{"id":352,"date":"2020-10-06T10:24:20","date_gmt":"2020-10-06T10:24:20","guid":{"rendered":"https:\/\/machine-learning.webcloning.com\/2020\/10\/06\/nvidia-ceo-outlines-vision-for-age-of-ai-in-news-packed-gtc-kitchen-keynote\/"},"modified":"2020-10-06T10:24:20","modified_gmt":"2020-10-06T10:24:20","slug":"nvidia-ceo-outlines-vision-for-age-of-ai-in-news-packed-gtc-kitchen-keynote","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2020\/10\/06\/nvidia-ceo-outlines-vision-for-age-of-ai-in-news-packed-gtc-kitchen-keynote\/","title":{"rendered":"NVIDIA CEO Outlines Vision for \u2018Age of AI\u2019 in News-Packed GTC Kitchen Keynote"},"content":{"rendered":"
\n

Outlining a sweeping vision for the \u201cage of AI,\u201d NVIDIA CEO Jensen Huang Monday kicked off this week\u2019s GPU Technology Conference<\/a>.<\/p>\n

Huang made major announcements in data centers, edge AI, collaboration tools and healthcare in a talk simultaneously released in nine episodes, each under 10 minutes<\/a>.<\/p>\n

\u201cAI requires a whole reinvention of computing \u2013 full-stack rethinking \u2013 from chips, to systems, algorithms, tools, the ecosystem,\u201d Huang said, standing in front of the stove of his Silicon Valley home.<\/p>\n

Behind a series of announcements touching on everything from healthcare to robotics to videoconferencing, Huang\u2019s underlying story was simple: AI is changing everything, which has put NVIDIA at the intersection of changes that touch every facet of modern life.<\/p>\n

More and more of those changes can be seen, first, in Huang\u2019s kitchen, with its playful bouquet of colorful spatulas, that has served as the increasingly familiar backdrop for announcements throughout the COVID-19 pandemic.<\/p>\n

\u201cNVIDIA is a full stack computing company \u2013 we love working on extremely hard computing problems that have great impact on the world \u2013 this is right in our wheelhouse,\u201d Huang said. \u201cWe are all-in, to advance and democratize this new form of computing \u2013 for the age of AI.\u201d<\/p>\n

This week\u2019s GTC is one of the biggest yet. It features more than 1,000 sessions\u2014400 more than the last GTC\u2014in 40 topic areas. And it\u2019s the first to run across the world\u2019s time zones, with sessions in English, Chinese, Korean, Japanese, and Hebrew.<\/p>\n

Accelerated Data Center\u00a0<\/b><\/h2>\n

Modern data centers, Huang explained, are software-defined, making them more flexible and adaptable.<\/p>\n

That creates an enormous load. Running a data center\u2019s infrastructure can consume 20-30 percent of its CPU cores. And as \u201ceast-west traffic, or traffic within a data center, and microservices increase, this load will increase dramatically.<\/p>\n

\u201cA new kind of processor is needed,\u201d Huang explained: \u201cWe call it the data processing unit.\u201d<\/p>\n

The DPU<\/a> consists of accelerators for networking, storage, security and programmable Arm CPUs to offload the hypervisor, Huang said.<\/p>\n

The new NVIDIA BlueField 2 DPU<\/a> is a programmable processor with powerful Arm cores and acceleration engines for at-line-speed processing for networking, storage and security. It\u2019s the latest fruit of NVIDIA\u2019s acquisition of high-speed interconnect provider Mellanox Technologies, which closed in April.<\/p>\n

\"\"<\/a><\/p>\n

Data Center \u2014 DOCA \u2014 A Programmable Data Center Infrastructure Processor<\/b><\/h2>\n

NVIDIA also announced DOCA, its programmable data-center-infrastructure-on-a-chip architecture.<\/p>\n

\u201cDOCA SDKs let developers write infrastructure apps for software-defined networking, software-defined storage, cybersecurity, telemetry and in-network computing applications yet to be invented,\u201d Huang said.<\/p>\n

