Question Intel's future after Pat Gelsinger

[DAPUNISHER]

No one knows actually what is the real issue beside Intel cause we just had another microcode looks like broken design for P cores btw are E cores affected as well with this issue did someone has to tone down E core clocks?

I am not an IEEE guy, and I am not certain how to answer that. You can follow the links in this article from last month, published by our sister site, and perhaps get the answer you are looking for - https://www.pcgamer.com/hardware/pr...-gen-raptor-lake-cpus/?utm_source=chatgpt.com

 

[DAPUNISHER]

Well, I understand that Intel has neglected to disclose information.
However, when the oxidation problem of the manufacturing process was announced, the Raptor problem I was confused because it was the same as the information disclosure of
I understand that the root cause of Raptor Lake's 13/14 generation problem is different from the oxidation problem.

Read the the list of mitigations and microcode updates. Then you tell me, do you think there is one root cause of failures?

 

[Io Magnesso]

I am not an IEEE guy, and I am not certain how to answer that. You can follow the links in this article from last month, published by our sister site, and perhaps get the answer you are looking for - https://www.pcgamer.com/hardware/pr...-gen-raptor-lake-cpus/?utm_source=chatgpt.com

New microcode released last monthI don't know why it was released
Maybe it's just a reflection of feedback from microcodes dealing with the root cause

 

[DAPUNISHER]

New microcode released last monthI don't know why it was released

"The new microcode fix—dubbed 0x12F and not to be confused with the 0x125, 0x129 and 0x12B patches that came before it—"further improves system conditions that can potentially contribute to Vmin Shift Instability on Intel Core 13th and 14th Gen desktop-powered systems."

Maybe it's just a reflection of feedback from microcode dealing with the root cause

LOL I don't even know what this is supposed to mean. If by feedback you mean continued reports of degradation and failures then all signs point to yes.

 

[Io Magnesso]

Read the the list of mitigations and microcode updates. Then you tell me, do you think there is one root cause of failures?

You can see the 4 scenarios that cause a Vmin shift I understand that
However, I understand that the root cause is due to the clock tree circuit in the IA core. Same for all scenarios

 

[Io Magnesso]

I'm sorry that I acted like an Intel believer.
I apologize again

 

[DAPUNISHER]

I'm sorry that I acted like an Intel believer.
I apologize again

No, you are good, and have nothing to apologize for. You can like anything you like. And this is an Intel thread so being pro Intel in here is more than OK. I only get involved to make certain misinformation does not get passed along as correct information.

 

[Io Magnesso]

No, you are good, and have nothing to apologize for. You can like anything you like. And this is an Intel thread so being pro Intel in here is more than OK. I only get involved to make certain misinformation does not get passed along as correct information.

Thank you very much Thank you from the bottom of my heart
It's very confusing Does the Raptor Lake degradation problem come from the oxidation problem of Intel 7? Or maybe it's a different issue for each?
Until now, I thought that these issues had nothing to do with each other and were independent.
If there is actually a relationship, it is not possible to explain the failure of Raptor Lake (Refresh)-S, which is the production lot since the oxidation problem was fixed.

 

[511]

I am not an IEEE guy, and I am not certain how to answer that. You can follow the links in this article from last month, published by our sister site, and perhaps get the answer you are looking for - https://www.pcgamer.com/hardware/pr...-gen-raptor-lake-cpus/?utm_source=chatgpt.com

After searching a bit online looks like E core and P core are different clock domains so looks like they don't share this issue looks like someone messed the P core implementation in RPL from my searching online. I don't have proof to back it up though just a speculation.

 

[Joe NYC]

Even Robert Hallock tried to do something similar for ARL

Still seems unreal that Robert Hallock went from selling Zen 5 to selling Arrow Lake. Probably the worst career move in history, unless Intel is paying him double.

 

[Joe NYC]

Even I would not recommend that despite my above average understanding of most medical matters. I'm not trained by professionals so no one should hire me to save someone's life.

But in a pinch, I bet you can do a quick appendectomy, no problem.

 

[Hitman928]

But in a pinch, I bet you can do a quick appendectomy, no problem.

Appendectomies are easy. It’s making sure the patient survives that’s the tricky part.

 

[igor_kavinski]

But in a pinch, I bet you can do a quick appendectomy, no problem.

