August 7, 2018 By Joseph P. Farrell

The Defense Advanced Research Projects Agency, or DARPA, or as we like to refer to it here, following a suggestion of Mr. J.B., the Diabolically Apocalyptic Research Projects Agency, has a plan to "reinvent" electronics, and specifically, the computer chip, and even to reinvent the way chip innovations are done. Mr. R.M. shared this story by Martin Giles from the MIT Technology Review, and there's something in it that caught my eye, particularly in view of yesterday's end-of-the-twig high octane speculation about steam locomotives and Union Pacific's "heritage fleet" of steam engines.

Here's the story:

DARPA has an ambitious $1.5 billion plan to reinvent electronics

There's the usual DARPA boilerplate about not only engineering, but engineering how engineering is done:

One project aims to radically reduce the time it takes to create a new chip design, from years or months to just a day, by automating the process with machine learning and other tools so that even relatively inexperienced users can create high-quality designs.

“No one yet knows how to get a new chip design completed in 24 hours safely without human intervention,” says Andrew Kahng of the University of California, San Diego, who’s leading one of the teams involved. “This is a fundamentally new approach we’re developing.”

“We’re trying to engineer the craft brewing revolution in electronics,” says William Chappell, the head of the DARPA office that manages the ERI program. The agency hopes that the automated design tools will inspire smaller companies without the resources of giant chip makers, just as specialized brewers in the US have innovated alongside the beer industry’s giants.

But then there was something that caught my eye, implying that DARPA wants an entirely new type of chip architecture:

Another ERI project will explore ways in which novel circuit integration schemes can eliminate, or at least greatly reduce, the need to shift data around. The ultimate goal is to effectively embed computing power in memory, which could lead to dramatic increases in performance.

On the chip architecture front, DARPA wants to create hardware and software that can be reconfigured in real time to handle more general tasks or specialized ones such as specific artificial-intelligence applications. Today, multiple chips are needed, driving up complexity and cost.

Some of DARPA’s efforts overlap with areas already being worked on extensively in industry. An example is a project to develop 3-D system-on-chip technology, which aims to extend Moore’s Law by using new materials such as carbon nanotubes, and smarter ways of stacking and partitioning electronic circuits. Chappell acknowledges the overlap, but he says the agency’s own work is “probably the biggest effort to make [the approach] real.” (Emphasis added)

Yesterday, you'll recall, I indulged in some really high octane speculations about typewriters and the ultimate in analog locomotive technology, the steam engine. My concern was couched in terms of "how would one haul long distance heavy freight in the event of an EMP attack on North America?" The answer, obviously, wasn't diesel-electric locomotives, but steam locomotives. The fly in that speculative ointment was not in the locomotive end of the spectrum, but in the railroad signalling and switching end, since most of this is now (and has long been) automated and therefore electrical.

So what's the relevance here? Methods of "hardening" electrical and digital systems have long been of concern to the military and therefore to DARPA. The internet, for example, in the form of the old ARPANET, emerged out of concerns about maintaining command and control of the country's military forces, and particularly its nuclear and thermonuclear forces, in the event of just such a "take-down." Decentralization of computer networks, burying cables, and hardening systems against the effects of EMP were the result, and many of these were and still are classified. But the point is that most of those systems, as far as I am aware, utilize methods that are not "built in" to the circuits they are trying to protect. To draw an analogy which might make this point clearer, cables are still buried in an effort to protect them, but the cables' circuitry themselves are not altered as a means of such hardening except in a case mentioned below.

But what if one could "harden" digital systems against such EMP attack, or other systemic breakdown, not by materials science alone, but by circuit architecture and in combination with materials science? What if a way could be found to distinguish between the "noise" of an EMP attack or other electromagnetic transient, and the actual "work" being performed by a chip or chips, and thus discerning the two, redirect the energy of the transient into the ground, or, in the ultimate electronic jiu jitsu, simply capture and store that energy in capacitors? In other words, what if one discovered a way to "partition electrical circuits" to do just what I've suggested? This would, in effect, make current "hardening methods" all but obsolete, and provide a measure of security to communications and industry it currently does not have, not to mention be a huge new industry. What if computer architecture and digital systems in general could be designed to respond to such attacks by more or less instantaneous modifications to its architecture to avoid the nasty results of such attacks? Nor need one think solely about the usefulness of such a capability simply in terms of deliberate EMP attacks. Other transients might be avoided, for example, the phenomenon of exploding transformers on power grids during electrical storms and lightning strikes. To a certain extent, the power grid already has safeguards against such things, and that occurs in its architecture, which, as we know, occasionally doesn't work.

In either case, I strongly suspect that behind DARPA's latest mission about doing an end run around Moore's Law, that much more is at stake, and much more is not being said. As usual, it's being said between the lines. It's that "electrical partitions" and "self-evolving circuit architecture" that the article implies, which indicates much more is afoot than meets the eye.

See you on the flip side...