François Piednoël Quits Intel

[PowerPC]

For the fanboys: Ryzen is still slower.

Cut this guy some slack. This is years and years of Intel indoctrination talking.

 

[notb]

What does that tell you about core (uarch) & Intel;s very own 14nm if anything at all?

Nothing.
Intel is preparing their next architecture, but in the meantime it would be great to respond to AMD's new CPUs. And Skylake has enough potential to do it.

Generally speaking, Intel's 14nm architecture is still more efficient than current Zen.
i7-7700 draws less power than Ryzen 5 1400 (idle and load) while being more powerful (both 4C/8T, TDP 65W).

Them or Apple, not many firms can afford someone like him, especially if he wants to be in charge of things.

Actually quite a lot of companies can afford him. It's just that he might be unwilling to work for a smaller company.

And then one day we wake up in a post human era, feeding our electrical energy to their power grid. Frankly I don't understand this obsession with AI, Skynet or whatever comes after will remember us for the fools we were.

Well... we've already woken up once with half of the world fighting with a weird guy with a mustache. And it wasn't the most fatal period in our history. We don't need AI to make problems.
However, we need AI to dream about shortening business hours or not having to do simple, time-consuming things (e.g. driving, cleaning, farming).

Sure, at some point AI will be able to build machines, program them and - potentially - fight humans. But even a possibility of that is still decades ahead. Today AI can only help.

 

[qubit]

I'll bet someone so central to the design of Intel CPUs leaving has something to do with Ryzen. It's clearly got Intel rattled so they need someone to blame and he's the scapegoat by the looks of it. It wouldn't surprise me if we start seeing Intel start to struggle to make competitive products in a couple of years.

Just my 2 cents, so no flaming please.

 

[EarthDog]

I'd don a flame suit for that comment too... jesus. Just bscause its .02 doesnt mean we cant take it out back and put it out of its misery.
Edit: being serious, anything is possible. From his twotter account though, i gkt the feeling they were going in different directions. Maybe Zen had something to do with it...


Want to bet hes going to be working on AI soon.... perhaps with apple... perhaps after his non compete is up? Anyone??

 

[john_]

Maybe that 5% every year on IPC + some extras from the frequency bumps wasn't what Intel wanted to give to the customers, restricting the good stuff, but what it could give. And as long as AMD was nowhere to be seen, probably François Piednoël was enjoying a calm working environment, where all Intel high ranking officials where happy with that progress.

But now with Ryzen maybe they started asking too much from him, pressuring him, or even blaming him of losing time and money all those last years. I can imagine a simple question to him
"How a company with 1/10 the size, resources, money, everything managed to get so close to our architecture? How can we gain that safe distance we had before Ryzen, in the near future? You have all the money and resources in the world at your disposal, how fast can you show results?".

It's just a possibility that, being a big name in the industry and probably a very wealthy person working all those years for Intel, maybe he decided to avoid a working environment full of pressure and anxiety for the coming years. Maybe he jumped ship because he values quietness more than a huge salary.

 

[Fluffmeister]

Maybe he has piles of cash and just wants to go fishing.

 

[Vya Domus]

Maybe he has piles of cash and just wants to go fishing.

I am pretty sure that's the case instead of the nonsense discussed on this thread.

 

[Th3pwn3r]

a guy like him assuredly has a termination/departure cause 3 miles long

he could't work for AMD if he wanted to

exactly, he is going to be so locked down its stupid. non-competes up the ass. Not to mention anything he touches will be looked at via intels lawyers the second they find out he touched anything that made it to the patent office.

AMD would't hire him if they could they IP/anti poaching lawsuit's would be a nightmare

There are always ways around these sorts of contracts. It just depends on how 'creative' you want to be. Example being, if I leave my union I'm not supposed to work in my current field and receive a wage. I can 'volunteer' and accept 'donations' of course. You can also hire someone you pay directly and then they have a 'team' or an employee they pay to do something via subcontracting.

 

[TheLostSwede]

A lot of discussions about a person none of you have ever met.
If you had, the discussion would be quite different I think.
You don't leave a job where you've been for almost 20 years just like that, so something has clearly happened to make him leave on what almost looks like a whim.

He posted this on LinkedIn a few days ago, which might help explain a little bit about why he left.

Being a Tech-leader does not mean being a strong minded not listening feared old man, it means enabling the young engineers to be successful, and lead them by example, accept when you are wrong, and enable learning out of it, and always seek the less senior point of views. Being innovating all the time for long period of time is only possible in those conditions. And If your environment or management does not enable that kind of ways, move on!

https://www.linkedin.com/feed/update/urn:li:activity:6293148635236564992

 

[R0H1T]

Maybe that 5% every year on IPC + some extras from the frequency bumps wasn't what Intel wanted to give to the customers, restricting the good stuff, but what it could give. And as long as AMD was nowhere to be seen, probably François Piednoël was enjoying a calm working environment, where all Intel high ranking officials where happy with that progress.

