• Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar
  • Skip to footer

Electrical Engineering News and Products

Electronics Engineering Resources, Articles, Forums, Tear Down Videos and Technical Electronics How-To's

  • Products / Components
    • Analog ICs
    • Connectors
    • Microcontrollers
    • Power Electronics
    • Sensors
    • Test and Measurement
    • Wire / Cable
  • Applications
    • Automotive/Transportation
    • Industrial
    • IoT
    • Medical
    • Telecommunications
    • Wearables
    • Wireless
  • Resources
    • DesignFast
    • Digital Issues
    • Engineering Week
    • Oscilloscope Product Finder
    • Podcasts
    • Webinars / Digital Events
    • White Papers
    • Women in Engineering
  • Videos
    • Teschler’s Teardown Videos
    • EE Videos and Interviews
  • Learning Center
    • EE Classrooms
    • Design Guides
      • WiFi & the IOT Design Guide
      • Microcontrollers Design Guide
      • State of the Art Inductors Design Guide
    • FAQs
    • Ebooks / Tech Tips
  • EE Forums
    • EDABoard.com
    • Electro-Tech-Online.com
  • 5G

Arm-based processors deliver 10x performance on AI and high-performance computing tasks

April 13, 2021 By Redding Traiger

NVIDIA announced its first data center CPU, an Arm-based processor that will deliver 10x the performance of today’s fastest servers on the most complex AI and high-performance computing workloads.

The result of more than 10,000 engineering years of work, the NVIDIA Grace CPU is designed to address the computing requirements for the world’s most advanced applications — including natural language processing, recommender systems, and AI supercomputing — that analyze enormous datasets requiring both ultra-fast compute performance and massive memory. It combines energy-efficient Arm CPU cores with an innovative low-power memory subsystem to deliver high performance with great efficiency.

Grace is a highly specialized processor targeting workloads such as training next-generation NLP models that have more than 1 trillion parameters. When tightly coupled with NVIDIA GPUs, a Grace CPU-based system will deliver 10x faster performance than today’s state-of-the-art NVIDIA DGX-based systems, which run on x86 CPUs.

While the vast majority of data centers are expected to be served by existing CPUs, Grace — named for Grace Hopper, the U.S. computer-programming pioneer — will serve a niche segment of computing.

The Swiss National Supercomputing Centre (CSCS) and the U.S. Department of Energy’s Los Alamos National Laboratory are the first to announce plans to build Grace-powered supercomputers in support of national scientific research efforts.

NVIDIA is introducing Grace as the volume of data and size of AI models are growing exponentially. Today’s largest AI models include billions of parameters and are doubling every two-and-a-half months. Training them requires a new CPU that can be tightly coupled with a GPU to eliminate system bottlenecks.

NVIDIA built Grace by leveraging the incredible flexibility of Arm’s data center architecture. By introducing a new server-class CPU, NVIDIA is advancing the goal of technology diversity in AI and HPC communities, where choice is key to delivering the innovation needed to solve the world’s most pressing problems.

CSCS and Los Alamos National Laboratory both plan to bring Grace-powered supercomputers, built by Hewlett Packard Enterprise, online in 2023.

Underlying Grace’s performance is fourth-generation NVIDIA NVLink interconnect technology, which provides a record 900 GB/s connection between Grace and NVIDIA GPUs to enable 30x higher aggregate bandwidth compared to today’s leading servers.

Grace will also utilize an innovative LPDDR5x memory subsystem that will deliver twice the bandwidth and 10x better energy efficiency compared with DDR4 memory. In addition, the new architecture provides unified cache coherence with a single memory address space, combining system and HBM GPU memory to simplify programmability.

Grace will be supported by the NVIDIA HPC software development kit and the full suite of CUDA and CUDA-X libraries, which accelerate more than 2,000 GPU applications, speeding discoveries for scientists and researchers working on the world’s most important challenges.

Availability is expected at the beginning of 2023.

You may also like:


  • Smart grids, software defined grids, and big data

  • What is exponential technology?

  • Measuring the performance of AI and its impact on society…

  • AI applications and emerging computing architectures – Virtual Conversation (part…

  • How “green” is your Artificial Intelligence?
DesignFast Banner version: 22e7f758

Filed Under: Applications, Artificial intelligence, Microcontroller Tips Tagged With: nvidia

Primary Sidebar

EE Training Center Classrooms

EE Classrooms

Featured Resources

  • EE World Online Learning Center
  • CUI Devices – CUI Insights Blog
  • EE Classroom: Power Delivery
  • EE Classroom: Building Automation
  • EE Classroom: Aerospace & Defense
  • EE Classroom: Grid Infrastructure
Search Millions of Parts from Thousands of Suppliers.

Search Now!
design fast globle

R&D World Podcasts

R&D 100 Episode 7
See More >

Current Digital Issue

April 2022 Special Edition: Internet of Things Handbook

How to turn off a smart meter the hard way Potential cyber attacks have a lot of people worried thanks to the recent conflict in Ukraine. So it might be appropriate to review what happened when cybersecurity fi rm FireEye’s Mandiant team demonstrated how to infiltrate the network of a North American utility. During this…

Digital Edition Back Issues

Sponsored Content

Positioning in 5G NR – A look at the technology and related test aspects

Radar, NFC, UV Sensors, and Weather Kits are Some of the New RAKwireless Products for IoT

5G Connectors: Enabling the global 5G vision

Control EMI with I-PEX ZenShield™ Connectors

Speed-up time-to-tapeout with the Aprisa digital place-and-route system and Solido Characterization Suite

Siemens Analogue IC Design Simulation Flow

More Sponsored Content >>

RSS Current EDABoard.com discussions

  • Help with Verilog replicate operator
  • ESP Serial Communication Problem with RS232
  • How to mark layer comments in CAP of spef file using StarRC
  • MAX5389 resetting by noise
  • Simulation of resonator in HFSS

RSS Current Electro-Tech-Online.com Discussions

  • Will Header and socket hold this PCB OK?
  • Relaxation oscillator with neon or...
  • software PWM
  • MPlab8 remove page breaks in list file
  • ATOM Diy module

Oscilloscopes Product Finder

Footer

EE World Online

EE WORLD ONLINE NETWORK

  • 5G Technology World
  • Analog IC Tips
  • Battery Power Tips
  • Connector Tips
  • DesignFast
  • EDABoard Forums
  • Electro-Tech-Online Forums
  • Engineer's Garage
  • Microcontroller Tips
  • Power Electronic Tips
  • Sensor Tips
  • Test and Measurement Tips
  • Wire & Cable Tips

EE WORLD ONLINE

  • Subscribe to our newsletter
  • Lee's teardown videos
  • Advertise with us
  • Contact us
  • About Us
Follow us on TwitterAdd us on FacebookConnect with us on LinkedIn Follow us on YouTube Add us on Instagram

Copyright © 2022 · WTWH Media LLC and its licensors. All rights reserved.
The material on this site may not be reproduced, distributed, transmitted, cached or otherwise used, except with the prior written permission of WTWH Media.

Privacy Policy