Huang also touched on a partnership with VMware, announced last week<\/a>, to port VMware onto BlueField. VMware \u201cruns the world\u2019s enterprises \u2014 they are the OS platform in 70 percent of the world\u2019s companies,\u201d Huang explained.<\/p>\n

\"\"<\/a><\/p>\n

Data Center \u2014 DPU Roadmap in \u2018Full Throttle\u2019<\/b><\/h2>\n

Further out, Huang said NVIDIA\u2019s DPU roadmap shows advancements coming fast.<\/p>\n

BlueField-2 is sampling now, BlueField-3 is finishing and BlueField-4 is in high gear, Huang reported.<\/p>\n

\u201cWe are going to bring a ton of technology to networking,\u201d Huang said. \u201cIn just a couple of years, we\u2019ll span nearly 1,000 times in compute throughput\u201d on the DPU.<\/p>\n

BlueField-4, arriving in 2023, will add support for the CUDA parallel programming platform<\/a> and NVIDIA AI \u2014 \u201cturbocharging the in-network computing vision.\u201d<\/p>\n

You can get those capabilities now, Huang announced, with the new BlueField-2X. It adds an NVIDIA Ampere architecture GPU to BlueField-2 for in-networking computing with CUDA and NVIDIA AI.<\/p>\n

\u201cBluefield-2X is like having a Bluefield-4, today,\u201d Huang said.<\/p>\n

\"\"<\/a><\/p>\n

Data Center \u2014 GPU Inference Momentum<\/b><\/h2>\n

Consumer internet companies are also turning to NVIDIA technology to deliver AI services.<\/p>\n

Inference \u2014 which puts fully-trained AI models to work \u2014 is key to a new generation of AI-powered consumer services.<\/p>\n

In aggregate, NVIDIA GPU inference compute in the cloud already exceeds all cloud CPUs, Huang said.<\/p>\n

Huang announced that Microsoft is adopting NVIDIA AI on Azure to power smart experiences on Microsoft Office<\/a>, including smart grammar correction and text prediction.<\/p>\n

Microsoft Office joins Square, Twitter, eBay, GE Healthcare and Zoox, among other companies, in a broad array of industries using NVIDIA GPUs for inference.<\/p>\n

Data Center \u2014 Cloudera and VMware\u00a0<\/b><\/h2>\n

The ability to put vast quantities of data to work, fast, is key to modern AI and data science.<\/p>\n

NVIDIA RAPIDS is the fastest extract, transform, load, or ETL, engine on the planet, and supports multi-GPU and multi-node.<\/p>\n

NVIDIA modeled its API after hugely popular data science frameworks \u2014 Pandas, XGBoost and ScikitLearn \u2014 so RAPIDS is easy to pick up.<\/p>\n

On the industry-standard data processing benchmark, running the 30 complex database queries on a 10TB dataset, a 16-node NVIDIA DGX cluster ran 20x faster than the fastest CPU server.<\/p>\n

Yet it\u2019s one-seventh the cost and uses one-third the power.<\/p>\n

Huang announced that Cloudera, a hybrid-cloud data platform that lets you manage, secure, analyze and learn predictive models from data, will accelerate the Cloudera Data Platform with NVIDIA RAPIDS, NVIDIA AI and NVIDIA-accelerated Spark.<\/p>\n

NVIDIA and VMware also announced a second partnership, Huang said.<\/p>\n

The companies will create a data center platform that supports GPU acceleration for all three major computing domains today: virtualized, distributed scale-out and composable microservices.<\/p>\n

\u201cEnterprises running VMware will be able to enjoy NVIDIA GPU and AI computing in any computing mode,\u201d Huang said. \u201c<\/p>\n

(Cutting) Edge AI\u00a0<\/b><\/h2>\n

Someday, Huang said, trillions of AI devices and machines will populate the Earth \u2013 in homes, office buildings, warehouses, stores, farms, factories, hospitals, airports.<\/p>\n