But now with Ryzen maybe they started asking too much from him, pressuring him, or even blaming him of losing time and money all those last years. I can imagine a simple question to him
"How a company with 1/10 the size, resources, money, everything managed to get so close to our architecture? How can we gain that safe distance we had before Ryzen, in the near future? You have all the money and resources in the world at your disposal, how fast can you show results?".

It's just a possibility that, being a big name in the industry and probably a very wealthy person working all those years for Intel, maybe he decided to avoid a working environment full of pressure and anxiety for the coming years. Maybe he jumped ship because he values quietness more than a huge salary.

Of course, that's like saying Atom was crap because Intel wanted it to be so, never mind the billions they burnt trying to take a chunk out of QC, MTK, Apple, Sammy et al.

Intel did it's best to make core into what it it is today i.e. Skylake, SKL-X is reason enough for me to believe that the next 5~15% IPC gains they may get out of this uarch are gonna come at a high enough cost, like temps for SKL-X, that it might force them to go for something more revolutionary. This doesn't mean that they'll get the 5~15% IPC gains easily with the next few iterations, clock speeds going to be a problem again on 10nm, but rather it's unlikely that Intel can milk their cash cow too much without having to reduce prices or their fat margins, now that AMD is back in the game.

 

[dicktracy]

Except for skylake-x lololololol. Poor intel.

1800x is overall slower than even a 7800x.
http://www.anandtech.com/show/11550...ew-core-i9-7900x-i7-7820x-and-i7-7800x-tested

AMD is still playing the budget brand like they have been doing for years and still gets outsold by Intel and Nvidia. But I guess AMD is attractive to the poorer PC users that seems to be the vocal minority. They have zero performance lead in both CPU and GPU.

 

[john_]

Of course, that's like saying Atom was crap because Intel wanted it to be so, never mind the billions they burnt trying to take a chunk out of QC, MTK, Apple, Sammy et al.

Intel did it's best to make core into what it it is today i.e. Skylake, SKL-X is reason enough for me to believe that the next 5~15% IPC gains they may get out of this uarch are gonna come at a high enough cost, like temps for SKL-X, that it might force them to go for something more revolutionary. This doesn't mean that they'll get the 5~15% IPC gains easily with the next few iterations, clock speeds going to be a problem again on 10nm, but rather it's unlikely that Intel can milk their cash cow too much without having to reduce prices or their fat margins, now that AMD is back in the game.

Did I mentioned Atom somewhere? Maybe you could show me where I was talking about Atom. Atom is a very peculiar case because you have to take an architecture that wasn't meant to run at 2-4 W TDP, throw away as much as you can and still end up with a chip that can run everything in the x86 library. And of course those billions didn't gone on R&D, but on tablet manufacturers to agree to use Intel chips. But as I said, where have you seen me talking about Atom? Did the fact that I mentioned Ryzen brought to your mind Atoms instantly?

As for the IPC, it is usually 5%, if not less in the latest generations of Intel processors. The rest is just frequency. So in a way you are repeating more or less what I wrote. As for temps, the problem with Intel cpus, is with the interface material between the cores and the heatspreader. The more revolutionary thing they can do, is find an acceptable way for them to solder the die on the heatspreader.

 

[R0H1T]

Did I mentioned Atom somewhere? Maybe you could show me where I was talking about Atom. Atom is a very peculiar case because you have to take an architecture that wasn't meant to run at 2-4 W TDP, throw away as much as you can and still end up with a chip that can run everything in the x86 library. Bus as I said, where have you seen me talking about Atom? Did the fact that I mentioned Ryzen brought to your mind Atoms instantly?

As for the IPC, it is usually 5%, if not less in the latest generations of Intel processors. The rest is just frequency. So in a way you are repeating more or less what I wrote. As for temps, the problem with Intel cpus, is with the interface material between the cores and the heatspreader. The more revolutionary thing they can do, is find an acceptable way for them to solder the die on the heatspreader.

I am, I'm clarifying what I think others see Intel doing i.e. some say that AMD wasn't competitive enough that's why we had measly IPC gains from Intel. The reality is IPC isn;t something that you pull out like a rabbit from the proverbial hat, KBL & now CFL have shown us that clock speeds & more cores is the way of the future, even for desktops. SKL-X tells me that Intel is running into a wall & that even a major rejig, mesh vs ring, will not guarantee them IPC gains across the board. As for solder I doubt an 18 core dissipating that much heat will get cooler just because of solder, probably bare die would be the best option for enthusiasts.

 

[john_]

AMD is still playing the budget brand like they have been doing for years and still gets outsold by Intel and Nvidia. But I guess AMD is attractive to the poorer PC users that seems to be the vocal minority. They have zero performance lead in both CPU and GPU.

