Eta Compute has announced its revolutionary neuromorphic computing System on Chip (SoC) for enabling ultra-low-power machine intelligence at the network edge. The SoC’s integrated DSP processor and controller architecture enables embedded machine intelligence for a wide range of applications. These include smart sensors, image processing, motion detection, and speech and video applications.
Utilising the TSMC 55-nm ULP process, the device is said to consume a fraction of the power of existing solutions and redefines the standard for ultra-low power embedded solutions.
“Our patented event-driven processor architecture (DIAL) is combined with our fully customizable neuromorphic algorithms,” said Nara Srinivasa Ph. D., CTO of Eta Compute. “These will be the foundation of a diverse and wide-ranging set of applications that deliver machine intelligence to the network edge.”
The new 55nm IP portfolio includes:
- The Arm Cortex-M3 processor, which consumes as little as 1uW while scaling up to 100MHz using our patented dynamic voltage scaling (DVS)
- NXP CoolFlux DSP operating on a variable workload based on DVS
12 bit SAR Analog-to-Digital converter that consumes only 1uW at 200ksps
- Power management and voltage references optimized to deliver high-efficiency voltage scaling
- Support for low power analog blocks such as power-on reset, brown out detector, oscillators, temp sensor, crystal oscillator, and RC oscillators
“With its rich machine learning software support, ease of programming and widest deployment, Arm Cortex-M processors are enabling applications that deliver machine learning at the edge,” said Thomas Ensergueix, senior director, Embedded and Automotive Line of Business, Arm. “Eta Compute’s ultra-low power Cortex-M3 based platform will speed time-to-market of more intelligent devices powered by energy harvesting and small coin cell batteries.”
More information: http://etacompute.com/