Download as PDF, PPTX
Uploaded byNVIDIA
GPU Technology Conference 2014 Keynote
The document discusses advancements in GPU computing, highlighting trends in machine learning, big data analytics, and GPU architecture evolution from 2003 to 2016. It mentions the impact of GPU technology on high-performance computing, energy efficiency, and large-scale data processing. Key developments include NVLink, 3D memory integration, and significant uses in AI research, particularly referencing projects like Google Brain.
More Related Content
![[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...](https://cdn.slidesharecdn.com/ss_thumbnails/cs264201106-cudaninjasharetmp-110301171948-phpapp01-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to GPU Technology Conference 2014 Keynote
GPU Technology Conference 2014 Keynote
- 2. 5 4 3 2 1 0 2003 2005 2007 2009 2011 2013 TeraFLOPS GPU CPU
- 3. GTC — GROWINGAND EXPANDING 2010 2012 2014 397 429 729 FASTEST GROWING TOPICS Big Data Analytics Machine Learning Computer Vision FASTEST GROWING TOPICS Energy Exploration Life Science & Genomics Molecular Dynamics #1 TOPIC HPC / Supercomputing
- 4. 2012 2013 2014 FOSTERING THE GPU ECOSYSTEM Big Data / Cloud / Computer Vision AudioStreamTV
- 5.
- 6. Takayuki Aoki GlobalScientific Information and Computing Center Tokyo Institute of Technology “ Large-scale CFD Applications and a Full GPU Implementation of a Weather Prediction Code on the TSUBAME Supercomputer ”
- 7. BANDWIDTH BOTTLENECKS CPU GPU PCIe PCI Express CPU Memory GPU Memory 16GB/sec 60GB/sec 288GB/sec
- 8. INTRODUCING NVLINK CPU GPU PCIe Differential with embedded clock PCIe programming model (w/ DMA+) Unified Memory Cache coherency in Gen 2.0 5 to 12X PCIe
- 9. 5X More Bandwidthfor Multi-GPU Scaling GPU PCIe SWITCH CPU GPU GPU GPU
- 10. 3D MEMORY 3DChip-on-Wafer integration Many X bandwidth 2.5X capacity 4X energy efficiency 0 200 400 600 800 1000 1200 2008 2010 2012 2014 2016 Memory Bandwidth
- 11. Blaise Pascal 1623-1662 Mechanical Calculator Probability Theory Pascal’s Theorem Pascal’s Law
- 12. PASCAL NVLink 3DMemory Module 5 to 12X PCIe 3.0 2 to 4X memory BW & size 1/3 size of PCIe card
- 13. SGEMM / WNormalized 2012 2014 2008 2010 2016 Tesla CUDA Fermi FP64 Kepler Dynamic Parallelism Maxwell DX12 Pascal Unified Memory 3D Memory NVLink 20 16 12 8 6 2 0 GPU ROADMAP 4 10 14 18
- 14. MACHINE LEARNING Branchof Artificial Intelligence Computers that learn from data person car helmet motorcycle bird frog person dog chair person hammer flower pot power drill
- 15. Machine Learning usingDeep Neural Networks Input Result
- 16. Building High-level FeaturesUsing Large Scale Unsupervised Learning Q. Le, M. Ranzato, R. Monga, M. Devin, K. Chen, G. Corrado, J. Dean, A. Ng Stanford / Google 1 billion connections 10 million 200x200 pixel images 1,000 machines (16,000 cores) 3 days
- 17. 1,000 CPU Servers2,000 CPUs • 16,000 cores 600 kWatts $5,000,000 GOOGLE BRAIN Today’s Largest Networks 1B connections 10M images ~3 days ~30 ExaFLOPS Human Brain ~100B neurons x 1000 connections 500M images 5,000,000X “Google Brain” ~150 YottaFLOPS ~40,000 “Google Brain-Years” SOURCE: Ian Goodfellow
- 18. Deep Learning withCOTS HPC Systems A. Coates, B. Huval, T. Wang, D. Wu, A. Ng, B. Catanzaro Stanford / NVIDIA • ICML 2013 STANFORD AI LAB 3 GPU-Accelerated Servers 12 GPUs • 18,432 cores 4 kWatts $33,000 Now You Can Build Google’s $1M Artificial Brain on the Cheap “ “ -Wired 1,000 CPU Servers 2,000 CPUs • 16,000 cores 600 kWatts $5,000,000 GOOGLE BRAIN
