Monday, 11 January 2021

Engineers create hybrid chips with processors and memory to run AI on battery-powered devices

Transactions between processors and memory can consume 95 percent of the energy needed to do machine learning and AI, which severely limits battery life. A team of engineers has designed a system that can run AI tasks faster, and with less energy, by harnessing eight hybrid chips, each with its own data processor built right next to its own memory storage.
at
Labels:

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)

How 'clean' does a quantum computing test facility need to be?

How to keep stray radiation from 'shorting' superconducting qubits; a pair of studies shows where ionizing radiation is lurking and ...

  • Arduino MCP2515 CAN Bus Interface Tutorial
    In this project, we will learn about the MCP2515 CAN Controller Module, how to interface the MCP2515 CAN Bus Controller with Arduino and fin...
  • Logic Gates And How They Work
    I was first introduced to logic gates when I was around 14 years old. I had heard that computers consisted of ones and zeroes. But I didn’t...
  • MOSFET Gate Resistor
    Do you need a MOSFET gate resistor? What value should it be? And should it go before or after the pulldown resistor? If you’re a bit impati...