Question DEGRADING Raptor lake CPUs

[Kocicak]

I noticed some reports about degrading i9 13900K and KF processors.

I experienced this problem myself, when I ran it at 6 GHz, light load (3 threads of Cinebench), at acceptable temperature and non extreme voltage. After only few minutes it crashed, and then it could not run even at stock setting without bumping the voltage a bit.

I was thinking about the cause for this and I believe the problem is, that people do not appreciate, how high these frequencies are and that the real comfortable frequency limit of these CPUs is probably at something like 5500 or 5600 MHz. These CPUs are made on a same process (possibly improved somehow) on which Alder lake CPUs were made. See the frequencies 12900KS runs at. The frequency improvement of the new process tweak may not be so high as some people presume.

Those 13900K CPUs are probably highly binned to be able to find those which contain some cores which can reliably run at 5800 MHz. Some of the 13900K probably have little/no OC reserve left and pushing them will cause them to degrade/break.

The conclusion for me is that the best you can do to your 13900K or 13900KF is to disable the 5800 MHz peak, which will allow you to offset the voltage lower, and then set all core maximal frequency to some comfortable level, I guess the maximum level could be 5600 MHz. With lowered voltage this frequency should be gentler to the processor than running it at original 5500 MHz at higher voltage. You can also run it at lower frequencies, allowing for even higher voltage drop, but then the CPU is slowly loosing its sense (unless you want some high efficiency CPU intended for heavy multithread loads).

Running it with some power consumption limit dependent on your cooling solution to keep the CPU at sensible temperature will help too for sure.

 

[Markfw]

I won't comment about the OPs specifics, but one thing is clear to me. Intel over the last few generations are binning closer than ever to try and win benchmarks. They factory overclock to the max. I could certainly see that whatever voltage is required to get 100 mhz over stock could degrade the CPU if done long term, and that could even be hours or days.

the 9600kf and the 1705t were NOT in that category. Any 13900k IS in that category.

 

[DrMrLordX]

Over single core turbo clock speed.

You can't expect every core on the chip to run even at single-core turbo in MT workloads, much less higher than that, without some significant increases in voltage which could be catastrophic at higher current levels. The additional heat would also complicate matters.

 

[Storm-Chaser]

You can't expect every core on the chip to run even at single-core turbo in MT workloads, much less higher than that, without some significant increases in voltage which could be catastrophic at higher current levels. The additional heat would also complicate matters.

Right, I have to admit, all cores running simultaneously at 5.8GHz is not the norm and would make it a challenge. Seems with delidding and lapping and liquid metal people are getting around 6GHz max all core performance. So the OC ceiling on these processors is relatively low, because like you said Intel factory overclocks them to the max.

But if you can manage the heat, the cores will operate normally with no thermal throttling. All I really know is the cores are tested viable for 5.8GHz. If you think 5.9 is going to kill it in two months you are definitely mistaken, even if that was the problem it would take at least another 6 months or more in all probability

OPs problem is OC configuration NOT degradation. He is simply reaching the limits of the CPU and misinterpreting that as degradation. The 13900K is no cupcake to tune/cool once you go above all core turbo.

 

[DrMrLordX]

If you think 5.9 is going to kill it in two months you are definitely mistaken, even if that was the problem it would take at least another 6 months or more in all probability

Sadly we have @Kocicak 's sample which seems to show otherwise. But it is arguable that he did not manage the heat and used too much voltage to compensate.

 

[Kocicak]

3 threads (out of 32) do not produce too much heat. I think the main problem is that I pushed the CPU too far, when I overclocked a possibly weaker core by 500 MHz, added a little bit higher voltage, the CPU could by chance make this particular core work hard, and the CPU may have a general tendency for quicker degradation as well.

Stating that degradation does not exist is not reasonable from any point of view.

It is not just me, Ichirou on overclock.net stated that he had multiple CPUs degraded (he seems to be an experienced overclocker, unlike me), and one member here stated that his 12700K degraded by 50mV per year.

