GTC 2018: A New AI Era Dawns

lovehao (25)in #technology • 6 years ago

GTC 20I8
“GTC was the introduction to the future of AI, a protector, a healer, a helper, a guardian, a visionary, and just a little slice of amazing.” —IT Business Edge
“Press quote.” —Publication “Clearly the adoption of GPU computing is growing.” —ZDNet
Deep Learning Institute 10,000 training instances Keynote 550K views Show Floor 150 exhibitors Inception Awards $1M to 3 AI startups “If the excitement at GTC this week is any indication, the market will need lots and lots of GPUs in the years to come.” —Datanami GTC is ground zero of the GPU computing movement. NVIDIA pioneered this computing model for those doing groundbreaking work that cannot be done without a supercharged computer. GTC 2018 was a four-day event with 8,300 registered attendees who were offered more than 600 technical sessions. More than 200 reporters and 80 industry analysts experienced first-hand NVIDIA’s lineup of announcements.
“NVIDIA’s CEO, Jensen Huang, is a consummate showman and he didn’t disappoint this year.” —IT Business Edge
Real-time ray tracing has been the dream of computer scientists since it was first described nearly 40 years ago. NVIDIA stunned the audience with its new NVIDIA RTX technology — a platform for real-time ray tracing. “Ray Tracing and Giant GPUs Electrify GTC 2018” —Electronic Design
The Quadro GV100 with NVIDIA RTX technology is the greatest advance in computer graphics of the past 15 years, since our introduction of the programmable shader. NVIDIA RTX is the culmination of 10 years of research, combining a new GPU architecture, algorithms, and deep learning in a way that no one else can. NVIDIA has reinvented computer graphics, again. “NVIDIA Is Really Throwing Down the Hammer with the Quadro GV100” —Hot Hardware
For 30 years, the dynamics of Moore’s law held true. But CPU performance scaling has slowed. GPU computing is defining a new, supercharged law. It starts with a highly specialized parallel processor called the GPU and continues through system design, system software, algorithms, and all the way through optimized applications. The world is jumping on board. “NVIDIA Is So Far Ahead of the Curve” —The Inquirer 370PF 20182013 Total GPU FLOPS of Top 50 Systems 15X in 5 Yrs 8M 20182013 820,000 20182013 CUDA Downloads 5X in 5 Yrs GPU Developers 10X in 5 Yrs Measured performance of Amber, CHROMA, GTC, LAMMPS, MILC, NAMD, Quantum Espresso, SPECFEM3D 1 10 100 2013 2018 GPU–Accelerated Computing CPU Moore’s Law
TITAN VNVIDIA AI InferenceNVIDIA GPU Cloud We are advancing GPU computing for deep learning and AI at the speed of light. We create the entire stack, and make it easily available in every computer, datacenter, and cloud. We supercharged NVIDIA AI with a new “double-sized” 32GB Volta GPU; announced the NVIDIA DGX-2, the power of 300 servers in a box; expanded our inference platform with TensorRT 4 and Kubernetes on NVIDIA GPU; and we built out the NVIDIA GPU Cloud registry with 30 GPU-optimized containers and made it available from more cloud service providers. “Creating Powerful System-level Solutions Will Give It an Edge Against Rivals Who Have Merely Developed a Good Chip” —TheStreet 30 GPU-Optimized Containers NEW TensorRT 4, TensorFlow, Kaldi, ONNX, WinML Out of stock!NGC Now on AWS, GCP, AliCloud, Oracle NEW 32GB NEW with V100 32GB NEW DGX-2 Tesla V100 DGX Systems Every Cloud
AI researchers want gigantic GPUs. We launched a breakthrough in deep learning computing with the introduction of NVIDIA DGX-2, the first single server capable of delivering two petaflops of computational power. DGX-2 features NVSwitch, a revolutionary GPU interconnect fabric which enables its 16 Tesla V100 GPUs to simultaneously communicate at a record speed of 2.4 terabytes per second. Programming DGX-2 is like programming “the largest GPU in the world.” “NVIDIA Gave a Look Inside Its DGX-2, the Star of This Year’s GTC” —EE Times