Intel Statement on Stability Issues: "Motherboard Makers to Blame"

[chrcoluk]

Intel wants the baseline profile to become the default profile by the end of May, article from benchlife.info:

Intel Baseline Profile can't solve the problem of 13th and 14th Gen processors, but it is a bigger problem for motherboard manufacturers and users - BenchLife.info

Intel's Baseline Profile can't solve the problem of 13th and 14th Gen processors, but it is a bigger problem for motherboard manufacturers and consumers​

By Chris.L on 2024-05-05 in Hardware Components, Processors

Is it the problem of the motherboard factory, or the problem of the Intel processor itself, in fact...

The problems with Intel 13th and 14th Gen desktop processors on the Raptor Lake-S and Raptor Lake-S Refresh platforms have a tendency to get bigger and bigger, and at this stage, we only feel that Intel has thrown all the problems to the motherboard factory, and has not suspected that it is the problem of its own processors.

Thinking about it from another angle, the 13th and 14th belong to the same architecture, and if there is a real problem, it should have erupted a long time ago, rather than a large-scale problem that has only occurred now, because it is not unreasonable for Intel to think that the BIOS settings of the motherboard factory are too aggressive.

However, Intel's approach to using the Baseline Profile is actually incomprehensible to motherboard manufacturers, and the Intel Baseline Profile will be renamed to Intel Default Settings in the future, but this is another story, and it doesn't have any impact, it's just a name anyway.


Intel Default Settings is tentatively scheduled to go live on May 31, 2024.

At this stage, Intel Baseline Profile is doing to reduce the PL value. Taking the PL2 = 253W of the Intel Corei9-14900K, you can make the Turbo Boost reach the operating conditions of 6.0GHz, but after applying the Intel Baseline Profile, the PL2 of the Intel Core i9-14900K will be set to 188W The Baseline Profile is all about making the Intel Core i9-14900K's PL2 = 188W.

What's funnier is that Intel's suggestion includes the phrase "For best performance use PL1 = 253W". This means that PL1 = PL2 = 253W, but unfortunately the Intel Baseline Profile will completely limit these settings.

PBP (Processor Base Power) is still at 125W, which has not changed, and we know that the CPU will not run at 125W at all.

On the other hand, according to the Performance setting of Intel specifications, the maximum values of Iccmax and Iccmax.app are 307A and 245A respectively, except for the PL2 = 253W of Intel Corei9-14900K, but after applying the Intel Baseline Profile, the maximum values are reduced to 249A and 200A, which is quite a limit to Intel Core i9-14900K Overall processor performance.

	Baseline	Performance	Extreme
Processor Base Power	125W	125W	125W
Iccmax	249A	307A	400A
Iccmax.app	200A	245A	320A
PL1	125W	125W	253W
PL2	188W	253W	253W
PL4	293W	380W	380W
iPL2	160A	200A	200A

As mentioned earlier, Intel Z790 chip motherboards have high-end, mid-range and entry-level motherboards, which are very different, but in the mid-range and high-end parts, the motherboard manufacturers use higher designs in the design of CPU power supply modules than recommended by Intel, and this time the 13th and 14th processors are not only downgraded to the same level as the non-"K" series processors due to the Intel Baseline Profile, but also make the motherboard manufacturers' motherboards useless.

After all, in addition to Performance and Baseline, the plan given by Intel also has an Extreme setting. In the Extreme Profile, there will be the PL1 = PL2 = 253W setting we mentioned earlier, which is also designed for high-end motherboards, after all, the "K" series processors can get better performance with high-end motherboards, which is one of the things that consumers expect.

The Iccmax and Iccmax.app in the Extreme Profile are 400A and 320A, respectively.

After the Intel Baseline Profile is applied, the Intel Thermal Velocity Boost Frequency time of the Intel Core i9-14900K, that is, the time that can reach 6.0GHz, will decrease, and the overall will be lower than 3s, assuming that we look at the number 3s, the Turbo Boost of the Intel Core i9-14900K It's like waste, useless.

