Jim Keller, a huge name in the semiconductor industry for his work at AMD and Apple, recently left AMD before the launch of the Zen architecture. This made us nervous, because when a big name leaves a company before a product launch, it could either be that their work is complete… or they're evacuating before a stink-bomb detonates and the whole room smells like rotten eggs.
It turns out a third option is possible: Elon Musk offers you a job making autonomous vehicles. Jim Keller's job title at Tesla will be Vice President of Autopilot Hardware Engineering. I could see this position being enticing, to say the least, even if you are confident in your previous employer's upcoming product stack. It doesn't mean that AMD's Zen architecture will be either good or bad, but it nullifies the earlier predictions, when Jim Keller left AMD, at least until further notice.
We don't know who approached who, or when.
Another point of note: Tesla Motors currently uses NVIDIA Tegra SoCs in their cars, who are (obviously) competitors of Jim Keller's former employer, AMD. It sounds like Jim Keller is moving into a somewhat different role than he had at AMD and Apple, but it could be interesting if Tesla starts taking chip design in-house, to customize the chip to their specific needs, and take away responsibilities from NVIDIA.
The first time he was at AMD, he was the lead architecture of the Athlon 64 processor, and he co-authored x86-64. When he worked at Apple, he helped design the Apple A4 and A5 processors, which were the first two that Apple created in-house; the first three iPhone processors were Samsung SoCs.
I doubt Tesla would develope
I doubt Tesla would develope their own chips, although what else would they need Keller for? They could be just using off the shelf components in unique ways though. Power efficient and sufficiently powerful AI may require a different architecture rather than just throwing more CPU or GPU cores at the problem.
I feel like this is a great
I feel like this is a great move for Keller.
He not only gets a Job that is pretty damn impactful.
He is also getting something that is completely new and challenging for him.
I feel like this will be the space to watch in the coming years.
If Tesla go in house and develop there own chips, nvidia are going to have some pretty brutal competition; considering Tesla continuously shows they are willing to do whatever it takes to get a job done.
There are loads of computers
There are loads of computers on Tesla’s and other autos, so Nvidia’s Tegras are only used for the dashboard graphics and others computers serve other functions. Maybe Musk needs a special form of CPU/SOC core that is/will be made to run Tesla’s autopilot, so Jim Keller is the go to guy with plenty of experience with ARM, and x86, the Alpha 21164 and 21264 processors(at DEC he worked on the ALPHA projects), Apple A4 and A5 SOC, and Keller Vice President of Engineering at P.A. Semi, a company specializing in low-power mobile processors. P.A. Semi was acquired by Apple(1).
I do not read anything into his leaving AMD, his contract was up in designing Both AMD’s Zen(x86 ISA) and K12(custom ARMv8A ISA) based CPU cores! Keller goes where there is challenging projects, so he’s now designing a custom SOC/CPU for the Tesla Company! That moving around as an expert for hire is Keller’s M.O., nothing new about that. So Keller is head of Tesla’s CPU/Autopilot unit that’s good for Musk, Keller is good to have on anyone’s team!
All that crap about him leaving AMD for other reasons is for Basement gamers on Mom’s dime! There are plenty of mobile phone makers making their own custom ARM/MIPS/other kit, so what is so special about an auto maker doing the same, there are plenty of lead engineers working on CPU cores at scores of companies.
(1)
https://en.wikipedia.org/wiki/Jim_Keller_(engineer)
P.S. You can be Damn sure
P.S. You can be Damn sure that Musk is not going to have a car’s autopilot made anywhere but in house using open code to do the job, the US/other regulatory agencies are going to want to audit any CPU hardware, and software/firmware before approving any computer’s complete control over something that weighs tons and travels any US/other based roads, same goes for the computers that control rockets, so Musk has some IP in house already from that other business. A Car’s autopilot will be like a rocket’s autopilot with loads of redundancies!
As you say, Elon Musk prefers
As you say, Elon Musk prefers to build in-house rather than buy. Perhaps the reluctance of vendors to allow life-critical usage of their products is limiting his choices a bit – though I’m sure AMD and Apple’s patent lawyers will look very closely at any new chips Tesla produces.
Those chips that go into the
Those chips that go into the space probes are very reliable and have been tested thoroughly many times over to get any errors out of the system. The same goes for the rocket control unit CPUs, with both the space probes and the rocket CPUs being hardened for space travel against radiation and other extremes, as well as the multiple redundancies.
The car autopilot CPUs may not be as radiation hardened but they will operate in temperature extremes. Any auto based autopilot computing system will have to have redundant CPUs that talk to each other and each CPU will have agree with the others before any microsecond decisions are made! This usually requires 3 or more CPUs/SOCs running the same software and then comparing their results to protect against errors, and that includes each system having ECC memory and error correcting hardware in each CPU/SOC. It is that extra redundancy, as well as the extra error correction hardware in the CPU’s cores that will make them a special order project under Jim Keller’s design team at Tesla.
These will not be as radiation hardened as the space probe CPUs/SOCs but they will have all of the error correction and temperature hardening of those space rated SKUs, as well as the redundant error correction and fail-safe circuitry that is expected for systems like airplanes/and other aerospace systems! Expect a lot more triple redundant types of systems for car autopilots, and even more so than on airplanes and space probes with the software/hardware/firmware certifications and extra error correction circuitry, simply because these car based autopilot based systems will be in millions of cars and the chances for any design errors to show up in use will go up with systems that are in millions of passenger cars even more so than systems that are in thousands of airplanes and a few space probes/rockets.
For Tesla the car autopilot system will definitely be done in house, even with other noncritical parts being contracted out to other makers. The car autopilot systems will expose Tesla to great liability, and you can bet that under Jim Keller there will be the exact same types of CPU/SOC/software/firmware control design engineers that are used for the aerospace markets, and Musk is in that transportation market as well as the earthbound one.
I don’t know if we can assume
I don’t know if we can assume that they are going to develope their own processors and such. They could just be exploring the idea of making their own solution verses building it out of off the self components.
Well, VP is ‘not working’ on
Well, VP is ‘not working’ on design, they need qualified IC designers, system architects and verification guys. And a lot of money! A lot! And time!
But, there is a lot of room for ASIC in image processing.
He’s going to reprise his
He’s going to reprise his role as Hercules in the TV show Hercules.
THAT’S IT! That’s who he
THAT’S IT! That’s who he looks like! It’s been nagging at me for like a year, now, every time I see a picture of Jim Keller I find myself thinking, “He looks like one of those ‘That Guy’ actors…. but who?”
Now I know. Kevin Sorbo.
http://i.imgur.com/Qtk9qAF.jpg
Haha, came in to say I keep
Haha, came in to say I keep thinking he’s Kevin Sorbo. You got this covered. Well done!
If you need to know anything
If you need to know anything else about Jim Keller, just give him a shout at 867-5309.