Surround Computing: The Future of Work & Play

THE SURROUND COMPUTING ERA

Mark Papermaster

Senior Vice President and CTO, AMD

Dell World 2012

Austin, Texas

December 13, 2012

3 billion Internet users want any content, anytime, any platform, anywhere

By 2020, more than 30 billion connected devices and sensors1

Explosion of data, much of it unstructured data

Index and search is fundamentally changing how organizations manage data

Demand for data center servers is expected to exceed 10 million units by 20162

1. Gartner Group 2013 Trends, 2012 I.T. Symposium

2. Worldwide and Regional Server 2012-2016 Forecast, IDC, May 2012

2 | The Surround Computing Era | Dell World – December 2012

A RAPIDLY CHANGING ENVIRONMENT

REVOLUTIONARY TRANSFORMATION

10 years ago The interactive computing revolution

Graphics acceleration enabled visual, interactive

computing

This made computing accessible to everyone

Amazing experiences from touch-screen phones to

cinematic 3D


Starting now The Surround Computing Era

Computers are everywhere!

Integrating into our environment

Computing is part of everyday life, not a

distinct activity

Cloud will be center of digital lives, for apps,

content and preferences

We are entering the

Surround Computing Era

It’s multi-platform –

eyeglasses to room-size

It’s fluid – realistic output,

natural human input

It’s intelligent – anticipates

our needs


SURROUND COMPUTING

Smarter clients enable realistic, natural human

communication

Smarter servers engage accelerators, lower power

usage and distribute workloads to

appropriate instruction sets

Smarter clouds orchestrate tens of billions of

devices in real-time


SURROUND COMPUTING:

PROFOUND IMPLICATIONS FOR COMPUTER ARCHITECTURE

The primary job of a client device is to interact with people

Surround computing requires increasingly natural interaction

Natural UI solutions are emerging (Siri, Google Voice, Kinect) but we have a long way to go

Truly “natural” UI is extremely

computationally intensive

Practical solutions limited by computational power

Real time processing

Acceleration is required


SMARTER CLIENTS Realism Realistic rendering, flawless video and audio

Gestures Truly “natural”, no training

Voice Speaker-independent natural language, anywhere in a room

Faces UIs take cues from subtle expressions and read lips

Eye Tracking Control computer actions by simply looking

Biometrics Identify users and behave appropriately


SMARTER SERVERS

The last 5 years have exploded the “one size fits all” model

Raw data is growing exponentially and we are at the beginning of the

big data revolution

Workloads have changed, and continue to change, at

unprecedented rates

The fastest growing are small and highly parallelizable workloads

Servers will be optimized for the workload

Consistent Experiences Across Multiple Devices

SMARTER CLOUDS


SMARTER CLOUDS

The Cloud Is

the “Backbone” of

Surround Computing

Consistent experiences across multiple devices

All your data and content – everywhere

Location, environment, activity, preferences

Identity, security, privacy

Threat detection, containment

Situation-driven behavior

Anticipatory behavior

Surround offload

Trust Context AI Compute

Surround Computing Cloud Services

Connected devices drive cloud computational loads Datacenter architecture must be optimized for Surround Computing

Scale – to support tens of billions of connected devices

Acceleration – app back-end – NUI, graphics, analytics

Security, privacy – consistent end-to-end architecture

Real time – latency is critical

Dense servers – optimized for low power and data center space

ISA choice – workloads dynamically distributed to optimal instruction

set (x86, 64-bit ARM …)


SMARTER CLOUDS – MANAGING UNSTRUCTURED DATA


THE WAY FORWARD

Surround computing requires smarter clients, servers and clouds,

efficient datacenters

– Heterogeneous engines that accelerate key client and server workloads

– Datacenters optimized for latency, scalability, efficiency and efficacy

– End-to-end trust and security

– Optimized for new and emerging workloads

Making all this work together requires a new silicon architecture –

Heterogeneous System Architecture (HSA)

Biometric Recognition

Secure, fast, accurate: face, voice, fingerprints

Augmented Reality

Superimpose graphics, audio, and other digital

information as a virtual overlay

AV Content Management

Searching, indexing and tagging of video and

audio. Multimedia data mining

Beyond HD Experiences

Streaming media, new codecs, 3D,

transcode, audio

Natural UI and Gestures

Touch, gesture, and voice

Content Everywhere

Content from any source to any display

seamlessly

This is about making a smarter client. Each of these experiences

improves the fluency and realism of human-computer interaction


APPLICATION AREAS WITH ABUNDANT PARALLEL WORKLOADS

HSA is designed to make the hardware directly

accessible to the software, using the high level

languages they already use

C, C++, Java, Python…even JavaScript, HTML5

ISA agnostic – e.g., x86, 64-bit ARM

GPU becomes equal to the CPU in terms of

processing capability

Full programming language features

Shared virtual memory: pointer is a pointer

Coherency

Context switching

HSA Foundation – an industry-wide initiative


CHANGING THE THINKING, CHANGING THE GAME

Power

Efficient

Easy to

Program

Future

Looking

Established

Technology

Foundation

Open

Standard

Industry

Support


BENEFITS OF HETEROGENEOUS SYSTEM ARCHITECTURE

While pushing the heterogeneous computing

envelope, forge ahead with CPU performance

AMD cores cater to a wide variety of workloads

Server: “big data” analytics, virtualization,

cloud, web hosting

Client: Independent multi-threading, GPU support

threads, legacy code base

The next generation “Steamroller” core

delivers greater efficiency in all areas


X86 CORE PERFORMANCE

AMD’s next-generation, multi-threaded

“Steamroller” Microarchitecture

Expand computation efficiency across design

Target “real-world” client and server applications

Maintain high-frequency engine

Shared L2 Cache

“STEAMROLLER”