To paraphrase Taiwanese folks, Intel Baseline Profile turns the Intel Core i9-14900K into a three-second man!

The 13th vs. 14th Intel Baseline Profile didn't actually prevent the problem from continuing to happen, but rather masked it.

In the course of our conversations with some motherboard manufacturers, we also learned that each motherboard manufacturer actually introduced Extreme Profile from the 9th Gen (Coffee Lake) platform, and all of them were unharmed until the 13th and 14th Core of the Raaptor Lake generation.

In fact, we also know that the settings of each motherboard manufacturer for high-end motherboards are higher than those of Intel Extreme Profile.

At the same time, there are also motherboard factories for Intel's instructions are not clear and distressed (the baby is bitter, but the baby can only do it in his heart), after all, it is not only Intel who faces consumers on the front line, but employees in various regions of the motherboard, and these practitioners have no way to solve the problems faced by consumers in the first time.

As for whether the problem is whether Intel itself has relaxed the acceptance criteria for commercially available processors, this can only be said to be speculation, after all, most of the processors in the hands of motherboard manufacturers are not commercially available, and most of the problems encountered are commercial versions, so that motherboard manufacturers are quite powerless when encountering feedback.

Fortunately, there is no such thing as 15th Gen in the future, and the new Intel Core Ultra 2 series (Arrow Lake-S) should allow Intel to change its life.

Intel is expected to issue an official statement before the end of May, and then see what different the x86 processor leader that has recently focused on Foundry will say.



I am not sure what a "three-second man" means, but the CPU at 125W certainly runs and with a heavy load runs very efficiently. I am not sure, if Intel in the beginning did not make a mistake (because they have not planned to use that limit anyway), but 125W is really a very low power for a 24 core CPU!

So if I understand this right, motherboard vendors saying all these misconfigured bios is not their fault, its intel telling them to do it.

Now intel wants them to put it as default, they have a problem with it as it makes their premium boards perform like the cheap boards out of the box, umm wasnt I trying to say this for the past couple of weeks, the board vendors want their products to outshine each other hence all of the tinkering and things like pre tuned optimised defaults.

 

[AnonymousGuy767]

So if I understand this right, motherboard vendors saying all these misconfigured bios is not their fault, its intel telling them to do it.

Now intel wants them to put it as default, they have a problem with it as it makes their premium boards perform like the cheap boards out of the box, umm wasnt I trying to say this for the past couple of weeks, the board vendors want their products to outshine each other hence all of the tinkering and things like pre tuned optimised defaults.

The board vendors want to be able to outshine themselves. There's no point in buying a higher end Z790 board (where they make profit) if your 14900K is going to only be a 14900K for 3 seconds before getting slammed down to i7 levels. I can guarantee you that the board vendors are really pissed about this whole situation, because Intel's the ones not delivering them a competitive product to run with and now throwing them under the bus like it's their fault that they tried to make these things not thermal throttle while hitting the advertised clock speeds. And board vendors already hated Intel for being years behind schedule on product releases.

 

[btk2k2]

So if I understand this right, motherboard vendors saying all these misconfigured bios is not their fault, its intel telling them to do it.

Now intel wants them to put it as default, they have a problem with it as it makes their premium boards perform like the cheap boards out of the box, umm wasnt I trying to say this for the past couple of weeks, the board vendors want their products to outshine each other hence all of the tinkering and things like pre tuned optimised defaults.

They stick to AMDs spec and AMD are still a smaller company so you would have thought Intel has more leverage over their motherboard partners.

This does feel like people trying to pin Intel's slack oversight of motherboard defaults onto the mobo makers rather than Intel's laissez-faire attitude.

 

[BoggledBeagle]

Power limits, maximal currents and LLC settings have been discussed, but what about THE EXTREME FREQUENCIES ? Not only are the CPUs running at those high frequencies power hungry and inefficient, but are not these extreme frequencies a big part of the problem? Or even a most important cause of any instabilities, cause of high voltages, leading to high currents, leading to high temps, leading to degradation?

