Friday, July 21st 2017

François Piednoël Quits Intel

Who is François Piednoël, you ask? Why, just the principal engineer at Intel for nearly 20 years now. He has been involved in the architecture development of CPUs, including Katmai, Conroe, Penryn, and Nehalem as well as SoCs in Sandy Bridge, Ivy Bridge, Haswell, Broadwell, Skylake, and Kaby Lake. Oh, and did I mention he was also strongly involved in the Intel Atom processor line and the massive shift in Intel's microarchitecture from Pentium 4 to Core? He has also supported development of CPUz, Intel Hyperthreading, and the Android x86 platform.

François Piednoël is a big name in the industry, so it was a big surprise to see him quit Intel today. Time will tell where he ends up next, but if his reply tweet to his announcement is anything to go by it is not AMD. Nonetheless, we wish him the very best in his future endeavors and also remind readers that this does not necessarily mean anything for the future of Intel or the ongoing CPU market share battle.
Source: François Piednoël via Twitter
Add your own comment

94 Comments on François Piednoël Quits Intel

#1
HimymCZe
I wonder if there is ANY connection with Intel's decision cutting production of G4560 and going full SELLOUT-mode with CPU specs that no one wants or ask for...
Posted on Reply
#2
micropage7
its interesting if he works on mobile chip. lets see
Posted on Reply
#3
seronx
Bulldozer designer insurrection?!

google:
site:linkedin.com "Soft Machines" Bulldozer
site:linkedin.com "Intel" Bulldozer

http://uploads.tapatalk-cdn.com/20161216/909d6751a535e0d126ac20bb359fef12.jpg
Intel vs AMD, right there.

So, when the AMD guys are getting promoted at Intel... Well. Ouch.

I am the only one apparently, ready for the great Bulldozer revival. Don't care who does it, as long as it isn't 2x(2EX/2AGLU) w/ 2FMAC/(1)2MMX.
Posted on Reply
#5
efikkan
And precisely how is this newsworthy? This happens all the time.
Posted on Reply
#6
FordGT90Concept
"I go fast!1!11!1!"
seronx, post: 3697019, member: 86156"
Bulldozer designer insurrection?!

google:
site:linkedin.com "Soft Machines" Bulldozer
site:linkedin.com "Intel" Bulldozer

http://uploads.tapatalk-cdn.com/20161216/909d6751a535e0d126ac20bb359fef12.jpg
Intel vs AMD, right there.

So, when the AMD guys are getting promoted at Intel... Well. Ouch.

I am the only one apparently, ready for the great Bulldozer revival. Don't care who does it, as long as it isn't 2x(2EX/2AGLU) w/ 2FMAC/(1)2MMX.
WTH? Bulldozer was all AMD and AMD admitted it failed as an architecture. AMD has a shiny new architecture that can stick it to Intel. AMD is going to exploit that for years to come.

People usually quit at a company for internal reasons that generally don't go public.

I wouldn't be surprised if AMD makes him an offer, along with half a dozen other companies. It'll be at least six months before he can take a competitive position though.
Posted on Reply
#8
natr0n
He should stick to his roots and make fancy baguettes.
Posted on Reply
#9
VSG
Editor, Reviews & News
efikkan, post: 3697035, member: 150226"
And precisely how is this newsworthy? This happens all the time.
I would love to know how "this" happens all the time. He's been one of the most important contributors to Intel's rise as a CPU giant in the past decade.
Posted on Reply
#10
Basard
Closet AMD fan...... Wants to go to AMD but Intel will kill him lol.... Probably helped invent the bitcoin... I have like 500 more conspiracy theories to explain this.... not a single one of them includes him being rich as fuck and just wanting to retire.
Posted on Reply
#11
TheGuruStud
Basard, post: 3697057, member: 33749"
Closet AMD fan...... Wants to go to AMD but Intel will kill him lol.... Probably helped invent the bitcoin... I have like 500 more conspiracy theories to explain this.... not a single one of them includes him being rich as fuck and just wanting to retire.
I just want to think of it as rats fleeing a ship as it sinks.
Posted on Reply
#12
Captain_Tom
This will undoubtedly mark the next era of Intel. This is the guy that fixed Intel's Netburst mentality and made them the unmovable CPU juggernaut of the previous decade. Then Ryzen comes out, and Intel doubles down on GHz while throwing efficiency to the wind once again. Then Francois quits.


Sorry, but it doesn't even matter why he quit - history will remember him quitting the second Intel chose to go Full Netburst again.
Posted on Reply
#14
cdawall
where the hell are my stars
Captain_Tom, post: 3697063, member: 155434"
This will undoubtedly mark the next era of Intel. This is the guy that fixed Intel's Netburst mentality and made them the unmovable CPU juggernaut of the previous decade. Then Ryzen comes out, and Intel doubles down on GHz while throwing efficiency to the wind once again. Then Francois quits.


Sorry, but it doesn't even matter why he quit - history will remember him quitting the second Intel chose to go Full Netburst again.
Cart is a bit in front of the horse on that one. Intel hasn't changed architecture since he did Nehalem. It could be as simple as they said hey amd is doing things we need new and he said fuck you this has worked for 10 years.
Posted on Reply
#15
Captain_Tom
cdawall, post: 3697072, member: 28601"
Cart is a bit in front of the horse on that one. Intel hasn't changed architecture since he did Nehalem. It could be as simple as they said hey amd is doing things we need new and he said fuck you this has worked for 10 years.
Yeah I am sure that's what happened /s
Posted on Reply
#16
cdawall
where the hell are my stars
Captain_Tom, post: 3697074, member: 155434"
Yeah I am sure that's what happened /s
I am not saying it is, but a lot of people don't like being told they need to change.
Posted on Reply
#17
Dave65
efikkan, post: 3697035, member: 150226"
And precisely how is this newsworthy? This happens all the time.
Are you really serious?
Posted on Reply
#18
Captain_Tom
cdawall, post: 3697076, member: 28601"
I am not saying it is, but a lot of people don't like being told they need to change.
Alright I will be fair: None of us know for sure what happened.

