When we found out that Jim Keller has joined Tesla, we were a bit confused. He is highly skilled in processor design, and he moved to a company that does not design processors. Kind of weird, right? There are two possibilities that leap to mind: either he wanted to try something new in life, and Elon Musk hired him for his general management skills, or Tesla wants to get more involved in the production of their SoCs, possibly even designing their own.
Now Peter Bannon, who was a colleague of Jim Keller at Apple, has been hired by Tesla Motors. Chances are, the both of them were not independently interested in an abrupt career change that led them to the same company. That seems highly unlikely, to say the least. So it appears that Tesla Motors wants experienced chip designers in house. What for? We don't know. This is a lot of talent to just look over the shoulders of NVIDIA and other SoC partners, to make sure they have an upper hand in negotiation. Jim Keller is at Tesla as their “Vice-President of Autopilot Hardware Engineering.” We don't know what Peter Bannon's title will be.
And then, if Tesla Motors does get into creating their own hardware, we wonder what they will do with it. The company has a history of open development and releasing patents (etc.) into the public. That said, SoC design is a highly encumbered field, depending on what they're specifically doing, which we have no idea about.
Keller is there to develop
Keller is there to develop specialized fault tolerant CPU cores for car autopilots! So probably the same for Peter Bannon with whatever he brings to the table. We are talking about autonomous cars so those CPU cores will be on functionality isolated redundant systems that confer with each other to come to a more error free decision, the same as the redundant CPU control systems for spacecraft/aerospace needs.
There is a lot more potential for serious liability in the car autopilot market what with cars sold by the millions, and no government agency is likely going to approve any system without the utmost in fault tolerance! Nobody wants to see a 2+ ton computer controlled car with a faulty system in charge of the driving, especially at superhighway speeds next to a 18 wheeler loaded with fuel, or that very same faulty autopilot system driving by a childrens’ playground!
These CPU cores will have to be able to interface with GPUs, cameras, and hundreds sensors of all types to be able to have a complete situational awareness all while having error rates as small a possible. So imagine the very same rocket science level of redundancy/error/fault tolerance for these CPU cores, along with some extra in the hardware features not found in phones/PC/laptops!
It’s also going to take smart roads with various systems that will be able to broadcast the road ahead condition! For example, sensors in the road to sense black ICE, or pot holes and such to warn the cars about such road hazards well in advance on any high speed roadways. There are just so many variables that will need to be reported that the road systems will have to have some form of built-in intelligence in order to provide as much advanced warning of the road’s actual real-time condition, and all the automakers will have to fund, along with the government, the smart road systems to make self driving cars a reality!
Musk can not afford to leave this design process up to any supplier, as the QC and QA of many suppliers can not be as fully under Tesla Moters control! Tesla has experience with CPUs for the aerospace industry so expect that Space-X’s technology will transfer over nicely to what Musk wants for his car autopolit systems!
Edit: Tesla has experience
Edit: Tesla has experience with CPUs for the aerospace industry
to: Space-X has experience with CPUs for the aerospace industry
I suspect that Tesla wants to
I suspect that Tesla wants to do self driving cars without needing a large camera, radar, or lidar system mounted on top of it like current Google self driving cars. Limiting the detection hardware may mean more processing power is needed. It is also unclear whether current “off the self” solutions are really sufficient to meet expectations. It would be interesting to see what hardware Google is using in their vehicles. Tesla seems to be going for complete autonomy; they want the car to be able to drive even without a person in it. This may require more hardware if only to guarentee a certain level of safety.
It is also possible that amount of power that the “off the self” solutions may take is a consideration. If you have a giant gasoline engine, then running a couple hundred watt self driving system may not be that much of an issue. The batteries on a current Tesla sedan are quite large, but it would still be a sizable power draw, mostly because the electric drive systems are so much more efficient. Turning on air conditioning or heat significantly reduces range. How much of a hit will it be for a fully autonomous self driving system? What they have now is more of just an advanced cruise control; it probably doesn’t take much power. Even if it is a small hit for a Tesla Model S, how big of an issue will it be for a Model 3? The Model 3 will be a lighter vehicle with a smaller battery.
Keller has been a chief architect, but he has also been in more of a management position at several companies. There is still no garantee that Tesla is planning on designing their own processors and such. It could just be a management position. Keller’s title isn’t chief architect, it is Vice President. You can’t really tell what is going on internally though. Another possibility is that they may be planning on using FPGAs in self driving systems due to the ability to almost completely reconfigure the device. This would allow you to get product on the road, and upgrade the abilities later. FPGAs can have parallel abilities close to GPUs. Programming an FPGA isn’t too far from hardware design. They are mostly programmed using hardware description languages, which is the first stages of almost any chip design. Keller’s knowledge and connections may be useful, if only to hire the right engineers to get the job done.
“Off the self” -> “off the
“Off the self” -> “off the shelf”
Posting from an iPhone 5 on this board is hard.