News [intel] Jim Keller resigns from Intel

[misuspita]

I'd love Amd ))

 

[DrMrLordX]

So regarding Jim Keller, where will he land next? Qualcomm? Back to Apple?

Who knows? Wherever he finds something that interests him. Nuvia maybe?

 

[Thibsie]

Who knows? Wherever he finds something that interests him. Nuvia maybe?

Yep. Nuvia came to my mind immediately when the news of him resigning from Intel broke.
That would be a hell of a dream team there.

 

[A///]

I was told by a source that despite his personal family issue, Mr. Keller has been hired by a company already. I'm lukewarm to this rumor because I'm not 100% positive what I should make of it given the severity of this family issue juxtaposed to his abrupt exit from Intel. Nuvia does seem likely.

 

[jpiniero]

I was told by a source that despite his personal family issue, Mr. Keller has been hired by a company already. I'm lukewarm to this rumor because I'm not 100% positive what I should make of it given the severity of this family issue juxtaposed to his abrupt exit from Intel. Nuvia does seem likely.

Page 6 - News - [intel] Jim Keller resigns from Intel

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

 

[A///]

Page 6 - News - [intel] Jim Keller resigns from Intel

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

Charlie is not my source and I knew before Charlie's announcement. Also hey, what's up?

 

[bononos]

Whats the timeframe for Intel to release their new gen Keller product? 2022? Its probably going to be very good going by his track record.

 

[Atari2600]

Whats the timeframe for Intel to release their new gen Keller product? 2022? Its probably going to be very good going by his track record.

Track record goes out the window when the intended process all the design assumptions and tradeoffs were based on is not available.

 

[DrMrLordX]

Whats the timeframe for Intel to release their new gen Keller product? 2022? Its probably going to be very good going by his track record.

It's still not 100% clear what exactly Keller did while at Intel. Some say he was more-involved in reorganization than working on a particular product. If he did have a hands-on approach with any of the designs, it would be with either Meteor Lake or whatever comes after Meteor Lake. Meteor Lake was supposed to be on Intel 7nm. Now all that is in doubt. 2023 at the earliest.

 

[name99]

It's still not 100% clear what exactly Keller did while at Intel. Some say he was more-involved in reorganization than working on a particular product. If he did have a hands-on approach with any of the designs, it would be with either Meteor Lake or whatever comes after Meteor Lake. Meteor Lake was supposed to be on Intel 7nm. Now all that is in doubt. 2023 at the earliest.

Read this thread:

https://twitter.com/i/web/status/1288402693770231809

and you can get a hint.

Probably brought in to fix the various managerial level issues described:
- too much emphasis on frequency and hand-crafted circuits
- too much baggage in design, with no interest in dropping any of it
- balkanized design flow because official pronouncements (unifications and splittings) never actually matched what lower product groups actually needed
- no communication between process design and processor design

Presumably (apparently) left in disgust that, even as the ship was sinking, too many people were unwilling to concede/accept that things needed to change.
(I see the same thing among Intel fans. They might concede that things don't look great, but the moment you suggest a single modification from the current way Intel does things, they'll find a million excuses not to make the change...)

 

[Doug S]

Thought part 1 was so good I had to look up the rest.

Part 2:

https://twitter.com/i/web/status/1288402697536720897

Part 3:

https://twitter.com/i/web/status/1291664039257333760

 

[clemsyn]

Thought part 1 was so good I had to look up the rest.

Part 2:

https://twitter.com/i/web/status/1288402697536720897

Part 3:

https://twitter.com/i/web/status/1291664039257333760

Interesting read. I hope Intel gets back on its feet but it will be awhile since AMD is firing on all cylinders and I am not sure if Jim Keller will consider returning.

 

[soresu]

Interesting read. I hope Intel gets back on its feet but it will be awhile since AMD is firing on all cylinders and I am not sure if Jim Keller will consider returning.

It will be interesting to see if he makes another stint at AMD though.

 

[bononos]

It will be interesting to see if he makes another stint at AMD though.

Theres probably some non-compete period before he can work for AMD after quitting from Intel.

 

[soresu]

Theres probably some non-compete period before he can work for AMD after quitting from Intel.

If that were true he wouldn't be able to work anywhere after Intel?

I doubt he would have signed on with such a specific non compete clause.

Given Koduri jumped right over to Intel it would be especially hypocritical of them.

 

[tamz_msc]

Read this thread:

https://twitter.com/i/web/status/1288402693770231809

and you can get a hint.

Probably brought in to fix the various managerial level issues described:
- too much emphasis on frequency and hand-crafted circuits
- too much baggage in design, with no interest in dropping any of it
- balkanized design flow because official pronouncements (unifications and splittings) never actually matched what lower product groups actually needed
- no communication between process design and processor design

Presumably (apparently) left in disgust that, even as the ship was sinking, too many people were unwilling to concede/accept that things needed to change.
(I see the same thing among Intel fans. They might concede that things don't look great, but the moment you suggest a single modification from the current way Intel does things, they'll find a million excuses not to make the change...)

