ARM Holdings PLC · NASDAQ:ARM

ARM — Technical Presentation (March 2017)

21 slides from ARM's technical presentation, Mar 20, 2017

  1. Slide 1: ARM DynamIQ: The future of multi-core computing

    ARM DynamIQ: The future of multi-core computing March 20, 2017 © ARM 2017

  2. Slide 2: What is ARM Announcing?

    What is ARM Announcing? What? ARM DynamIQ technology: New multicore microarchitecture for all new ARM Cortex-A processors beginning this year Target markets: Automotive, networking, server and Where? primary compute devices Redefining flexible multicore and heterogenous compute Why? to enhance the e

  3. Slide 3: Extraordinary growth of compute – from chip to cloud

    Extraordinary growth of compute – from chip to cloud 5 years 4 years 22 years 1991 2013 2017 2021 50 100 50 billion billion chips shipped billion chips shipped chips projected to ship 3 © ARM 2017

  4. Slide 4: Building the ecosystems for the next 100 billion

    Building the ecosystems for the next 100 billion 100 billion chips Millions of lines of code Innovation on common architecture 1000+ community partners 450+ silicon partners 4 © ARM 2017

  5. Slide 5: ARM Total Computing

    ARM Total Computing Transforming solutions everywhere compute happens. 5 © ARM 2017

  6. Slide 6: ARM has changed the compute landscape

    ARM has changed the compute landscape • 3.5 billion people using ARM-based primary compute devices today • More than 100x increase in Work Play mobile CPU performance since 2009 • Cortex-A processors driving performance from chip to cloud Connect Learn 6 © ARM 2017

  7. Slide 7: 2020: Looking ahead from chip to cloud

    2020: Looking ahead from chip to cloud Ubiquitous A.I. Safe and autonomous Mixed Reality Hyper-efficient Secure private compute The future requires a new approach to CPU design… 7 © ARM 2017

  8. Slide 8: Technology for the next era of compute

    Designed from the Massive system More intelligent ground up performance uplift systems Technology for the next era of compute 8 © ARM 2017

  9. Slide 9: ARM DynamIQ: Multicore redefined

    ARM DynamIQ: Multicore redefined Greater flexibility with Redesigned memory Advanced compute New single cluster design or without big.LITTLE sub-system capabilities 9 © ARM 2017

  10. Slide 10: Accelerating AI adoption everywhere

    Accelerating AI adoption everywhere DynamIQ boosting AI/ML performance both on CPU and in system Dedicated processor instructions Improved access to for AI acceleration More than 50x AI performance Up to 10x quicker boost on the CPU in the next 3-5 years response to accelerators 10 © ARM 2017

  11. Slide 11: Safer autonomous systems

    Safer autonomous systems DynamIQ supports safety critical industrial and automotive systems (ASIL D) Resilient systems Allowing systems to operate safely under failure Adaptive compute Uncompromised performance for ADAS Faster responsiveness Quicker safety critical decision-making 11 © ARM 2017

  12. Slide 12: Scalable compute, tailored solutions

    Scalable compute, tailored solutions Nearly Enables previously “Unlimited” unused design points design spectrum Up to 8 CPUs Multicore Covers all existing use cases Up to 4 CPUs Up to 4 CPUs Single cluster Dual cluster DynamIQ cluster Homogenous processing Heterogeneous processing Dynamic flexibilit

  13. Slide 13: Efficient compute, Fast responsiveness

    Efficient compute, Fast responsiveness DynamIQ big.LITTLE  A new single cluster design for big.LITTLE  Increased efficiency from shared memory between CPUs  Higher performance through faster task migration 13 © ARM 2017

  14. Slide 14: Efficient compute, Fast responsiveness

    Efficient compute, Fast responsiveness DynamIQ big.LITTLE  A new single cluster design for big.LITTLE  Increased efficiency from shared memory between CPUs  Higher performance through faster task migration 14 © ARM 2017

  15. Slide 15: Efficient compute, Fast responsiveness

    Efficient compute, Fast responsiveness DynamIQ big.LITTLE  A new single cluster design for big.LITTLE  Increased efficiency from shared memory between CPUs  Higher performance through faster task migration 15 © ARM 2017

  16. Slide 16: Efficient compute, Fast responsiveness

    Efficient compute, Fast responsiveness DynamIQ big.LITTLE  A new single cluster design for big.LITTLE  Increased efficiency from shared memory between CPUs  Higher performance through faster task migration 16 © ARM 2017

  17. Slide 17: Efficient compute, fast responsiveness

    Efficient compute, fast responsiveness DynamIQ big.LITTLE  A new single cluster design for big.LITTLE  Increased efficiency from shared memory between CPUs  Higher performance through faster task migration 17 © ARM 2017

  18. Slide 18: Intelligent power savings

    Intelligent power savings C P U C P U CPU Power States CPU memory Finer-grained speed control Faster on/sleep/off Autonomous power management 18 © ARM 2017

  19. Slide 19: ARM DynamIQ technology

    ARM DynamIQ technology Faster responsiveness • Most versatile compute technology Intelligent power management • For all markets, all devices Accelerating AI • Higher performance and efficiency levels Increased scalability • Smarter capabilities for AI/ML Improved big.LITTLE Advanced safety 19 © ARM

  20. Slide 20: 100 billion chips

    Redefining the Changing how Built with Architecture for future of computing devices operate security in mind total computing helping to power the next 100 billion chips 20 © ARM 2017

  21. Slide 21: Trademark notice

    The trademarks featured in this presentation are registered and/or unregistered trademarks of ARM Limited (or its subsidiaries) in the EU and/or elsewhere. All rights reserved. All other marks featured may be trademarks of their respective owners. Copyright © 2017 ARM Limited © ARM 2017

ARM — Technical Presentation (March 2017) on 6ix