Considering how I program (does it blow up? OK, phew, didn't blow up), I would have to be a russian surgeon experimenting on their dispensable soldiers or prisoners before I'm able to get one right.

 

[Io Magnesso]

Thank you very much Thank you from the bottom of my heart
It's very confusing Does the Raptor Lake degradation problem come from the oxidation problem of Intel 7? Or maybe it's a different issue for each?
Until now, I thought that these issues had nothing to do with each other and were independent.
If there is actually a relationship, it is not possible to explain the failure of Raptor Lake (Refresh)-S, which is the production lot since the oxidation problem was fixed.

At least the Intel 7 (Enhanced) processor produced after 2024 seems to have fixed the oxidation problem of the via.
Laptop processors are said to have no problems
At least I've never heard of a laptop encountering a 13/14 generation problem.
And it seems that control problems can be avoided by applying a microcode patch to the New Raptor Lake.
This is what I understand at the moment
It's difficult, isn't it... this

 

[Doug S]

Appendectomies are easy. It’s making sure the patient survives that’s the tricky part.

Being able to tell the appendix and the liver apart also helps.

 

[DrMrLordX]

Only Intel knows the truth

Not to beat a dead horse, but Intel could have released lot numbers for all CPUs affected by the oxidation issue and chose not to do so. We'll never know exactly how many CPUs were affected. Pretty embarassing. At least they seem to have ironed out the issue since only Intel 7 products were affected.

Not sure how prospective foundry customers would really feel about that. They'll want to know how Intel will prioritize foundry customer orders vs. internal orders.

 

[Io Magnesso]

Not to beat a dead horse, but Intel could have released lot numbers for all CPUs affected by the oxidation issue and chose not to do so. We'll never know exactly how many CPUs were affected. Pretty embarassing. At least they seem to have ironed out the issue since only Intel 7 products were affected.

Sure, it would have been better to publish the affected lots.
Since the release date has only been one year since the release from Raptor, you can receive the warranty period. Maybe there was no need to take special measures It's my opinion
I wouldn't say it's good…

 

[Thibsie]

Sure, it would have been better to publish the affected lots.
Since the release date has only been one year since the release from Raptor, you can receive the warranty period. Maybe there was no need to take special measures It's my opinion
I wouldn't say it's good…

If I was a potentiel foundry customer, I'd be scared. If Intel can't acknowledge such problem for their own use, I would not trust them with my products.

 

[Io Magnesso]

If I was a potentiel foundry customer, I'd be scared. If Intel can't acknowledge such problem for their own use, I would not trust them with my products.

It's good to be anxious
To be honest, I can understand from the bottom of my heart that I can't trust people
It's good to feel that you can't trust

 

[511]

If I was a potentiel foundry customer, I'd be scared. If Intel can't acknowledge such problem for their own use, I would not trust them with my products.

Do you think Intel didn't tell intel themselves 🤣. I am sure they must have know what were the affected lots better way would be to check with OEMs what info did they got.

 

[Thibsie]

Do you think Intel didn't tell intel themselves 🤣. I am sure they must have know what were the affected lots better way would be to check with OEMs what info did they got.

I'm not even convinced foundry guys told/admited to the other guys.
Rotten culture is rotten culture.

 

[511]

I'm not even convinced foundry guys told/admited to the other guys.
Rotten culture is rotten culture.

Yeah the foundry under BK/Swan were similar in rotten culture they kept saying 10nm is fine we all know how fine turned out to be.

 

[Io Magnesso]

Yeah the foundry under BK/Swan were similar in rotten culture they kept saying 10nm is fine we all know how fine turned out to be.

At least we didn't provide it to external customers until Intel 7...
It's definitely different from Intel 4(3) or later, because we'll provide it to our customers.
I can't tell lies At least I think Intel himself is aware of it.

 

[StefanR5R]

Re: Intel Foundry not winning customers besides Intel itself —

I have read the argument that IP developers (e.g. CPU IP developers) are not keen on showing their IP to a foundry which is owned by another IP developer. To myself as a member of the great unwashed masses, this sounds plausible. But is it really?

 

[511]

Well Intel 3 is Industry Standard EDA Compatible whoever wants to use it besides Intel ofc.
The biggest attraction for Intel process are ironically UMC 12/Intel 16 🤣