 

[Storm-Chaser]

I've explained my position on matter more than once, if you guys want to insist its degradation that's fine. But it doesn't make it true.

 

[coercitiv]

Sadly we have @Kocicak 's sample which seems to show otherwise.

No, it does not. The way @Kocicak understands to perform overclocking is bound to lead into this kind of situation. He lacks method and knowledge, critical ingredients for stable overclocks and meaningful collection of data. For all we know, his initial OC config was not stable to begin with.

And that all without addressing the elephant in the room, which is the sample size of 1. We had someone on the forum with a dead stock 5800X3D in a matter of days, seemingly at stock voltage. Buildzoid's 7950X died during an overclocking stream. Does that make Zen 3 and Zen 4 prone to failure due to extreme degradation? This is the kind of discourse and FUD we're dealing with in this thread, just throwing failure events at the wall and conjuring explanations with no investigation attached to validate the conclusion.

Bonus: here's a thread about Zen 3 failure rate. Weak silicon everywhere.

 

[Hulk]

Right, I have to admit, all cores running simultaneously at 5.8GHz is not the norm and would make it a challenge. Seems with delidding and lapping and liquid metal people are getting around 6GHz max all core performance. So the OC ceiling on these processors is relatively low, because like you said Intel factory overclocks them to the max.

But if you can manage the heat, the cores will operate normally with no thermal throttling. All I really know is the cores are tested viable for 5.8GHz. If you think 5.9 is going to kill it in two months you are definitely mistaken, even if that was the problem it would take at least another 6 months or more in all probability

OPs problem is OC configuration NOT degradation. He is simply reaching the limits of the CPU and misinterpreting that as degradation. The 13900K is no cupcake to tune/cool once you go above all core turbo.

I thought only the two "favored" cores could boost to 5.8GHz?

Intel® Core™ Processors - View Latest Generation Core Processors

Find the latest Intel® Core™ processor family with specifications, benchmarks, features, Intel® technology, reviews, pricing, and where to buy.

www.intel.com

How Intel Technologies Boost Your CPU's Performance - Intel

Innovations like Adaptive Boost Technology make 11th Gen CPUs faster. Learn how Turbo Boost, Thermal Velocity Boost, and ABT work together.

www.intel.com

 

[Storm-Chaser]

I thought only the two "favored" cores could boost to 5.8GHz?

Intel® Core™ Processors - View Latest Generation Core Processors

Find the latest Intel® Core™ processor family with specifications, benchmarks, features, Intel® technology, reviews, pricing, and where to buy.

www.intel.com

How Intel Technologies Boost Your CPU's Performance - Intel

Innovations like Adaptive Boost Technology make 11th Gen CPUs faster. Learn how Turbo Boost, Thermal Velocity Boost, and ABT work together.

www.intel.com

Exactly what are you trying to say? The article seems to reference the 11900K not the 13900K, so I don't see where you are going with this.

 

[Storm-Chaser]

For all we know, his initial OC config was not stable to begin with.

I think you hit the nail on the head.

 

[DrMrLordX]

No, it does not. The way @Kocicak understands to perform overclocking is bound to lead into this kind of situation.

He put on 1.45v in MT workloads and tried for all-core of what, 6 GHz? You're darn right that was a problem.

He lacks method and knowledge, critical ingredients for stable overclocks and meaningful collection of data. For all we know, his initial OC config was not stable to begin with.

Maybe not, but it's becoming less stable over time, which is the point. Even his maybe-ragtag methodology for assessing stability is picking up problems now that didn't appear before.

3 threads (out of 32) do not produce too much heat. I think the main problem is that I pushed the CPU too far, when I overclocked a possibly weaker core by 500 MHz, added a little bit higher voltage, the CPU could by chance make this particular core work hard, and the CPU may have a general tendency for quicker degradation as well.

