You are here : Home > INDUSTRIAL INNOVATION > MARKETS > Advanced Computing

Article | Articles & files


Advanced computing



Advanced Computing, formerly linked to multiphysics simulation, underpins new consumer services that require Big Data and Data Analytics for processing large data volumes. Energy efficiency for a constant energy envelope, low latency data access, real-time computing and reliability are essential, while maintaining a low cost ownership. Today’s ultra-low-power computing node, server, and micro-server architectures and cutting-edge computing paradigms demand advanced technologies and a suitable software environment.
Published on 6 February 2024

 

Meeting computing challenges in a connected world

Exponential growth in personal and object information has prompted a digital data explosion and a need for fast data processing.

Advanced Computing, formerly linked to multiphysics simulation, underpins new consumer services that require Big Data and Data Analytics for processing large data volumes. Energy efficiency for a constant energy envelope, low latency data access, real-time computing and reliability are essential, while maintaining a low cost ownership. Today’s ultra-low-power computing node, server, and micro-server architectures and cutting-edge computing paradigms demand advanced technologies and a suitable software environment.

Today, Leti capitalizes on its silicon and 3D integration technologies, many-core architecture and embedded software to improve energy efficiency and facilitate compute node scaling. Leti investigates and develops energy-efficient, high-performance architectures, design solutions and technologies for servers and micro-servers. Design and manufacturing performance are there by enhanced at minimum cost. 

By 2020, Leti technologies will have further raised compute node energy efficiency by introducing optical transmission and non-volatile memories close to the computer. 
CoolCubeTM 3D sequential technology and high-density integration will also increase compute density. System-on-chip neural networks will be developed for accelerating specific applications.