I believe that 14900K commonly runs P cores at 5700 MHz, if temperatures allows that, would not dropping the frequencies alone solve most of the problems? Dropping 5700 to say 5200 takes away 9% of the performance, solves instability issues, probably prolongs the life of the CPU by years, increases efficiency, what else? Is not it worth it?

 

[chrcoluk]

They stick to AMDs spec and AMD are still a smaller company so you would have thought Intel has more leverage over their motherboard partners.

This does feel like people trying to pin Intel's slack oversight of motherboard defaults onto the mobo makers rather than Intel's laissez-faire attitude.

Previous testing has shown they dont stick to AMD spec, I even see it on my b450 pro 4.

The board vendors want to be able to outshine themselves. There's no point in buying a higher end Z790 board (where they make profit) if your 14900K is going to only be a 14900K for 3 seconds before getting slammed down to i7 levels. I can guarantee you that the board vendors are really pissed about this whole situation, because Intel's the ones not delivering them a competitive product to run with and now throwing them under the bus like it's their fault that they tried to make these things not thermal throttle while hitting the advertised clock speeds. And board vendors already hated Intel for being years behind schedule on product releases.

hey tried to make these things not thermal throttle while hitting the advertised clock speeds.

Maybe so, but the chips are designed to thermal throttle like that, and turbo clocks have never been a guaranteed speed, thats on them for trying to make the chips into something they not.

I see no issue with them having one touch Asus/Asrock/Gigabyte/MSI whatever pre tuned so the user can activate, but I think it should need end user intervention to activate it.

They could also have a dual switch bios on the board, with the second one using their own defaults, but the boards should be shipped on first switch. there is options open to them.

 

[AnonymousGuy767]

Maybe so, but the chips are designed to thermal throttle like that, and turbo clocks have never been a guaranteed speed, thats on them for trying to make the chips into something they not.

The turbo frequencies are part of the specs though, and thus should be guaranteed. It's obviously not guaranteed under all conditions but the general idea is: if you have power and thermal headroom the chip should run at these speed.

We need to be very clear what the problem is: we're not talking about thermal throttling or inefficiencies or motherboard vendors goosing limits (these were all talked about last year): we're talking about these chips *point blank* not being capable of operating at their advertised speeds without degrading very quickly. And what Intel is doing is trying to say "well we never really advertise those speeds lasting for more than 3 seconds". Intel isn't entitled to silently redefining the last 10 years of how turbo limits have actually been implemented.

And don't let that distract you from the fact they're only saying any of this because they know that the chip is incapable of sustaining the advertised clock speeds without degrading. They're trying to dance a line and invent excuses so they don't get a class action lawsuit over their marketing running past their engineering.

 

[Crackong]

So the actual actual 'baseline' profile is PL2 = 188W now? official?

 

[dir_d]

The baseline is already documented in the whitepaper. The Intel-recommended values are stated in section 4.4, page 98 of the data sheet.

13th Generation Intel® Core™ and Intel® Core™ 14th Generation Processors Datasheet, Volume 1 of 2

Isso descreve todas as especificações divulgadas publicamente, incluindo características elétricas, mecânicas e funcionalidades de componentes, uma lista de principais recursos, uma descrição funcional e uma visão geral arquitetônica.

www.intel.com.br

For example, the i9 KS chips:

S-Processor 8+16 150 W: PL1 253 W, PL2 253 W
S-Processor 8+16 150 W, Extreme Config: PL1 320 W, PL2 320 W

Or the i5 chips:

S-Processor 6+8 125 W: PL1 125 W, PL2 181 W

The data sheet is concise and complete regarding Tau length, recommended current and wattage for all models and specifications, you just need to know how to correlate the subtype with the marketed name. S-Processor 150 W means i9 KS, S-Processor 125 W means i9 K, etc.