There is a difference in taking a 1T workload, pegging it to a weak core, and using extra voltage to make it stable at a ~500 MHz overclock and making every core clocked to 5.8GHz-6 GHz (or wherever you had it) during a current-heavy MT workload. Even your "good" cores would not enjoy that experience, unless you have cooling to compensate.

 

[Storm-Chaser]

Maybe not, but it's becoming less stable over time, which is the point. Even his maybe-ragtag methodology for assessing stability is picking up problems now that didn't appear before.
There is a difference in taking a 1T workload, pegging it to a weak core, and using extra voltage to make it stable at a ~500 MHz overclock and making every core clocked to 5.8GHz-6 GHz (or wherever you had it) during a current-heavy MT workload. Even your "good" cores would not enjoy that experience, unless you have cooling to compensate.

You are defending a flawed supposition where OP conclusion is based purely off speculation. Less stable over time? His OC config was probably wrong both times! it's next to impossible to get more than 1-200MHz out of a 13900K without extreme cooling like at minimum, a chiller. And he doesn't have that, does he?

using extra voltage to make it stable at a ~500 MHz overclock

I thought he was going from 5.8 to 5.9GHz, no? Where are you getting this 500MHz number?

 

[Kocicak]

Jesus, I wrote it like 5 times already, it was THREE threads load. THREE out of 32, that is VERY LIGHT LOAD. The CPU was free to shuffle the load around on the 8 P cores as it wished.

When you have 6 cores limited to 5.5 and 2 to 5.8 GHz, and you OC all to 6, you OC some by 500 and some by 200 MHz!

There is a difference in taking a 1T workload, pegging it to a weak core, and using extra voltage to make it stable at a ~500 MHz overclock and making every core clocked to 5.8GHz-6 GHz (or wherever you had it) during a current-heavy MT workload. Even your "good" cores would not enjoy that experience, unless you have cooling to compensate.

I had an air cooler on it, I tested only what that cooler can cool - that is the reason I ran so low number of threads. I hoped that the CPU will randomly assign the load to all cores and all cores will get tested with only 3 threads.

If I remember correctly, 3 cores pull about 80W, that is a half of what the cooler can handle.

I did not overheat anything.

 

[coercitiv]

He put on 1.45v in MT workloads and tried for all-core of what, 6 GHz? You're darn right that was a problem.

3 cores, he put stress on 3 cores.

 

[Storm-Chaser]

3 cores, he put stress on 3 cores.

WOW this is computer comedy at it's finest. LOL

When you have 6 cores limited to 5.5 and 2 to 5.8 GHz, and you OC all to 6, you OC some by 500 and some by 200 MHz!

don't mean to be harsh, but I think you failed your overclocking exam.

 

[Hulk]

Exactly what are you trying to say? The article seems to reference the 11900K not the 13900K, so I don't see where you are going with this.

I didn't know that any of the cores are capable to boost to 5.8GHz. Unless Turbo Boost 3.0 has changed Intel described it as boosting two favored cores and then directing threads TO them. Not moving threads around to boosted cores. From the second link in my original post.

"Due to production differences, processor cores vary in maximum potential frequency. Intel® Turbo Boost Max Technology 3.0 identifies up to two of the fastest cores on your CPU, known as “favored cores.” Then it applies a frequency boost to those cores (or that core) and directs critical workloads to them."

Intel seems pretty clear that this is only for two of the favored cores. I was just wondering where you read any of the cores can use Turbo Boost Max Technology 3.0?

The link wasn't specific to 11th generation but to Intel's Turbo Max Technology 3.0?

So where did you read that any of the cores is capable of 5.8GHz?
I've noticed on my 13900K that only two will hit 5.8GHz.

 

[DrMrLordX]

3 threads (out of 32) do not produce too much heat.

That must have been lost in the noise then. Didn't you run MT workloads @ 1.45v at any point?

