News Intel finally announces a solution for CPU crashing errors — claims elevated voltages are the root cause; fix coming by mid-August

[bit_user]

In spite of one word being a bit of taking liberty, it doesn't change the fact that the situation had to be corrected by AMD just as this situation will have to be corrected by Intel. I feel pretty confident that a company with their lineage of fine products and history as well as market cap will make this right one way or the other.

But why are they taking so long? That's one of the more troubling aspects of this whole fiasco.

Also, I am not trying to "fanboi" one over the other, just mention to point out that both of these manufacturers are trying really hard not to be Icarus.

From what I understand, AMD never wanted nor recommended manufacturers to use such high voltages, in which case it would seem misplaced to label them Icarus.

 

[pocketdrummer]

An "overclockable" CPU that can't be overvolted? Nice, Intel.

 

[-Fran-]

Would it surprise you if amd has done it as well? It would surprise me. Amd fans are doing the shilling non stop for 0$, amd doesn't need to spend any money on this 😁

Yes. I don't think AMD has, because for most part of the last decade they've had no money to hire actors. The best we got here in Tom's forums was an official PR person before the Bulldozer launch and you can imagine how that went down.

Fanbois are one thing, and it's somewhat easy to flush them out, but distinguishing between a paid actor and someone dumb enough to represent Company interest without being paid. A tad harder to* go that one extra level deeper.

Are you getting paid?

Regards.

 

[punkncat]

But why are they taking so long? That's one of the more troubling aspects of this whole fiasco.


From what I understand, AMD never wanted nor recommended manufacturers to use such high voltages, in which case it would seem misplaced to label them Icarus.

To fix a problem one has to know the cause. Maybe they do, perhaps not, but working at it is better than no response at all.

As far as AMD is concerned, and admittedly I have not been deep diving research on their CPU since around 2xxx/3xxx. Early on they were quite adamant about not going farther than 1.4V (IIRC) because it would cause damage. I don't recall there being much insistence afterward, but leaving pretty much everything they offer "unlocked" says a lot by itself. Both of the manufacturers want to 'aim for the sun' and in both cases their wings melted, just as Icarus and a fine analog to that myth.

 

[-Fran-]

Did you actually watch Buildzoid's video?

He doesn't in any way suggest that what Intel is saying is wrong. He tossed out some humor with regards to TVB, but also says repeatedly he doesn't know what exactly Intel is changing. There's a lot of talk about runaway voltages especially under lower power/heat due to the way the VF curves effectively work.

No one is saying what Intel will do is wrong; my position is they're leaving stuff out conveniently in order to save a buck.

Lowering the voltage suplied will alleviate most of the issues, specially for systems not being taxed hard; thing is, it's also been demonstrated it won't fix anything by itself.

In what BZ analyzed there were 2 key elements which will help the overall discussion:
1.- Intel upped the voltage limit guidance to 1.72v (1.52v+offset, as they described it before, now it's just 1.72v), but everyone knows the practical limit is 1.5v and any motherboard which is supplying up to 1.5v is within official spec. Which is kind of funny?
2.- If seeing the voltage so high on a simple inspection, why have they taken so long to come up with this response? Also, keep in mind they've already been toying around with lowering power and voltage. Even on the SuperMicro motherboard he uses to analyse the situation he spitballs under 1.5v operation.

Don't get me wrong, I want Intel to do right by its customers. I am just doubting what they're saying in this communication is going to fix things. Everything keeps pointing that way.

EDIT: Forgot to mention one element of #1 -> when the max VID for the boost is disabled the failure rate went down from 100% to ~30%. So yeah, it helps, but it doesn't fix.

EDIT2: Jay is saying people are sending him emails saying Intel is refusing RMAs. Grain of salt in terms of volume and specifics, but welp. EDIT 2.1: It was for CPUs in pre-built PCs, so that makes more sense.

View: https://www.youtube.com/watch?v=nrWQLFWbQY8


Regards.

 

[Matriaz]

Its brilliant, they have just invalidated all evidence companies will continue to see as it doesn't repair the CPU.

It will be impossible to identify if crash reports are from new CPU's or the old degraded ones. To prove them right or wrong you would have to buy a new CPU now and run it for a year or so before you could discover if the fix actually worked with this deflecting the problem years down the line.