Again, it must be stressed that 13th and 14th Generation Core as well as the Xeon E-2400 series CPUs have the exact same denomination and stepping: "Raptor Lake-S", and they do not have any differences whatsoever between them. They have the exact same stepping and hardware revision, if you compare a i9-13900K and a i9-14900KS the sole difference between them is their clock table and silicon quality, functionally and at a technical level, they are the exact same processor unchanged.



There will be no further "generations" to the Core i processor line. The next will be Core Ultra series 2, and it should be radically different compared to the existing Raptor Lake chips.

This was like 6 pages back replying to me about waiting to here from Intel about Baselines. It seems you were wrong, and we needed to wait. Now that Intel has said what is baseline, no one but Gigabyte got it right.

 

[Dr. Dro]

This was like 6 pages back replying to me about waiting to here from Intel about Baselines. It seems you were wrong, and we needed to wait. Now that Intel has said what is baseline, no one but Gigabyte got it right.

It's more like Intel pulled some revised numbers out of their behind because the new values they're pushing completely mismatch with the documentation. Ugh. It's a mess.

 

[btk2k2]

Previous testing has shown they dont stick to AMD spec, I even see it on my b450 pro 4.

Last time they (ASUS) didn't stick to spec AMD released a new AGESA to force them too. I didn't see the motherboard vendors kick up a fuss about that when AMD forced their hand so why would they when Intel does? Would be interested in what spec your board is not sticking to.

As long as they are all on the same playing field so one vendor does not gain an advantage over the others they won't care all that much.

 

[chrcoluk]

The turbo frequencies are part of the specs though, and thus should be guaranteed. It's obviously not guaranteed under all conditions but the general idea is: if you have power and thermal headroom the chip should run at these speed.

We need to be very clear what the problem is: we're not talking about thermal throttling or inefficiencies or motherboard vendors goosing limits (these were all talked about last year): we're talking about these chips *point blank* not being capable of operating at their advertised speeds without degrading very quickly. And what Intel is doing is trying to say "well we never really advertise those speeds lasting for more than 3 seconds". Intel isn't entitled to silently redefining the last 10 years of how turbo limits have actually been implemented.

And don't let that distract you from the fact they're only saying any of this because they know that the chip is incapable of sustaining the advertised clock speeds without degrading. They're trying to dance a line and invent excuses so they don't get a class action lawsuit over their marketing running past their engineering.

I linked on here not long ago a document from intel that clearly states turbo speeds are not assured, of course media has embedded it into people's heads the expectation they should be able to get some kind of locked turbo clock performance. However they are merely a max possible burst speed and nothing more.

we're talking about these chips *point blank* not being capable of operating at their advertised speeds without degrading very quickly.

No we not talking about that, I dont know if you have some kind of vendetta, but that is not the subject. Chips can hit the specified clocks under the right conditions without degrading.

motherboard vendors goosing limits

This is what we are actually talking about.

Last time they (ASUS) didn't stick to spec AMD released a new AGESA to force them too. I didn't see the motherboard vendors kick up a fuss about that when AMD forced their hand so why would they when Intel does? Would be interested in what spec your board is not sticking to.

As long as they are all on the same playing field so one vendor does not gain an advantage over the others they won't care all that much.

It did get resolved eventually yes, but ASUS still did it, and not only did they do it, they started editing their website to remove liability, and basically rewrote history, horrible behaviour.

We dont know yet what the end game will be here, what we do know is Intel has asked the board vendors to make spec the cmos default, whether they comply or not remains to be seen.

 

[mxthunder]

What a mess. Although it kind of has been for last 4-5 generations. They finally just found the limit before the average end user experiences crashing.
I wont even let a new intel system boot without going into the BIOS and tweaking everything how I want it anymore.
The default voltage curves are almost scary on some of the Gigabyte boards.

 

[BoggledBeagle]

It seems that Intel has nothing to do with the "baseline profile" and suggests using Intel defaults settings with the highest power profile the motherboard can handle.