You are defending a flawed supposition

I've damaged CPUs with overvoltage before, notably a memory controller on an old a64 x2. Intel memory controllers were also subject to damage from excessive vDIMM in some generations. It happens.

I thought he was going from 5.8 to 5.9GHz, no? Where are you getting this 500MHz number?

His weaker cores were not capable of 5.8 GHz, but were rather capable of ~5.4 GHz or whatever was the maximum speed he could get out of them (probably no TVB kicking in).

edit: removed broken formatting

 

[Kocicak]

That must have been lost in the noise then. Didn't you run MT workloads @ 1.45v at any point?

No, there is no point in doing that, in MT you are limited by power draw, you cannot run the CPU at so high frequency, which would require such voltage. I used that voltage only to test what is the highest frequency the two best cores can run at. And for that 3 thread run.

 

[Kocicak]

I didn't know that any of the cores are capable to boost to 5.8GHz.
...
So where did you read that any of the cores is capable of 5.8GHz?
I've noticed on my 13900K that only two will hit 5.8GHz.

I believe that 5.5 GHz limit is there mainly to limit power consumption in certain load intensity. If they got the limit to e.g. 5.7, the CPU may draw too much power during gaming for example, which would be tested in reviews and Intel would look even worse than they look now (power draw wise).

Those two 5.8 cores are probably selected somehow, but many of the other cores are capable of the same frequency or even more.

Unfortunatelly nobody knows (except Intel), what which core is capable of and what it can be safely subjected to during testing.

I guess most of the 5.5 cores can run at 5.7 GHz without much trouble.

 

[Storm-Chaser]

I didn't know that any of the cores are capable to boost to 5.8GHz. Unless Turbo Boost 3.0 has changed Intel described it as boosting two favored cores and then directing threads TO them. Not moving threads around to boosted cores. From the second link in my original post.

"Due to production differences, processor cores vary in maximum potential frequency. Intel® Turbo Boost Max Technology 3.0 identifies up to two of the fastest cores on your CPU, known as “favored cores.” Then it applies a frequency boost to those cores (or that core) and directs critical workloads to them."

Intel seems pretty clear that this is only for two of the favored cores. I was just wondering where you read any of the cores can use Turbo Boost Max Technology 3.0?

The link wasn't specific to 11th generation but to Intel's Turbo Max Technology 3.0?

So where did you read that any of the cores is capable of 5.8GHz?
I've noticed on my 13900K that only two will hit 5.8GHz.

My bad, I didnt read it. I just assumed it was like my Broadwell Xeon E5 2696 v4 lol which I've been really studying turbo core behavior on --- but in hindsight the chips couldn't be much more different.

guess I should have done some homework first. So all core turbo is 5.5, right? So if he goes to 6GHz, thats 500MHz OC.

I would like to get a HWINFO64 read out for a 13900K to study just how much variation in clock speed

But back to the main point, I do not believe that he degraded the processor in that short a timeframe.

 

[igor_kavinski]

I used that voltage only to test what is the highest frequency the two best cores can run at. And for that 3 thread run.

If any of those cores could hit 6 GHz long term, you wouldn't have got such a CPU in the first place because Intel would be saving them for the KS SKU. Just admit defeat and stop torturing CPUs for your pleasure. Accept 5.8 GHz as the limit. You really think Intel would let slip super quality silicon into relatively cheaper K SKUs? Even if your CPU degraded due to overvolting, you did it yourself. Intel didn't tell you to do that or tell you officially that these CPUs are capable of flawless operation beyond 5.8 GHz. I just have a hard time understanding your thought process.

 

[Storm-Chaser]

That must have been lost in the noise then. Didn't you run MT workloads @ 1.45v at any point?



I've damaged CPUs with overvoltage before, notably a memory controller on an old a64 x2. Intel memory controllers were also subject to damage from excessive vDIMM in some generations. It happens.