- 19. DEMO: MACHINE LEARNING,SIMPLE TRAINING SET
- 20. 1.2M 1000 2 7 25 Image training set Classes Weeks of training GPUs EXAFLOPS total to train DEMO: MACHINE LEARNING, NYU OVERFEAT
- 21. CUDA for MACHINELEARNING Talks @ GTC Image Detection Face Recognition Gesture Recognition Video Search & Analytics Speech Recognition & Translation Recommendation Engines Indexing & Search Use Cases Early Adopters Image Analytics for Creative Cloud Image Classification Speech/Image Recognition Recommendation Hadoop Search Rankings
- 22. Big Data &Infinite Compute Turbocharge Deep Learning SOURCE: KPCB/Mary Meeker, company data. Unstructured data: IDC's Digital Universe Study. 800M photos uploaded per day 100 hours of video uploaded per minute Unstructured data exploding 0 100 200 300 400 500 600 700 800 900 2007 2008 2009 2010 2011 2012 2013 2014 Facebook Instagram Snapchat Flickr 0 20 40 60 80 100 120 2007 2008 2009 2010 2011 2012 2013 Hours (YouTube) Millions 1,104 5,379 0 1,000 2,000 3,000 4,000 5,000 6,000 2010 2015 Exabytes of data
- 24. DEMO: TITAN ZREVEAL
- 25. 5,760 CUDA cores 12GB memory 8 TeraFLOPS $2999
- 26. STANFORD AI LAB 1 Titan Z-Accelerated Server 3 Titan Zs • 17,280 cores 2 kWatts $12,000 1,000 CPU Servers 2,000 CPUs • 16,000 cores 600 kWatts $5,000,000 GOOGLE BRAIN 300X energy efficiency 400X lower cost Fits next to a desk
- 27. RenderMan with programmableshading 1.5 hours to render each frame CCI 6/32 minicomputer First CGI Film Nominated for an Academy Award®
- 28. State-of-the-art water simulator48 hours to simulate the base water 250 hours to render each frame 2013 Academy Award® Winner BEST VISUAL EFFECTS
- 29.
- 30.
- 31.
- 32.
- 33. One is aphoto, One is Iray…
- 34. Bunkspeed Maya Catia 3ds Max IRAY VCA SCALABLE GPU RENDERING APPLIANCE 8 Kepler-class 12GB per GPU 23,040 2 x 1GigE 2 x 10GigE 1 x InfiniBand GPUs GPU memory CUDA cores Network
- 35. DEMO: IRAY /HONDA
- 36. 0 20 40 60 80 Relative Performance CPU-only Workstation Quadro K5000 Workstation Iray VCA Bunkspeed Maya Catia 3ds Max IRAY VCA SCALABLE GPU RENDERING APPLIANCE MSRP $50,000
- 38. GRID GPU inthe Cloud
- 39. Ben Fathi ChiefTechnology Officer Horizon DaaS Platform
- 40.
- 41. “10 of theTop 10” Greenest Supercomputers Powered by CUDA GPUs
- 42. Unify GPU andTegra Architecture 192 fully programmable CUDA cores 326 GFLOPS 4X energy efficiency over A15 TEGRA K1 Mobile Super Chip MOBILE ARCHITECTURE Maxwell Kepler Tesla Fermi Tegra 3 Tegra 4 Tegra K1 GPU ARCHITECTURE
- 43. Computer Vision onCUDA Feature Detection / Tracking ~30 GFLOPS @ 30 Hz Object Recognition / Tracking ~180 GFLOPS @ 30 Hz 3D Scene Interpretation ~280 GFLOPS @ 30 Hz
- 44. JETSON TK1 1stMOBILE SUPERCOMPUTER FOR EMBEDDED SYSTEMS 192 CUDA cores 326 GFLOPS VisionWorks SDK $192
- 45. VISIONWORKS COMPUTER VISIONON CUDA Driver Assistance Computational Photography Augmented Reality Robotics CUDA Jetson TK1 VisionWorks Primitives Your Code Sample Pipelines Object Detection / Tracking Structure from Motion … Classifier Corner Detection …
- 46. Single Precision GFLOPS/ W Normalized 80 60 0 40 2013 2014 2011 2012 2015 Tegra 2 Tegra 3 Tegra 4 Tegra K1 Kepler GPU CUDA 64b & 32b CPU Erista Maxwell GPU 20 TEGRA ROADMAP
- 47. Andreas Reich Headof Audi Pre-Development
- 48.
- 51. CUDA EVERYWHERE PASCAL PC CLOUD MOBILE
- 52. DEMO: PORTAL ONSHIELD