Ryzen Architect Jim Keller Joins Intel

[R0H1T]

Well, that's shit. Everyone worth anything leaving from AMD to join Intel...

Except the boss lady, also their current CTO, BK looks like a joke as compared to her.

 

[bug]

First Raja, now Jim. What?

Jim didn't leave AMD to join Intel. He went to Tesla first. Plus, he works as a consultant, why would his job changes be "shit"?

 

[R0H1T]

Jim didn't leave AMD to join Intel. He went to Tesla first. Plus, he works as a consultant, why would his job changes be "shit"?

You're taking his comment out of context & probably a bit too literally, FYI he said that's sh!t perhaps in the context of this being bad for AMD & IMO rightfully so.

 

[sutyi]

First Raja, now Jim. What?

Well if I recall correctly, Mr. Keller himself stated he would be only joining AMD for Zen and then bugger off if it looks good. So near launch he went on his marry way to Tesla and now he is back at Intel again. Would have been nice to see him stay for a couple of more years, but apparently he values new and interesting challanges and developement opposed to long(er) term "polishing" type projects. Intel is probably is planing to do interesting iGPU work with Mr. Coduri so Mr. Keller went there after the Tesla gig.

On the RTG front I'm not really sad that Mr. Coduri left the helm. There should've been fork in architecture at RTG some years back. GCN based Polaris and Vega cards are just simply not efficient enough on the gaming front and I mean that in manufacturing and power consumption fashion.

Hopefully Lisa Su can shake up RTG the sameway as they did on the CPU front.

 

[R0H1T]

Well if I recall correctly, Mr. Keller himself stated he would be only joining AMD for Zen and then bugger off if it looks good. So near launch he went on his marry way to Tesla and now he is back at Intel again. Would have been nice to see him stay for a couple of more years, but apparently he values new and interesting challanges and developement opposed to long(er) term "polishing" type projects. Intel is probably is planing to do interesting iGPU work with Mr. Coduri so Mr. Keller went there after the Tesla gig.

On the RTG front I'm not really sad that Mr. Coduri left the helm. There should've been fork in architecture at RTG some years back. GCN based Polaris and Vega cards are just simply not efficient enough on the gaming front and I mean that in manufacturing and power consumption fashion.

Hopefully Lisa Su can shake up RTG the sameway as they did on the CPU front.

Tbh I don't think Keller has as much to do with Zen as many of us think he does, he left AMD what 3 years back & in that span Zen (design) must have been updated a few more times. The current team lead by Mark Papermaster(?) deserve a lot more credit.

 

[bug]

You're taking his comment out of context & probably a bit too literally, FYI he said that's sh!t perhaps in the context of this being bad for AMD & IMO rightfully so.

I'm taking his comment out of context and you're guessing what he meant. What could possibly go wrong?

I just said I don't understand why he put Jim and Raja in the same basket only to conclude this is "shit". Raja could be seen as an AMD guy, but Jim is very different from a career point of view.

Tbh I don't think Keller has as much to do with Zen as many of us think he does, he left AMD what 3 or 4 years back & in that span Zen (design) must have been updated a few times. The current team lead by Mark Papermaster(?) deserve a lot more credit.

Jim left AMD exactly when his work on Zen ended. He only designs this stuff, implementation, tweaking and manufacturing aren't part of his job descriptions.

 

[sutyi]

Tbh I don't think Keller has as much to do with Zen as many of us think he does, he left AMD what 3 years back & in that span Zen (design) must have been updated a few more times. The current team lead by Mark Papermaster(?) deserve a lot more credit.

AFAIK Jim Keller and Micheal Clark were team lead on ZEN from 2012 and Keller bailed around fall (?) 2015.

 

[R0H1T]

I'm taking his comment out of context and you're guessing what he meant. What could possibly go wrong?

I just said I don't understand why he put Jim and Raja in the same basket only to conclude this is "shit". Raja could be seen as an AMD guy, but Jim is very different from a career point of view.

Well, that's shit. Everyone worth anything leaving from AMD to join Intel...

I don't think I am but let's leave it at that.

Jim left AMD exactly when his work on Zen ended. He only designs this stuff, implementation, tweaking and manufacturing aren't part of his job descriptions.

Yes but again I wouldn't credit Jim anymore than Mark, for Zen.

 

[trparky]

Well there goes the neighborhood.

 

[bug]

I don't think I am but let's leave it at that.

Like I said, even that doesn't hold, because Jim didn't leave AMD for Intel. He is now with Intel, but he left for Tesla. This goes back to my original question of putting Jim and Raja in the same basket. But ok.

Yes but again I wouldn't credit Jim anymore than Mark, for Zen.

You're free to credit whoever you want, but it's Jim that has a proven track record of designing CPUs having great perf/W ratios. I'm sure Mark wasn't just pushing papers around, but his resume is not as closely tied to CPU design as Jim's.

 

[ssdpro]

Would have been nice to see him stay for a couple of more years, but apparently he values new and interesting challanges and developement opposed to long(er) term "polishing" type projects. Intel is probably is planing to do interesting iGPU work with Mr. Coduri so Mr. Keller went there after the Tesla gig.

If that premise is true it looks like AMD is set for another 10 years of stagnation.

 

[john_]

Zentel architecture is coming.

 

[bug]

This is good news because it means Intel sees AMD as a threat again. I'm hoping that as a result, we see something surprising from Intel in a few years.