The part about the Graphics Team working with TMG on PPA characteristics while the CPU team was faltering does not line up with the facts - it was after all the graphics portion of the chip which was borked in Cannon Lake.

 

[dmens]

The part about the Graphics Team working with TMG on PPA characteristics while the CPU team was faltering does not line up with the facts - it was after all the graphics portion of the chip which was borked in Cannon Lake.

Nah, that is just because they needed to release something/anything to pretend 10nm is healthy and GPU functionality is not necessary for that. The guy is legit, his knowledge of the CAD situation at Intel is spot-on.

Now imagine that level of dysfunction also at the design and verification domains. Especially verification, that makes the CAD/methodology situation look healthy.

 

[Thibsie]

Given Koduri jumped right over to Intel it would be especially hypocritical of them.

I may be wrong but I somehow remember that it is illegal for AMD (because the state they are based in) to use such contracts but was legal for Intel.
I dunno remember where someone explained that, it may be BS.

 

[Doug S]

I may be wrong but I somehow remember that it is illegal for AMD (because the state they are based in) to use such contracts but was legal for Intel.
I dunno remember where someone explained that, it may be BS.

Some states (like California) make such non-competes illegal, but I'm not sure if it depends on where the company's HQ is or where the employee lives/works.

Someone at his level also has more leverage in the employee/employer relationship than most of us do, so he could insist on changes to the "standard" contract of employment that is a take it or leave it deal for those of lower rank.

 

[beginner99]

The part about the Graphics Team working with TMG on PPA characteristics while the CPU team was faltering does not line up with the facts - it was after all the graphics portion of the chip which was borked in Cannon Lake.

I would argue that they needed to ship a product (due to claims made9 and shipping one without a cpu was obviously not an option. Since yield was the main issue, making skus without iGPU was the only option to get some product out I would assume.

 

[soresu]

I may be wrong but I somehow remember that it is illegal for AMD (because the state they are based in) to use such contracts but was legal for Intel.
I dunno remember where someone explained that, it may be BS.

Pretty sure that they are both based in California?

AMD has several locations in the US, but the one most mentioned is Santa Clara - and according to Google so is Intel.

SC is right smack in the middle of Silicon Valley, so they should definitely be under the same state law, and likely more or less the same county law too if they even differ.

 

[soresu]

Someone at his level also has more leverage in the employee/employer relationship than most of us do, so he could insist on changes to the "standard" contract of employment that is a take it or leave it deal for those of lower rank.

To a point maybe, though more likely any leverage would be over wage negotiation and/or stock options - that latter of which he would likely have forfeited upon leaving the company early.

From a lawyers perspective of a company like Intel it would be better to lose a bit of money/stock than to live with the consequence of someone like Keller getting tired of the corporate bureaucracy at Intel and jumping straight to AMD.

At a time like this Keller going back to AMD would be the sucker punch that really twists the knife after the 7nm delay announcement only added to their current problems.

Having said that, whatever the potential impact of such an action, I don't think AMD need him at the moment considering how well it looks like he left the spot - depending on just how good their planning for future generations of Zen is of course.

If they run out of steam by the time Intel gets back up to full throttle they might just hire him back for another handful of years.

I do wonder just how much growth potential Zen has, and whether AMD has an exit/successor plan for it - especially given how much of a bad move that Bulldozer was, though I think the likelihood of them going for such a different/risky design again is minimal to non existent.

 

[Atari2600]

Sometimes people like that are best motivated by asking them back in and telling them to "go design something you think is worth making" - the only requirement we give you is that it fits into our ecosystem, either via socket, IF, PCIe, CXL, DRAMbus, NVMe or ANOther.

(and if it is a CPU, that it support x86 instructions - either directly or via emulation - up to and including AVX2)

Letting folks with deep technical backgrounds and experience out of the box can lead to "interesting" results.

 

[Kenmitch]

I do wonder just how much growth potential Zen has, and whether AMD has an exit/successor plan for it - especially given how much of a bad move that Bulldozer was, though I think the likelihood of them going for such a different/risky design again is minimal to non existent.

It looks like they have some big/little plans for something? Would be interesting if Jim came back to AMD to help them out.

AMD Files Patent for big.LITTLE-esque Hybrid Computing Technique

Big performance, low power

www.tomshardware.com

I'd imagine he'll choose wisely next time he accepts a job. He's probably got all the money he needs already anyways. If he held on to AMD stock he's probably rolling in the $'s as we type.

 

[soresu]

I'd imagine he'll choose wisely next time he accepts a job. He's probably got all the money he needs already anyways. If he held on to AMD stock he's probably rolling in the $'s as we type.

I think the generalism is that people at his level make way more than the average person needs to live on comfortably for the rest of their lives.

It seems fairly clear that they continue to work for the prestige and/or challenge of any given job.

They say that people with nothing left to do die faster and more miserably, so there's that also.