This Computer Is Tiny but Mighty

Photo: Getty Images/krystiannawrocki

You’ve heard about the Internet of Things. Picture miniature computers aiding medical devices, automating factories, watering crops, monitoring infrastructure—all seamlessly integrated with other systems. 

Image of chip compared to the head of a pencil eraser, and an enlargement of the chip.

Now Stanford electrical engineers and computer scientists have demonstrated a powerful computer-on-a-chip that can serve as the foundation for this everywhere computing. The team, led by professors Subhasish Mitra and H.-S. Philip Wong and assistant professor Mary Wootters, worked alongside scientists from the CEA-LETI research institute in Grenoble, France, to design a functional prototype. It’s a proof of concept for RRAM (resistive random access memory), a new data storage technology that’s more compact, more energy efficient, faster and more reliable than its predecessor options.

Illustrated cross-section of the chip showing thick metals, the memory stack, transistor and silicon.


What’s so special about this chip?


It’s tiny. Engineers were able to build RRAM on top of a processing circuit, so you get data storage and computing in one chip.

It works. CEA-LETI scientists were able for the first time to integrate RRAM with com­mer­cial silicon, making the breakthrough scalable. It provides 10 times better energy efficiency for the same processing speed.

It’s got a great memory. To make RRAM store more data, the team invented a way to accommodate five values—not just zeros and ones—into each memory cell.

RRAM can go long. When memory cells wear out, they scramble data and make mistakes. The team came up with an algorithm that constantly shuffles which memory cells are written on, spreading the wear and tear of data storage across the chip’s 150,000 memory cells.

This chip gets to the point. It’s self-contained, can be set up on-site and sustains itself for an estimated 10 years.


Jill Patton, ’03, MA ’04, is the senior editor for Stanford.