Intel and Nvidia have better brand recognition. Not so much NVidia, but definitely Intel. So even if AMD creates better products, it will still be cheaper in the market. Think it like this. Coca Cola is the best known brand with Pepsi being probably second. Even if Pepsi creates a new cola, better than anything on the market, the average consumer will still have a problem to buy a Pepsi can when a cheaper Coca Cola is sitting just next to it. So Pepsi will have to sell at the same price or lower it's product, even if it is better.

As for poorer people. It's obvious that there are also poorer people in the brain, who love to express their problem with the vocal minority, that ruin the neighborhood obviously.

 

[dwade]

Jim Keller left the sinking ship first.

 

[john_]

As for solder I doubt an 18 core dissipating that much heat will get cooler just because of solder, probably bare die would be the best option for enthusiasts.

Well, it wouldn't make things worst. Right now is like having a thermal barrier between the cores and the heatspreader. Removing that, will move the problem to the cooler. A good cooler will definitely do a better job than what it can do today. As things are now, the problem is under the heatspreader and the better cooling solution benefits are not fully exploited.

 

[Deathlokke]

For those saying he had to sign a non-compete contract and won't be able to work for any other company, that may not be the case. In California, non-competes are completely unenforceable. Here's a good article on what is and isn't allowed: http://www.huffingtonpost.com/entry...ompete-agreements_us_58af1626e4b0e5fdf6196f04. Intel's HQ is in Santa Clara, which I assume is where most of their senior staff operate from. He can't work for both Intel and AMD at the same time, but he can move to any other company he wants to, as long as he doesn't divulge company secrets.

 

[Devon68]

Good for him. Leaving must have been a big decision but if he felt that was what he needs he did the right thing. It might not even been Intel's fault he decided to leave. Maybe he found something that he want's to do more. As we mature in life, we might find that something we loved doing in the past might be replaced by something else we enjoy more.

 

[yotano211]

Maybe the dude has a hot wife that he wants to bang more before he leaves mother Earth, who knows but him.

 

[EarthDog]

A lot of discussions about a person none of you have ever met.
If you had, the discussion would be quite different I think.
You don't leave a job where you've been for almost 20 years just like that, so something has clearly happened to make him leave on what almost looks like a whim.

He posted this on LinkedIn a few days ago, which might help explain a little bit about why he left.

https://www.linkedin.com/feed/update/urn:li:activity:6293148635236564992

winner winner... chicken dinner.

His twitter account posted a few things as well as i mentioned earlier.

The guy and intel were going different directions.

 

[Yukikaze]

You might want to write-in a correction: He is "a principal engineer", not "the principal engineer".

 

[TheGuruStud]

1800x is overall slower than even a 7800x.
http://www.anandtech.com/show/11550...ew-core-i9-7900x-i7-7820x-and-i7-7800x-tested

AMD is still playing the budget brand like they have been doing for years and still gets outsold by Intel and Nvidia. But I guess AMD is attractive to the poorer PC users that seems to be the vocal minority. They have zero performance lead in both CPU and GPU.

Maybe for all those intel tuned apps or ones that only care about individual core speed...but skylake-x is a joke as this shows. Inter-core communication is DREADFUL. No, thanks. I can see the real IPC here.

 

[dicktracy]

Referencing an outlier review from the same guy that managed to burn his 7800x. If anything, that reviewer is too incompetent to draw conclusions from. Everywhere else Skylake-X is faster in games than any Ryzen. And that's not even considering the ridiculous overclock headroom vs Ryzen's tiny ceiling of 4Ghz (with a golden chip).

 

[Captain_Tom]

Referencing an outlier review from the same guy that managed to burn his 7800x. If anything, that reviewer is too incompetent to draw conclusions from. Everywhere else Skylake-X is faster in games than any Ryzen. And that's not even considering the ridiculous overclock headroom vs Ryzen's tiny ceiling of 4Ghz (with a golden chip).

Good lord you are pathetic. "Outlier review". It is practically unanimous that Skylake-X is a garbage line-up of fail products. Go read some more reviews fanboy, or maybe don't. We all know you will either cherry pick a few luke-warm reviews, or just put your head back in the sand.

 

[notb]

Good lord you are pathetic. "Outlier review". It is practically unanimous that Skylake-X is a garbage line-up of fail products. Go read some more reviews fanboy, or maybe don't. We all know you will either cherry pick a few luke-warm reviews, or just put your head back in the sand.

It's unanimous where? In the AMD coffee room?

Some more reviews (I hope I'm not cherry picking, since these are all major CPU reviewing sites):
http://www.anandtech.com/show/11550...core-i9-7900x-i7-7820x-and-i7-7800x-tested/17
https://www.bit-tech.net/reviews/tech/intel-core-i9-7900x-and-x299-chipset-revie/8/
http://hexus.net/tech/reviews/cpu/107017-intel-core-i9-7900x-14nm-skylake-x/?page=10

From Hexus:
"...a 10-core powerhouse offering excellent IPC performance and outstanding multi-core potential in a single $999 chip armed with plenty of overclocking headroom."
That doesn't exactly mean "garbage line-up of fail products" in my book.