 

[vanadiel007]

OEM's reaction will be interesting, as they are responsible for warranty and not Intel, since they are OEM and not retail CPU's.

They will blame it on everything else and have you reinstall Windows and drivers forever and a day, before they will replace your CPU due to "degradation".

Intel should do the right thing and replace all CPU's of the affected generations regardless of them working or not.
You bet the second hand market is going to explode for years to come with "Intel CPU's" being sold as working when in reality they are damaged.

I am glad I stuck with my AM4 platform, with everything working well for my needs. I feel sorry for those with Intel CPU's because you have no way of knowing if yours has damage or not...

 

[thestryker]

In what BZ analyzed there were 2 key elements which will help the overall discussion:
1.- Intel upped the voltage limit guidance to 1.72v (1.52v+offset, as they described it before, now it's just 1.72v), but everyone knows the practical limit is 1.5v and any motherboard which is supplying up to 1.5v is within official spec. Which is kind of funny?

It's been that high since 8th gen this is nothing new.

EDIT: Forgot to mention one element of #1 -> when the max VID for the boost is disabled the failure rate went down from 100% to ~30%. So yeah, it helps, but it doesn't fix.

They turned off TVB which doesn't actually fix the VF curve, but it's basically a slight downclock.

2.- If seeing the voltage so high on a simple inspection, why have they taken so long to come up with this response? Also, keep in mind they've already been toying around with lowering power and voltage. Even on the SuperMicro motherboard he uses to analyse the situation he spitballs under 1.5v operation.

The KS SKUs go over 1.5 volts regularly which means it's not as simple as "oh 1.5v+ bad" which he also explained in the video.

No one is saying what Intel will do is wrong; my position is they're leaving stuff out conveniently in order to save a buck.

Please explain how leaving stuff out saves them money if this actually fixes the problem?

I certainly agree if this doesn't fix the problem then it's because they're hiding what the real problem is.

 

[Gururu]

Do you know what electromigration is?

Intel does, AMD does, they have been designing complex electronically dense circuits for decades. Do you genuinely believe that they will not adapt the paper designs that they fabricate/present for fabrication to minimise the effects of electromigration?

There is always a minimal amount of damage and yes pushing for the last possible bit of clock speed will exacerbate the problem BUT ITS A KNOWN PROBLEM, run within specs and the device will last years, run slightly out of spec and it should last a little bit less… run excessively out of spec… BOOM.

The problems being discussed in the thread should not be happening. ICs do last for many years, look at your lcd tv, your washing machine, your car, your Casio watch and if run in spec they should and will continue to do so.

For intel, something has gone wrong, whether it is a design bug, a power implementation bug .. whatever.. Intel has a problem that needs addressing.

Arguably chips like the “ks” chips encourage people to push the operational envelope for those devices… push it just a little harder… little more V, make a little more available power to draw… rinse and repeat… and it’s fun getting the extra speed. A new chip shouldn’t be approaching the margins out of the box, they never used to. It was a user’s choice to push the chip into its danger zone.

A car dealer asked me how many years I expected to have my car and I told him 20 years. He was shocked. I think chip manufacturers would be shocked if I said I needed my CPU for five years. For them it’s maybe a couple of years tops. No surprise we need 1000w power supplies now. Is it bad engineers making a quick buck by depending on ever increasing voltage to hit faster quotas? Will we ever have smarter, faster chips with less voltage?

 

[bit_user]

I think chip manufacturers would be shocked if I said I needed my CPU for five years. For them it’s maybe a couple of years tops.

I was curious what warranty terms major OEMs would offer. I checked Dell.com and found they offer up to a 4-year warranty on Alienware and XPS machines offered with an i9-14900K (the default is 1 year).

What you might find more interesting is that Intel has an IoT group which (among other things) promotes the use of some of their CPU models for specialized applications, including:

• Retail, Banking, Education, and Hospitality - Integrated graphics supports immersive and interactive digital signage, video walls, AI-driven in-store advertising, and interactive flat panel displays (IFPDs) for services and storefronts.
• Healthcare - Performance for more devices, apps, and multitasking—alongside built-in AI acceleration—support more diagnostics and medical procedures, ultrasound imaging, medical carts, endoscopy, and clinical devices.
• Industrial - Enable machine vision use cases on the factory floor as well as real-time capabilities for critical workloads in AI-based industrial process control (AIPC), industrial PCs, and human-machine interfaces (HMIs).
• Smart Cities and Transportation - Support network video recorder (NVR) solutions with AI box and roadside units (RSUs) for computer vision, smart city, and smart transportation use cases with Intel® UHD Graphics and fast CPU image classification performance.

