September Update: Cruise, CHIPS, FP8, and McLaren

Wow, September has flown by. It's already the last Friday of the month, the last day of the month in fact, and so time for a monthly update.

Kaufman Award

The 2022 Kaufman Award honors Giovanni (Nanni) De Micheli of École Polytechnique Fédérale de Lausanne (EPFL) in Lausanne, Switzerland. This is not really an update of course, the news was only announced a couple of days ago. I will interview him in the coming days and write a full post, probably next week.

Cruise Designing Their Own Chips

When I rejoined Cadence in 2015, I wrote a couple of posts about what, at the time, we called System Design Enablement (SDE). In my post The SDE Circle of Life Today, I wrote about the automotive industry:

Do they build their own silicon so that they can deliver a unique driving experience? My prediction: the leaders will design their own silicon.

That prediction was in 2017, so five years ago. In 2019, Tesla announced that they were doing just that. See my post Tesla Drives into Chip Design. Since then...crickets. But a couple of weeks ago, Cruise (owned by GM) announced that they had designed a family of chips. See this article from Reuters Upset by high prices, GM's Cruise develops its own chips for self-driving cars. It seems there are four chips:

Cruise had developed four in-house chips so far - a computing chip called Horta, the main brains of the car, Dune, which processes data from the sensors, a chip for the radar, and one that it would announce later, Jenkins said.

It has all been happening more slowly than I expected, but I continue to expect more and more car companies to design their own chips. You can't be "the ultimate driving machine" if you use the same chips as Volkswagen or Ford.

NHTSA Cyber Update

Sticking with automotive, NHTSA issued NHTSA Updates Cybersecurity Best Practices for New Vehicles. The actual document is Cybersecurity Best Practices for the Safety of Modern Vehicles (Updated 2022). If you have anything to do with automotive security (meaning if you have anything to do with automotive, period, since safety and security are intertwined), then you need to read this document.

CHIPS

Details have started to emerge about how the CHIPS act will be implemented. See the government press release Biden Administration Releases Implementation Strategy for $50 Billion CHIPS for America program. There is still a lack of real detail, but the three main goals will be:

The program supports three distinct initiatives:

• Large-scale investments in leading-edge manufacturing: The CHIPS incentives program will target approximately three-quarters of the incentives funding, around $28 billion, to establish domestic production of leading-edge logic and memory chips that require the most sophisticated manufacturing processes available today.
• New manufacturing capacity for mature and current-generation chips, new and specialty technologies, and for semiconductor industry suppliers: The CHIPS incentives program will increase domestic production of semiconductors across a range of nodes, including chips used in defense and in critical commercial sectors such as automobiles, information and communications technology, and medical devices.
• Initiatives to strengthen US leadership in R&D: The CHIPS R&D program will invest $11 billion in a National Semiconductor Technology Center, a National Advanced Packaging Manufacturing Program, up to three new Manufacturing USA Institutes, and in NIST metrology research and development programs.

If you want to look at the details of all the fabs under construction (or planned) in the US, see Here’s the latest hot list of cool chip fabs coming to the US.

Eight-bit Floating Point

In my posts about this year's HOT CHIPS, I wrote about how some of the latest training and inference chips are using eight-bit floating point (FP8), of which there are two or three different flavors depending on how the exponent and mantissa are split. See my post HOT CHIPS Day 2: AI...and More Hot Chiplets. Now there is a new paper FP8 Formats for Deep Learning published by engineers at NVIDIA, Arm, and Intel seeking to standardize these formats so that they can be implemented in hardware.

AI Images and Getty

You've probably seen some of those amazing AI-generated images. If not, see my post What Is a Lagrange Point in Space? And DALL·E 2. Getty images have banned AI-generated artwork since it is unclear if it can be copyrighted or not. I haven't thought it all through, but it seems silly if such images cannot be copyrighted. For a deeper dive, see.Getty Images bans AI-generated content over fears of legal challenges.

McLaren

Cadence has a technology partnership with McLaren racing. See my post Cadence Shifts into High Gear with the McLaren Formula 1 Team. We recently announced Verisium (see my post Verisium AI-Driven Verification Platform). The day before, we had an invitation-only event at Levi Stadium to pre-announce the product under embargo. One of the speakers was Zak Brown, CEO of McLaren racing. We also recently produced a book on automotive (see my post Books: Hyperscale and Automotive), so what an opportunity to get a photo of Zak holding my book:

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.

.