But by your own statement "people don't like being told they need to change", and this guy was all about efficiency. Skylake-X is grossely inefficient. Connect the dots.
Posted on Reply
#19
cdawall
where the hell are my stars
Captain_Tom, post: 3697081, member: 155434"
Alright I will be fair: None of us know for sure what happened.

But by your own statement "people don't like being told they need to change", and this guy was all about efficiency. Skylake-X is grossely inefficient. Connect the dots.
Skylake x is being sold outside of its efficiency curve. The process is no less efficient than any of the others.
Posted on Reply
#20
thesmokingman
Captain_Tom, post: 3697063, member: 155434"
This will undoubtedly mark the next era of Intel. This is the guy that fixed Intel's Netburst mentality and made them the unmovable CPU juggernaut of the previous decade. Then Ryzen comes out, and Intel doubles down on GHz while throwing efficiency to the wind once again. Then Francois quits.

Sorry, but it doesn't even matter why he quit - history will remember him quitting the second Intel chose to go Full Netburst again.
cdawall, post: 3697072, member: 28601"
Cart is a bit in front of the horse on that one. Intel hasn't changed architecture since he did Nehalem. It could be as simple as they said hey amd is doing things we need new and he said fuck you this has worked for 10 years.
You guys are both arguing the same thing but different. I agree with Tom, that years from now this could be the first step of Intel slipping. It doesn't happen all at once, but it has to start somewhere... thinking back to P4 days. But regardless, you don't kick someone like Piednoël to the curb or let them walk unless the direction at the top has really changed. Piednoël is not a maverick like Jim Keller moving from interests to interests so it is quite alarming for Intel imo. Piednoël is lifer and when a lifer leaves, wtf?
Posted on Reply
#21
phanbuey
As was said -- this is like when Jim Keller left AMD. A clear sign of mismanagement if you lose talent like that.

Prepare yourselves ... for Netburst II

efikkan, post: 3697035, member: 150226"
And precisely how is this newsworthy? This happens all the time.
I believe the analogy is that its news worthy because of what happened to AMD - not a bad one. 20 Year old vets who contributed to the upswing leaving so unceremoniously is an ominous sign.

I know corporations try to buffer against it... but a 20 year veteran with that much knowledge and involvement leaving with no transition HURTS. Hell a 20 year veteran secretary leaving with no transition hurts, it's amplified as the position becomes more technical/complex.

We had a dev leave like that leave a 5 man team and debugging/reverse engineering his stuff set us back at least a month and a half.
Posted on Reply
#22
ensabrenoir
HimymCZe, post: 3697013, member: 168756"
I wonder if there is ANY connection with Intel's decision cutting production of G4560 and going full SELLOUT-mode with CPU specs that no one wants or ask for...
finally found one for $75 bucks at B&H. wished i got it ealier when it was in the $60 range. Upgrading my sons pc.

But someone had to take the blame for intel's kerfluffle indirectly or directly. New blood new ideas needed. But with a resume like that i think he has little to worry about.
Posted on Reply
#23
Hood
Perhaps the art of CPU design is getting too complicated, heading beyond the point where human brains can conceive meaningful and cost effective designs. The next truly new architecture might be designed by AI, a super-computer that can run simulations of all possible materials, foundry methods, and architectures, until it finds a breakthrough design. Skynet approaches...
Posted on Reply
#24
phanbuey
Hood, post: 3697093, member: 110365"
Perhaps the art of CPU design is getting too complicated, heading beyond the point where human brains can conceive meaningful and cost effective designs. The next truly new architecture might be designed by AI, a super-computer that can run simulations of all possible materials, foundry methods, and architectures, until it finds a breakthrough design. Skynet approaches...
All AI (at the moment) cannot frame the problem - that's the job of the designer. It's just layers iterations of algorithms that learn through massive datasets and infers solutions (by reducing error - goal seek) to a framed problem. If the data-set changes the error rate increases massively; if the problem is not properly framed the solution is useless. AI can shorten the design cycle, but it can't change the rules of the game (yet).
Posted on Reply
#25
thesmokingman
Hood, post: 3697093, member: 110365"
Perhaps the art of CPU design is getting too complicated, heading beyond the point where human brains can conceive meaningful and cost effective designs. The next truly new architecture might be designed by AI, a super-computer that can run simulations of all possible materials, foundry methods, and architectures, until it finds a breakthrough design. Skynet approaches...
Nooo, automated design cannot replace humans. We have see this lesson played out with AMD's Bulldozer. Automating the tuning aspect of silicon does not pay dividends. And in AMD's case they found that automated design resulted in 20% larger and 20% slower designs as compared with human designs. When AMD moved to automated design which is more common in gpu design, it was at a time when their management were making seriously suspect moves. They were reducing their engineer force and replacing them with automated design. And the rest is history and we all know how AMD flatlined after that.

http://news.softpedia.com/news/Ex-AMD-Engineer-Blames-Bulldozer-s-Low-Performance-on-Lack-of-Fine-Tuning-227816.shtml
Posted on Reply
Add your own comment