Many of those are products which have a service life much longer than a couple years. Probably at least 5 years, which is how long they guarantee availability of the i9-13900 and i9-13900TE.

In the Gen 13 models being promoted for use in such applications, they include the i9-13900, i9-13700, and i7-13700T. I wonder if the i9-13900T used to be included, but was subsequently removed.

Details, here:

https://www.intel.com/content/www/us/en/products/details/embedded-processors/core/13thgen.html
​

So, it's clear that Intel certainly intended these CPUs to be stable beyond the boxed retail CPU warranty of 3 years.

 

[Phaaze88]

Update from GN:

View: https://www.youtube.com/watch?v=OVdmK1UGzGs

 

[SyCoREAPER]

Place your bets:

A) Generic "Weeewere sowwy and promise to do better in the future.

B) Bribe with incentives, discounts and free stuff.

C) Change power settings which will just delay the inevitable.

D) Dick around with PR Statements, tapper off until people move along.

E) Exchange for a refab. Running a new batch of dies id imagine has to be cheaper than issuing everyone a refund.

F(unny) Full Refund.


Option E) would be the most economical because people otherwise have to rebuild no matter what.
The only way they could outdo that, which they absolutely will not is take Option B to the extreme. 15th-Gen CPU, Motherboard and RAM for those with DDR4.

Intel is screwed. Top bin 13 and 14th Gen hospice patients in the wild with no amount of code that can fix physical defects.

Edit
Completely OT, I was taking caffeine powder out of capsules and I'm 85% sure I touched my face at some point during the process

 

[Phaaze88]

A) Generic "Weeewere sowwy and promise to do better in the future.

You omitted, 'while rubbing their nipples'. Like in that one episode of South Park, where the cable company reps are doing so while apologizing.

My bet is on D, as they're kinda doing that already...

 

[TheHerald]

Honestly we are all speculating right now because we have no idea what the microcode will do. It may be negligible.

However if the microcode limits the speed of the chips to say max boost of 5.5 ghz while intel's marketing said 5.8 ... will people be happy or will there be lawsuits for false advertising? I ask specific to boost clocks because the "fixes" that are going around the interwebs right now basically cut max frequency and thus overall performance often times by more than just a few %.

Oh definitely both. People will be happy (who cares about the ST boost?) but there will be lawsuits too.

 

[TheHerald]

instead there was a dialog for selecting the power of your cooler

YES, that's what I'm talking about. If your CPU is running with no power limits it's because of your choice in that dialogue is all im saying.

 

[TheHerald]

Are you getting paid?

If you mean by Intel, I wish.

But at the end of the day does it matter? That's the definition of an ad hominem argument. Does it matter if someone is paid by PR or not? If what he is saying is correct, it's correct, and if it's wrong, it's wrong, whether or not he is getting paid to say it doesn't change the validity (or not) of what someone's saying.

 

[bit_user]

Does it matter if someone is paid by PR or not? If what he is saying is correct, it's correct, and if it's wrong, it's wrong, whether or not he is getting paid to say it doesn't change the validity (or not) of what someone's saying.

I'd like to weigh in on this, although my remarks shouldn't be taken as an accusation of any kind.

The reason I think it matters is that compensation comes with strings that both limit and force what one says. Even when a true statement is made, it could be a half-truth or carefully worded to avoid relevant caveats and qualifiers. When people get an answer from someone who's a PR mouthpiece, they deserve to know this, so they can weight it accordingly.

That's why I think any paid representative of a company should be required to disclose this fact. In the past, I've seen forum accounts that have a label on them, indicating a brand affiliation. Not recently, but it was more common back when the site had AMA (Ask Me Anything) sessions with manufacturers. For example:

AMA - Intel Optane AMA - Starts Tuesday at 12pm

Tom's Hardware is proud to announce the upcoming guest in our featured Community Series -- ASK ME ANYTHING. On Tuesday, March 12th through Wednesday, March 13th the Intel Optane Team will join us in a live AMA. Have a question about Optane? Now’s your chance to speak with Intel directly. Log...

forums.tomshardware.com

 

[TheHerald]

I'd like to weigh in on this, although my remarks shouldn't be taken as an accusation of any kind.