I wouldn't infer Intel sees AMD as a threat from this move.
My takeaway is Intel's in-house engineering team was unable to come up with a significantly better x86 implementation on their own, so they had to hire some help. Either that or Intel could be working on something new entirely and is building a new team to do it.

 

[HD64G]

This is good news because it means Intel sees AMD as a threat again. I'm hoping that as a result, we see something surprising from Intel in a few years.

Exactly! Tables are turning now and Intel needs new arch to compete with upcoming Zen 2 on 7nm. We might see then AMD on the top until Jim's core is launched and polished, that means for at least 2-3 years from 2019 and on. And Intel fanboys cannot underestimate Ryzen cpus now, just because of Intel hiring the father of this arch to make a better one for them. That proves that they are a bit behind in ideas atm. Big news imho.

 

[Slizzo]

I think a couple of you are putting way too much stock into what exactly Jim did at AMD on Zen. Yeah, he was lead architect, but be aware that the rest of the design team is still there, and it's not like they're not talented folk. Let's see how Zen2 shapes up before we make sweeping generalizations, eh?

And Raja is talented, yes. But we'll see how RTG does with what they have going now.

 

[HD64G]

I think a couple of you are putting way too much stock into what exactly Jim did at AMD on Zen. Yeah, he was lead architect, but be aware that the rest of the design team is still there, and it's not like they're not talented folk. Let's see how Zen2 shapes up before we make sweeping generalizations, eh?

And Raja is talented, yes. But we'll see how RTG does with what they have going now.

You have to understand though but if Jim is hired from Intel to lead their next CPU arch disigning team, thinks can be interpreted only in one way and I described it in my previous post. Not that you didn't speak the truth in yours but such facts cannot be ignored either.

 

[efikkan]

I seriously doubt this is gong to matter at all. Large companies seem to "recycle" management at this level, and even inside companies this kind of people seem to be shuffled around constantly, and they rarely even touch the technical details of the projects they lead. Both AMD, Intel and Nvidia have lost key people in project management the last few years, but what really matters is the focus, direction and cooperation of the core team developing the products.

 

[Vya Domus]

I think a couple of you are putting way too much stock into what exactly Jim did at AMD on Zen. Yeah, he was lead architect, but be aware that the rest of the design team is still there, and it's not like they're not talented folk. Let's see how Zen2 shapes up before we make sweeping generalizations, eh?

And Raja is talented, yes. But we'll see how RTG does with what they have going now.

Well , that's how people think. They feel the need to put a face on the success/failure of a company. Of course, both Keller and Koduri did little in the grand scheme of things but regardless people want to project expectations of miracles onto their departure/hiring.

 

[bug]

Well , that's how people think. They feel the need to put a face on the success/failure of a company. Of course, both Keller and Koduri did little in the grand scheme of things but regardless people want to project expectations of miracles onto their departure/hiring.

Well, if the rest of the team was so capable, then why was Jim brought on board? Obviously the guy has unique knowledge and even if he doesn't tackle everything himself, he is at least needed to assess the status quo and come up with a plan for getting from point A to point B. Doesn't make the rest of the team any less relevant, but it does put him in a unique position.

 

[BiggieShady]

No Jimmy, No! Bad boy!

 

[Paganstomp]

This is how it really works guys.

 

[Fluffmeister]

View attachment 100263
This is how it really works guys.

Business 101. But GPP still = tears.

 

[Xaled]

If you can't beat them, hire them.

 

[eidairaman1]

Well if I recall correctly, Mr. Keller himself stated he would be only joining AMD for Zen and then bugger off if it looks good. So near launch he went on his marry way to Tesla and now he is back at Intel again. Would have been nice to see him stay for a couple of more years, but apparently he values new and interesting challanges and developement opposed to long(er) term "polishing" type projects. Intel is probably is planing to do interesting iGPU work with Mr. Coduri so Mr. Keller went there after the Tesla gig.

On the RTG front I'm not really sad that Mr. Coduri left the helm. There should've been fork in architecture at RTG some years back. GCN based Polaris and Vega cards are just simply not efficient enough on the gaming front and I mean that in manufacturing and power consumption fashion.

Hopefully Lisa Su can shake up RTG the sameway as they did on the CPU front.

On the IGP front Vega does well, I think a Vega 32 would be something plausible. But yes GCN is a general purpose part, it doesn't master anything today other than being a mining die unlike 2013s inception of it.


Jim Keller only stuck around for Ryzen that was a part of his contract so as soon as his contract was up he left. He is going to move around the industry wherever the money is at that point of time so obviously it doesn't matter where he goes

 

[sergionography]

I wouldn't infer Intel sees AMD as a threat from this move.
My takeaway is Intel's in-house engineering team was unable to come up with a significantly better x86 implementation on their own, so they had to hire some help. Either that or Intel could be working on something new entirely and is building a new team to do it.

From the news i read, jim keller wont be involved in a new x86 architecture and is rather going to be an soc engineer. Intel as an soc maker is rather behind the industry as it doesnt have a solid standardized interconnect for building socs. AMD has infinity fabric which is probably as big of an achievement they made as the zen core itself. It means they can connect ip building blocks on a standard fabric that expedites design time tremendously, lowers cost, and paves the way for multi chip module implementation. This right now is perhaps more important than die shrinks moving forward.