His weaker cores were not capable of 5.8 GHz, but were rather capable of ~5.4 GHz or whatever was the maximum speed he could get out of them (probably no TVB kicking in).

[/QUOTE]
[/QUOTE]
I understand it can happen. Not out of the question. Just has not happened to me as of yet and I've pushed some chips pretty hard over the years.

 

[Kocicak]

If any of those cores could hit 6 GHz long term, you wouldn't have got such a CPU in the first place because Intel would be saving them for the KS SKU. Just admit defeat and stop torturing CPUs for your pleasure. Accept 5.8 GHz as the limit. You really think Intel would let slip super quality silicon into relatively cheaper K SKUs? Even if your CPU degraded due to overvolting, you did it yourself. Intel didn't tell you to do that or tell you officially that these CPUs are capable of flawless operation beyond 5.8 GHz. I just have a hard time understanding your thought process.

The resiliency for heavy 24/7 load is something different than for reasonable occasional load. Even a CPU which would not be suitable for the first use case could happily serve in the second.

5.8 or 5.5 GHz cannot be the limit by definition, because this is unlocked special version of the chip INTENDED FOR OVERCLOCKING. By definition it must have some headspace to be oveclocked to.

If these CPUs cannot be overclocked at all, they cannot be sold as special CPUs for overclocking. Hope this is clear...

So far I came to conclusion that these CPUs can be overclocked by 200 MHz in all frequencies (two best P cores to 6 GHz, the rest of P cores to 5.7 GHz and E cores to 4.5 GHz).

As I said two of the three CPUs I had/have could pass 1T CNB run at 6200 MHz on the two best cores. That is already 400 MHz more than stock.

The signs of degradation so far may not be just a result of a mistake, but something else, and that is why I proposed earlier that somebody should test it properly so that people have some real info to work with if they want to overclock these CPUs (what voltages, temperatures and frequencies to avoid) and on the other hand if they want run the CPUs more efficiently (then they need to know by what amount they can lower the voltage and when they should retest the CPU if it does not need more, etc).

 

[DrMrLordX]

No, there is no point in doing that, in MT you are limited by power draw, you cannot run the CPU at so high frequency, which would require such voltage. I used that voltage only to test what is the highest frequency the two best cores can run at. And for that 3 thread run.

Let's lay some facts out here.

Yes, at lower current levels, some higher voltages are "safe" in that they may not be fatal to the chip or cause any degradation. My 3900x (for example) can push voltages as high as 1.5v for ST workloads, but it must retreat to at most ~1.325v (according to FIT tables extracted by The_Stilt; these table values actually can vary from one sample to the next) for intense MT workloads. Part of the reason for this is to stay within power constraints, which in the case of a 13900k can be pretty generous. After all, some motherboards stock will push you to 350W if the cooling is there.

If you set manual voltages, you may wind up end-running the power limitations set by the UEFI (PL2 in particular), though I'm not 100% sure about that. It may well be that PL values will constrain how much current you can draw if you use a manual voltage setting, not that I would encourage you to test for such a thing. On older platforms - AMD or Intel - setting manual voltage would blow past any power-based restrictions and let the chip draw however much power was necessary to handle a particular workload while maintaining the indicated voltage.

In other words, I'm pretty sure that if you manually set 1.45v (or whatever) and tried to hit 5.8GHz on all 8 P-cores, that you would disable power restrictions and bring some serious heat. At that point we could talk about degradation. Maybe.

 

[igor_kavinski]

If these CPUs cannot be overclocked at all, they cannot be sold as special CPUs for overclocking. Hope this is clear...

True. But Intel guarantees nothing.

You may find this interesting:

His P-cores are all running at 6200 MHz but clock down to 6100 when the load increases. For lightly threaded loads like game threads, this is working. But he is also using a crazy water block. You, on the other hand, insist on pushing the cores to 100% above 5800 MHz. What you are doing is more torturous. I hope I do not need to point out that even a seasoned overclocker isn't doing what you are trying to do.