The reason I think it matters is that compensation comes with strings that both limit and force what one says. Even when a true statement is made, it could be a half-truth or carefully worded to avoid relevant caveats and qualifiers. When people get an answer from someone who's a PR mouthpiece, they deserve to know this, so they can weight it accordingly.

That's why I think any paid representative of a company should be required to disclose this fact. In the past, I've seen forum accounts that have a label on them, indicating a brand affiliation. Not recently, but it was more common back when the site had AMA (Ask Me Anything) sessions with manufacturers. For example:

​

AMA - Intel Optane AMA - Starts Tuesday at 12pm

Tom's Hardware is proud to announce the upcoming guest in our featured Community Series -- ASK ME ANYTHING. On Tuesday, March 12th through Wednesday, March 13th the Intel Optane Team will join us in a live AMA. Have a question about Optane? Now’s your chance to speak with Intel directly. Log...

forums.tomshardware.com

​

If someone says a half truth it should be exposed, regardless of whether he is getting paid or not or what have you. Saying that who the individual is matters is basically a logical fallacy on it's own.

 

[bit_user]

If someone says a half truth it should be exposed, regardless of whether he is getting paid or not or what have you. Saying that who the individual is matters is basically a logical fallacy on it's own.

Yes, I would never take issue with who is saying something, if I agree with what they're saying. Heck, even when I don't, I try to focus on the statement and facts, trying to take each issue or point of contention on its own.

Where the conflict of interest becomes relevant is when you don't know enough to determine whether you're being told the full and accurate truth. That's when you'd really like any sort of strong biases to be disclosed.

BTW, I also believe people are likely to be better behaved when they're wearing the badge of their employer.

 

[TheHerald]

Yes, I would never take issue with who is saying something, if I agree with what they're saying. Heck, even when I don't, I try to focus on the statement and facts, trying to take each issue or point of contention on its own.

Where the conflict of interest becomes relevant is when you don't know enough to determine whether you're being told the full and accurate truth. That's when you'd really like any sort of strong biases to be disclosed.

BTW, I also believe people are likely to be better behaved when they're wearing the badge of their employer.

Ok, we are in agreement now.

But the issue is there is so much bias from fans of a specific company that any kind of normal discussion is just not an option. Everyone pretends there is no issue but of course there is, and if you can't see it you are probably part of the issue.

There are so many content creators that have said that they cannot make content that's not positive towards a specific company. There are creators like mlid and all those similar channels that gets hundreds of thousands of views spreading nonsense just because the nonsense they spread is pro that specific company.

You are against missinformation but you havent once corrected anyone when they keep claiming the usual "inefficiency problem", when all they are doing is ignoring the actual efficient skus and just focus on unlocked unlimited chips running amok on unlocked unlimited mobos. Does that seem like an unbiased argument to you?

The fact (it's a fact) is that if you take ALL of the existing skus of both brands and put them on an efficiency chart for ST and one for MT, you won't find a single amd chip on the top 10 of those charts out of the box. But everyone is trying to convince me they are super efficient. That to me is madness, and if I didn't know how crazy the fans of this specific company are I'd thought they were being paid.

 

[KyaraM]

Default motherboard settings was configured before. All vendors doing the same. Same Z690 boards used on 12900K without degradation.

That default Bios values where endorsed by Intel because they allowed the 13900k to be the performance king over AMD. All reviews and benchmarks sold the CPU with those parameters. Only after massive degradation was reported, Intel started to promote the "Intel defaults" where was those "Intel defaults" parameters before this issue arised?

Clearly if those parameters existed Intel keep them as hidden as possible to avoid losing the crown.

Now, Intel is reporting that the issue is with the eTVB voltage algorithm, so their own product was degrading by itself. And you keep saying that this degradation is somewhat "normal" because customers where applying too much voltage? Be serious.

