don't.

Degradation is going from 1.25v to 1.33v on a 24/7 low overclock in one month from running high voltage and high LLC repeatedly during stress tests with higher overclocks (like example 1.40v + LLC6). THAT Is degradation.

Thank you for taking the time to answer that and your point makes sense.

Lesson learnt: Buy a new one every ~three years if you want it to keep performing ~15% better than it was intended to 🙂 As soon as comet lake-s hits stores this chip will be retired. Let’s hope it can hold 4.7 for a few more months then.