Jim Keller has resigned from Intel

[WYP]

The chip architect has left due to "personal reasons".



Read more about Jim Keller's resignation from Intel.

 

[AlienALX]

TBH after what I've heard about the inner workings at Intel this doesn't surprise me at all.

From what I know of the guy he's a completely free spirit, does what he needs to and moves on. Fair play.

 

[DARKSF]

I thought he would bring Intel up to speed but obviously they are doomed stuck at their skylake ++++++++++ variation.Same way AMD Radeon division can't escape from their power hungry GCN architecture for 8 years already.

 

[NeverBackDown]

This isn't anything new. He does what he sets to do then leaves. Did the same with AMD.

 

[AlienALX]

I thought he would bring Intel up to speed but obviously they are doomed stuck at their skylake ++++++++++ variation.Same way AMD Radeon division can't escape from their power hungry GCN architecture for 8 years already.

He left AMD after about two years. His job was done. It then took them three more years to bring it to fruition.

That is probably exactly what he did for Intel. When you are a genius like that you don't need to stay at a boring 9-5.

 

[dihartnell]

I thought he would bring Intel up to speed but obviously they are doomed stuck at their skylake ++++++++++ variation.Same way AMD Radeon division can't escape from their power hungry GCN architecture for 8 years already.

2 years is nowhere near enough time to make changes, and then see them come to market. Keller joined AMD 2012, left 2015 and it was another 2 years later that the product actually made it to market. His handywork at Intel not likely to see light of day until 2022/23.


FYI, GCN at AMD has been replaced by RDNA, and RDNA 2 later this year so they aren't stuck on GCN

 

[AlienALX]

2 years is nowhere near enough time to make changes, and then see them come to market. Keller joined AMD 2012, left 2015 and it was another 2 years later that the product actually made it to market. His handywork at Intel not likely to see light of day until 2022/23.


FYI, GCN at AMD has been replaced by RDNA, and RDNA 2 later this year so they aren't stuck on GCN

That is exactly why they said this....

https://www.techradar.com/news/intel-admits-it-wont-catch-up-with-amds-7nm-chips-until-2021

Remember, they have a lot more money than AMD so can probably make it happen faster, but the fact is if Keller has left that means his work is done and the rest is up to father time.

 

[Kleptobot]

I have no doubt that Keller is a very talented engineer, but you can't have a single data point and call it a pattern. People have practically fetishized him as some kind of CPU architecture wizard.

 

[AlienALX]

I have no doubt that Keller is a very talented engineer, but you can't have a single data point and call it a pattern. People have practically fetishized him as some kind of CPU architecture wizard.

lol are you for real?

Career
Jim Keller worked at DEC until 1998, where he was involved in designing the Alpha 21164 and 21264 processors.[4][5] In 1998 he moved to AMD, where he worked to launch the AMD Athlon (K7) processor and was the lead architect of the AMD K8 microarchitecture,[18] which also included designing the x86-64 instruction set and HyperTransport interconnect mainly used for multiprocessor communications.[4]

In 1999, he left AMD to work at SiByte to design MIPS-based processors for 1 Gbit/s network interfaces and other devices.[5][13][19] In November 2000, SiByte was acquired by Broadcom,[20] where he continued as chief architect[10] until 2004.[4]

In 2004 he moved to serve as the Vice President of Engineering at P.A. Semi,[4][12] a company specializing in low-power mobile processors.[5] In early 2008 Keller moved to Apple. P.A. Semi was acquired by Apple shortly afterwards, reuniting Keller with his prior team from P.A. Semi.[7][18] The new team worked to design the Apple A4 and A5 system-on-a-chip mobile processors. These processors were used in several Apple products, including iPhone 4, 4S, iPad and iPad 2.

In August 2012, Jim Keller returned to AMD, where his primary task was to lead development of new generation of x86-64 and ARM microarchitectures called Zen and K12.[16][15] After years of being unable to compete with Intel in the high-end CPU market, new generation of Zen processors is hoped to restore AMD's position in the high-end x86-64 processor market.[4][14] On September 18, 2015, Keller departed from AMD to pursue other opportunities, ending his three-year employment at AMD.[21]

The guy is a genius. And has form for being a genius LMFAO hark at you going on like he's a one trick pony

 

[NeverBackDown]

I have no doubt that Keller is a very talented engineer, but you can't have a single data point and call it a pattern. People have practically fetishized him as some kind of CPU architecture wizard.

Do you not know the history of Keller? The dude has a ton of achievements.

Co-author of : x86-64
Started AMD with K7 and eventually lead to being lead architect of K8(which helped AMD immensely)
Then moved to Apple and started them on their A4/5 CPUs for all their devices and helped push Apple towards developing their own CPUs.
Went back to AMD to work on Zen and ARM stuff(AMD opted out of this though)
Then went to Tesla to help advance their AI chips which is why they don't use Nvidia anymore
Then went to Intel to fix their architecture from the ground up.


So yeah he's pretty much worked at every major technology company developing microprocessors. Intel, AMD, Apple, Tesla (being new), Broadcom and a few others. Outside of Nvidia since he doesn't seem to do Graphic Architecture he's done quite a bit. He's a renown architect for a reason.


Edit: Alien beat me to it(was typing while he posted) but I didn't copy a wiki page

 

[Kleptobot]

I am aware of his long and storied history, i was specifically referring to the 'two years and he's done' sentiment.

 

[NeverBackDown]

I am aware of his long and storied history, i was specifically referring to the 'two years and he's done' sentiment.

The way it was written could be interpreted the wrong way as demonstrated...

But he did it with Tesla, AMD, AMD back in 1999, Intel, and Broadcom.

He pretty much moved companies anywhere from 1-3 years with the exception of Apple in which he stayed for 4 years. He has multiple data points. He is the type of person who focuses on one thing then needs to do something else to stay out of boredom. Only assuming but based off his history it's plausible.

Now that he is available for hire, assuming he doesn't retire, whatever company gets him is going to definitely benefit. It will be very interesting.

 

[RobM]

I have no doubt that Keller is a very talented engineer, but you can't have a single data point and call it a pattern. People have practically fetishized him as some kind of CPU architecture wizard.

That's because he is every bit CPU wizard, pretty much anything he has touched has been brilliant

 

[grec]

I think generally people miss the fact that the difference between a good engineer and a bad engineer (Particularly in electronic engineering) often comes down to their understanding of psychology and sociology rather than science, the science is usually the easy bit everyone in the field is taught already(Since we almost always work with already accepted and proven scientific concepts, and you're not really an engineer at all if you don't understand the underpinning science of your respective field, in stark contrast to scientists who generally work with new and emerging yet to be proven or discovered concepts), the hard bit is extrapolating how the science will interact with the most complex element of its environment, the human mind.

From interviews it seems that a lot of Jim Keller's value doesn't just come in his engineering skills, which are obviously still top tier, but his team leadership and philosophy around people and corporate hierarchies in general. He is a problem solver and analyser in a way that goes far beyond designing circuits.

 

[Avet]

90/10 rule. 90% of the important work is done by up to 10% of the people. Some times by only one individual. So having someone like Jim Keller leave might create a stagnation.

 

[WYP]

I have heard from a source that a member of his family is "very sick", which explains his sudden departure.