This site may earn affiliate commissions from the links on this page. Terms of use.

Silicon wafer

It's no undercover that general-purpose compute performance has been largely stuck in idle these last few years, with pinnacle-end improvements limited to unmarried-digit increases. DARPA has a program for revitalizing CPU performance by fundamentally changing how we design chips — and it's an approach that could yield real benefits in at least some areas.

One DARPA concept calls for an overhaul of how we recollect about CPU architectures. It'due south related to a concept we discussed v years ago, in which specific applications could actually be implemented in hardware, potentially offering tremendous performance improvements. It'south besides a partial respond to the problem of dark silicon. Moore's Law has increased transistor density, only without Dennard Scaling we don't have the ability to plow all of that silicon on at once. One mode to use information technology effectively is to map very specific functions to information technology that would typically run in the groundwork, only could benefit from running extremely efficiently.

Other architectural concepts existence explored include software-divers hardware — hardware that can reconfigure itself at fine-grained timescales to optimally execute workloads based on its analysis of the type of piece of work yous're doing. The other, dubbed "domain specific system on chip," is an initiative to explore what kinds of common building blocks would need to be present on an SoC to enable constructive, flexible, on-the-fly reconfiguration — for, say, radio communication. Are specialized accelerator blocks required, or tin can the work be done in other ways?

GreenDroid

The GreenDroid compages was designed to use dedicated, specialized co-processors for many tasks.

Only CPU architectures aren't the only place where DARPA wants to reinvent computing. As IEEE Spectrum details, its new Arts and crafts initiative wants to empower pocket-size design teams to build much larger hardware, and to slash the difficulty of moving to a new node past as much as 80 per centum. In that location's also a ready of semiconductor design programs, Idea and Posh, that are attempting to capture the intersection between Electronic Blueprint Automation (EDA) and car intelligence. Ideally, as you design SoCs, your design software would itself get more intelligent and get better at its job. Posh, meanwhile, is focused on building open up source hardware and ensuring a foundation of building blocks is created that can exist the ground for any future projection.

There's even more than details of how these new initiatives fit into existing IEEE Spectrum projects and how the organization hopes to kickstart research into areas of computing. The article is well worth reading in its entirety. Most of the coverage y'all run into in semiconductors these days often grapples with how engineers are fighting an always-more difficult battle to wring less and less functioning out of silicon. It makes for grim reading on the whole and the twenty-four hour period-to-day developments don't practise much to modify that.

Initiatives like these emphasize that in that location are all the same ideas for how to ameliorate compute performance. They're still heavily caveated — information technology's not at all clear, for instance, that anyone will e'er build a CPU that actually checks email or performs awarding functions in hardware. Simply the try to build software and hardware that piece of work together more intimately to better functioning is probably a primal method we'll use in the future.

DARPA has previously stated that from 1978 to the nowadays solar day, CPU architectural improvements account for a 50x improvement in CPU performance, compared with the 1,000x comeback delivered by increased CPU clocks. With CPU clocks stuck below 5GHz and no known well-nigh-term improvements expected on that front, finding ways to employ the clock speed and configuration options we have more than effectively is the only fashion we'll keep improving CPU operation.