Fetch

FP Scheduler

L1 DCache L1 DCache

Pip

elin

e

Pip

elin

e

Pip

elin

e

Pip

elin

e

Integer

Scheduler

Integer

Scheduler

Decode Decode

Pip

elin

e

Pip

elin

e

Pip

elin

e

Pip

elin

e

128-b

it F

MA

C

128-b

it F

MA

C

MM

X U

nit


AMD’S “STEAMROLLER” CORE

Feed the cores faster

Improve single-core execution

Push on performance/watt


AMD WILL DEVELOP 64-BIT ARM PROCESSORS FOR SERVERS

Production of ARM technology-based AMD Opteron™

processors planned for servers in 2014

Unique offering with supercomputing style fabric

AMD will continue to design x86 for servers

The right solutions for leading server workloads

APU Accelerated

Software Applications

Serial and Task Parallel Workloads

HSA Accelerated Processing Unit

Data Parallel Workloads


EFFECTIVE COMPUTE OFFLOAD IS MADE EASY BY HSA

Unleash the best compute elements for parallel programs

10 years of server CPU leadership shipping in millions of systems

Enterprise class IP portfolio – memory, I/O, RAS, design methodologies, tools, platforms

Fabric technology for data center-level scalability

World-class 64-Bit microprocessor technology leadership

AMD is the only company

bringing world-class server

and systems expertise to both

the x86 and ARM ecosystems

Power efficiency and

Open Source ecosystem ARM

Performance and

established workloads x86


PROVIDING CUSTOMER CHOICE – UNIQUE POSITION IN THE MARKET

Power efficiency is Fundamental

Long battery life, sleek light weight form

factors, cool and quiet computation

Lower energy consumption and utility bills in

data center; lower TCO

Multi-faceted attack beyond process

technology is required:

Efficient design methodologies

Intelligent on die power management

Optimize hardware with software applications


SMART DESIGNING FOR LOW POWER

Great processors need comparably great

interconnect fabrics


FAST FABRICS TIE EVERYTHING TOGETHER

Optimally process unstructured data

Able to connect massive numbers of processors

Lowest possible overhead

AMD off–chip interconnect fabric IP

designed to enable significantly lower TCO

Links hundreds thousands of SoC modules

Shares hundreds of TBs storage and virtualizes I/O

160Gbps Ethernet Uplink

Instruction Set: ARM & x86

SoC

Module

Off-chip Interconnect Fabric

I/O VT TIO

Supercompute Fabric

ASIC

Fabric

CPU/APU

DRAM DRAM

DRAM DRAM

DRAM DRAM

DRAM DRAM

Ethernet


AMD FREEDOM FABRIC™ TECHNOLOGY


ENABLES DENSE SERVER OPTIMIZATION

End to End System View

Acceleration of the application stack

Agile delivery of tailored solutions

leveraging differentiated IP

Workload distributed to appropriate

ISA based on task

Surround computing requires

Supercomputing approach at

commercial cost and power


THE PURE SPEEDS AND FEEDS RACE IS OVER -

IT’S ABOUT THE SOLUTION

THE NEXT WAVE – SURROUND COMPUTING REVOLUTION

AMD’s products will enable the transition HSA – Consistent programming model for heterogeneous processors

Ambidextrous – flexibility of configuration and ISA

Fast fabrics – massive scalability, lowest latency

Relentless focus on power efficiency

Over the past 10 years, AMD technology inspired the

interactive computing revolution

Over the next 10 years, AMD will lead the way to surround

computing


THANK YOU

DISCLAIMER The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions and

typographical errors.

The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to

product and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences

between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. AMD assumes no obligation to update or

otherwise correct or revise this information. However, AMD reserves the right to revise this information and to make changes from time to

time to the content hereof without obligation of AMD to notify any person of such revisions or changes.

AMD MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE CONTENTS HEREOF AND ASSUMES NO

RESPONSIBILITY FOR ANY INACCURACIES, ERRORS OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION.

AMD SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE. IN

NO EVENT WILL AMD BE LIABLE TO ANY PERSON FOR ANY DIRECT, INDIRECT, SPECIAL OR OTHER CONSEQUENTIAL DAMAGES

ARISING FROM THE USE OF ANY INFORMATION CONTAINED HEREIN, EVEN IF AMD IS EXPRESSLY ADVISED OF THE POSSIBILITY OF

SUCH DAMAGES.

ATTRIBUTION

© 2012 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, Radeon, and combinations thereof

are trademarks of Advanced Micro Devices, Inc. Other names and logos are used for informational purposes only and may

be trademarks of their respective owners.

Surround Computing: The Future of Work & Play

Technology

Transcript of Surround Computing: The Future of Work & Play