Intel Statement: Intel äußert sich und empfiehlt das Baseline-Profil nicht - Hardwareluxx

Intel Statement: Intel äußert sich und empfiehlt das Baseline-Profil nicht

www.hardwareluxx.de

Power limits, maximal currents and LLC settings have been discussed, but what about THE EXTREME FREQUENCIES ? Not only are the CPUs running at those high frequencies power hungry and inefficient, but are not these extreme frequencies a big part of the problem? ....

BTW has anyone heard about an unstable 14600K? If not, why is that? Is it because it runs at sane frequencies (5300/4000 MHz)? Is it because thanks to its lower core count and the limited frequencies is has limited power draw and it does not overheat? With lower current density and lower temperature leading to slower degradation?

 

[londiste]

BTW has anyone heard about an unstable 14600K? If not, why is that? Is it because it runs at sane frequencies (5300/4000 MHz)? Is it because thanks to its lower core count and the limited frequencies is has limited power draw and it does not overheat? With lower current density and lower temperature leading to slower degradation?

Pretty sure it is the frequencies.

 

[BoggledBeagle]

Pretty sure it is the frequencies.

I wonder what will persuade Intel to accept that they are simply running the CPUs quicker and with way more power than what the silicon can actualy handle.

 

[AusWolf]

So basically, changing any setting from default results in instability and is not recommended. What a load of ****.

Why doesn't Intel enforce stricter current and voltage curves so that even changing the aforementioned settings wouldn't cause any problems?

 

[londiste]

So basically, changing any setting from default results in instability and is not recommended. What a load of ****.

Why doesn't Intel enforce stricter current and voltage curves so that even changing the aforementioned settings wouldn't cause any problems?

But that is not the case at all here. It is specific couple of settings lowering the voltages that cause instability.

Stupid power limits are fine as long as you can cool the thing and if you cannot it throttles. Considering what settings have been pushed to these as defaults I would even go as far as saying these have been remarkably stable from the side of pushing way too much power into it.

 

[ArcanisGK507]

This must be stupid... you buy a K series CPU and a motherboard with open profiles with the intention of doing Minimum OC or having the system do it automatically to gain a little performance in games or specific software and you end up with an unstable system...

I better buy a core i3/i5 and I already know that it comes limited and capped from the factory and I skip the instability problems, knowing that I can't have what I'm paying for... it loses all sense and I go for a stable AMD. ..

 

[:D:D]

I am seeing the bigger picture, sadly there is so much hatred for Intel, and for some reason some kind of unexplained loyalty to ASUS and co that people are not seeing things straight.

No hate from me, I just gave my opinion, someone should ask Intel why they allow such a setting of 115C for Tjmax

Most CPU's are set for around 100C and Tjmax itself cannot be changed, just a negative offset that can be used. My own Xeon v3 is set at 78C by Intel depending on which microcode patch is being used.

Had a quick look for that Gigabyte link but this one for 9th Gen and default 115C should do I think.

All of the motherboard vendors like to get a one-up, which is why we see features like Multi-Core Turbo (which we reported on in 2012) sometimes defaulting to 'on'. Motherboard manufacturers prefer the 'Unlimited PL2' route, because it puts their results at the top of benchmark lists.

Some reasoning from https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

 

[chrcoluk]

No hate from me, I just gave my opinion, someone should ask Intel why they allow such a setting of 115C for Tjmax

Most CPU's are set for around 100C and Tjmax itself cannot be changed, just a negative offset that can be used. My own Xeon v3 is set at 78C by Intel depending on which microcode patch is being used.

Had a quick look for that Gigabyte link but this one for 9th Gen and default 115C should do I think.

Some reasoning from https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

Do Intel code the bios limits then? The spec sheet states 100C, that to me says they dont allow it if you expect your CPU to stay in warranty. Which matches up with the response they told the customer in the tech support thread I linked to.

I think anything bios related cannot be rationally pointed at anyone other than the board vendors, unless its something Intel has told them to do or provided as a baseline, I think we can be reasonably confident they havent provided a baseline of 115C tjmax.