My mainboard (MSI PRO Z690-A) prompted me to choose power limits based on cooling solutions. One option was default limits, one was big air cooler, one was AIO (or something similar; been a while since I last looked and resetted to factory settings). Default should be Intel limits. Air was PL1 = PL2 = 190W for my 12700k iirc.

Also, at least for power limits, it seriously isn't hard to find those online. Intel literally gives you both PL1 and PL2 limits for all their CPUs on their website, so you can look them up and adher to them (or not) on your own volition.
https://www.intel.com/content/www/u...-25m-cache-up-to-5-00-ghz/specifications.html
I didn't dig through the datasheet, so no idea if anything else is said in there regarding voltages. But power limits? I'm sorry, if you can't even look up this much, I don't know what to say. It's really not hard, took me all but 5 seconds.

And there are reviews at default power limits...

 

[TheHerald]

My mainboard (MSI PRO Z690-A) prompted me to choose power limits based on cooling solutions. One option was default limits, one was big air cooler, one was AIO (or something similar; been a while since I last looked and resetted to factory settings). Default should be Intel limits. Air was PL1 = PL2 = 190W for my 12700k iirc.

Also, at least for power limits, it seriously isn't hard to find those online. Intel literally gives you both PL1 and PL2 limits for all their CPUs on their website, so you can look them up and adher to them (or not) on your own volition.
https://www.intel.com/content/www/u...-25m-cache-up-to-5-00-ghz/specifications.html
I didn't dig through the datasheet, so no idea if anything else is said in there regarding voltages. But power limits? I'm sorry, if you can't even look up this much, I don't know what to say. It's really not hard, took me all but 5 seconds.

And there are reviews at default power limits...

The best "review" ive seen is one from gnexus. He was trying to show how terribly inefficient intel chips are even at iso power. So he limited the 14700k to 90 watts and run blender. It literally topped the entire chart in efficiency (in an amd friendly test as well) only losing to a 5.999$ threadripper. Still that didn't phase him and he doubled down saying the result is bad cause performance is lower than when running unlimited.

I mean... what the hell?

 

[Peksha]

YES, that's what I'm talking about. If your CPU is running with no power limits it's because of your choice in that dialogue is all im saying.

What makes you think there are no restrictions? as already written here, there are three options - air (less than 200W), stationary (about 200W), aio (about 253W).

You have no experience with these motherboards)

Both boards didn’t even know that for KS pl=320W and iccmax=400A.

 

[-Fran-]

It's been that high since 8th gen this is nothing new.

They turned off TVB which doesn't actually fix the VF curve, but it's basically a slight downclock.

The KS SKUs go over 1.5 volts regularly which means it's not as simple as "oh 1.5v+ bad" which he also explained in the video.

Please explain how leaving stuff out saves them money if this actually fixes the problem?

I certainly agree if this doesn't fix the problem then it's because they're hiding what the real problem is.

Well, two general things:
1- Going over 1.5v is bad. Specially if it's for long periods of time. What BZ tried to understand is under what circumstances that would happen, which he didn't have a clear answer. The fact Intel upped the limit in newer specs is weird (in hindsight), but motherboard vendors are not going "out of spec" by pumping over 1.5v even if anyone that has ever run an Intel CPU knows you do not want to pump that much voltage into it.
2.- This saves money because it would lower the RMA count and image. If they say "ah, yeah; we screwed up the CPU VID tables so we need to recall them" (for example; manufacturing defect is another example) that is more expensive for sure and it implies they'll have to flat out replace full SKU lines. If they tell people they can fix it via microcode by capping voltages and lowering performance, then they'll delay the innevitable and consequences. I thought this would be somewhat obvious.

If you mean by Intel, I wish.

But at the end of the day does it matter? That's the definition of an ad hominem argument. Does it matter if someone is paid by PR or not? If what he is saying is correct, it's correct, and if it's wrong, it's wrong, whether or not he is getting paid to say it doesn't change the validity (or not) of what someone's saying.

Uh... It obviously matters. Do I really need to explain it? Really?

Regards.

 

[TheHerald]

Uh... It obviously matters. Do I really need to explain it? Really?

Yes, please. If you say something wrong it's wrong regardless of whether you are paid, you are a fanboy or just mistaken. Just change the validity of your argument. It's literally a 101 logical fallacy.