TiECon Florida keynote - New opportunities for entrepreneurs using GPU & CUDA

SHANKER TRIVEDI Vice President, Worldwide Sales (PSG) | September 2013

NVIDIA THE VISUAL COMPUTING COMPANY

NVIDIA Growth Drivers

~$800M / 15% CAGR ~$1B / 9% CAGR ~$200M / 68% CAGR ~$2B / 32% CAGR

CUDA & GRID 20M Artists & Designers

100M PLM users

SHIELD & GRID 400M Gamers

TESLA & GRID 5-10M Servers

GPU, TEGRA & IP Billions of Devices

PRO-VISUALIZATION & DESIGN

GAMING HPC & BIG DATA ANALYTICS

OEM

Revenue based on current run rates. CAGR from FY10-FY13.

From A Modest Beginning

To Almost Unbelievable Reality

The March of GPUs

0  

50  

100  

150  

200  

250  

2007   2008   2009   2010   2011   2012  

GBytes/s  Peak  Memory  Bandwidth    

M1060  

Nehalem    3  GHz  

Westmere  3  GHz  

8-­‐core    Sandy  Bridge  

3  GHz  

Fermi  M2070  

Fermi+  M2090  

0  

200  

400  

600  

800  

1000  

1200  

1400  

2007   2008   2009   2010   2011   2012  

Gflops/s  Peak  DP  FP    Performance  

Nehalem  3  GHz  

Westmere  3  GHz  

Fermi  M2070  

Fermi+  M2090  

M1060  

8-­‐core  Sandy  Bridge  

3  GHz  

NVIDIA  GPU  (ECC  off)   x86  CPU  Double  Precision:  NVIDIA  GPU   Double  Precision:  x86  CPU  

Kepler   Kepler  

GPUs Enable Next Generation Graphics (with Computing) Deferred Lighting, Depth of Field and Post-Processing | Facial Performance | Wave and Fire Simulation | Interactive Global Illumination | Volumetric Lighting | iRAY on CUDA

CPU GPU + CPU

MotionDSP – Example Image Stabilization & Rectification

Aerodynamics/CFD

Electromagnetics

Satellite Imaging

Computer Vision

Video Enhancement

Signal Processing

GPU Apps Run 10x-100x Faster

Traditional CPU Cluster GPU Workstation

Nodes Required 12 1 (4 GPUs)

Interconnect QDR IB None

Time to complete 4.98 days 2.25 days

Power Consumption 5.7 kW (681.3 kWh) 1.0 kW (54.0 kWh)

System Cost (per day) $96,800 ($88.40) $5200 ($4.75)

Simulation Cost (681.3 * 0.18) + (88.40 * 4.98) (54.0 * 0.18) + (4.75 * 2.25)

$562.87 $20.41

AMBER 4 simultaneous simulations, 23,000 atoms, 250ns each

GPU Apps are 5-25X Cheaper To Deploy

Source: San Diego Supercomputing Center

Better Simulation - Toto Example

MEDIA & ENT. OIL AND GAS RESEARCH MANUFACTURING GOVERNMENT

The buyer plans for including accelerators in their next technical computing server purchase has more than doubled from 29% to over 65% in last 20 months.

IDC Market Research April, 2013

“

”

GPUs becoming mainstream in the Data Center For Scientific & Technical Computing

CUDA Developer Momentum

2008 2013

4,000 Academic Papers

150K CUDA Downloads

60 University Courses

100M CUDA –Capable GPUs

1 Supercomputer

430M CUDA-Capable GPUs

50 Supercomputers

1.6M CUDA Downloads

640 University Courses

37,000 Academic Papers

Easiest Way to Learn CUDA

50K Registered

127 Countries

$$

Learn from the Best

Anywhere, Any Time

It’s Free!

Engage with an Active Community

Big Data Analytics

Machine Learning

Computer Vision

Visual Shopping

Real-time Video Delivery

Searching Audio

Analyzing Twitter

Beyond Scientific & Technical Computing

The New Enterprise Mobile | BYOD | Massive Data

NVIDIA GRID

Virtualized GPU Servers

Visual Computing Appliances

Meet Ira!

The Next Generation of Visual Computing

http://www.youtube.com/watch?v=Vx0t-WJFXzo

Console Games Can Now Target Mobile

What’s the Next Big Thing?

Augmented Reality

Healthcare

Autonomous Vehicles, Drones, Robots

Deep Machine Learning

Visual Analytics of Big Data

GPU & CUDA Easy & Cheap

What’s the Next Big Thing?

Augmented Reality Autonomous Robots Computer Vision Deep Machine Learning Visual Analytics Biological Computing

Resources http://www.slideshare.net/shankert/nvidia-intro-0513 http://blogs.nvidia.com/ https://www.facebook.com/gpucomputing http://www.nvidia.com/object/race-for-better-science.html http://www.nvidia.com/object/gpuventures.html http://www.gputechconf.com/page/home.html http://www.gputechconf.com/page/gtc-learning-center.html http://www.slideshare.net/shankert/nvidia-gtc-computerecapbrochure http://sarahtariq.com/presentations.html http://www.slashgear.com/nvidia-shield-hands-on-with-parrot-ar-drone-2-0-26288075/ https://developer.nvidia.com/ http://www.nvidia.com/object/tesla_software.html http://gpu.meetup.com/ http://gpgpu.org/ https://www.udacity.com/course/cs344 https://www.coursera.org/course/hetero