The NVIDIA EGX AI platform<\/a> makes it easy for the world\u2019s enterprises to stand up a state-of-the-art edge-AI server quickly, Huang said. It can control factories of robots, perform automatic checkout at retail or help nurses monitor patients, Huang explained.<\/p>\n

Huang announced the EGX platform is expanding<\/a> to combine the NVIDIA Ampere architecture GPU and BlueField-2 DPU<\/a> on a single PCIe card. The updates give enterprises a common platform to build secure, accelerated data centers.<\/p>\n

Huang also announced an early access program<\/a> for a new service called NVIDIA Fleet Command. This new application makes it easy to deploy and manage updates across IoT devices, combining the security and real-time processing capabilities of edge computing with the remote management and ease of software-as-a-service.<\/p>\n

Among the first companies provided early access to Fleet Command is KION Group, a leader in global supply chain solutions, which is using the NVIDIA EGX AI platform to develop AI applications for its intelligent warehouse systems.<\/p>\n

Additionally, Northwestern Memorial Hospital, the No. 1 hospital in Illinois and one of the top 10 in the nation, is working with Whiteboard Coordinator<\/a> to use Fleet Command for its IoT sensor platform.<\/p>\n

\u201cThis is the iPhone moment for the world\u2019s industries \u2014 NVIDIA EGX will make it easy to create, deploy and operate industrial AI services,\u201d Huang said.<\/p>\n

\"\"<\/a><\/p>\n

Edge AI \u2014 Democratizing Robotics<\/b><\/h2>\n

Soon, Huang added, everything that moves will be autonomous. AI software is the big breakthrough that will make robots smarter and more adaptable. But it\u2019s the NVIDIA Jetson AI computer<\/a> that will democratize robotics.<\/p>\n

Jetson is an Arm-based SoC designed from the ground up for robotics. That\u2019s thanks to the sensor processors, the CUDA GPU and Tensor Cores, and, most importantly, the richness of AI software that runs on it, Huang explained.<\/p>\n

The latest addition to the Jetson family, the Jetson Nano 2GB, will be $59<\/a>, Huang announced. That\u2019s roughly half the cost of the $99 Jetson Nano Developer Kit announced last year.<\/p>\n

\u201cNVIDIA Jetson is mighty, yet tiny, energy-efficient and affordable,\u201d Huang said.<\/p>\n

\"\"<\/a><\/p>\n

Collaboration Tools<\/b><\/h2>\n

The shared, online world of the \u201cmetaverse\u201d imagined in Neal Stephensen\u2019s 1992 cyberpunk classic, \u201cSnow Crash,\u201d is already becoming real, in shared virtual worlds like Minecraft<\/i> and Fortnite<\/i>, Huang said.<\/p>\n

First introduced in March 2019, NVIDIA Omniverse<\/a> \u2014 a platform for simultaneous, real-time simulation and collaboration across a broad array of existing industry tools \u2014 is now in open beta<\/a>.<\/p>\n

\u201cOmniverse allows designers, artists, creators and even AIs using different tools, in different worlds, to connect in a common world\u2014to collaborate, to create a world together,\u201d Huang said.<\/p>\n

Another tool NVIDIA pioneered, NVIDIA Jarvis conversational AI, is also now in open beta, Huang announced. Using the new SpeedSquad benchmark, Huang showed it\u2019s twice as responsive and more natural sounding when running on NVIDIA GPUs.<\/p>\n

It also runs for a third of the cost, Huang said.<\/p>\n

\u201cWhat did I tell you?\u201d Huang said, referring to a catch phrase he\u2019s used in keynotes over the years. \u201cThe more you buy, the more you save.\u201d<\/p>\n

Collaboration Tools \u2014 Introducing NVIDIA Maxine<\/b><\/h2>\n

Video calls have moved from a curiosity to a necessity.<\/p>\n

For work, social, school, virtual events, doctor visits \u2014 video conferencing is now the most critical application for many people. More than 30 million web meetings take place every day.<\/p>\n

To improve this experience, Huang announced NVIDIA Maxine<\/a>, a cloud-native streaming video AI platform for applications like video calls.<\/p>\n

Using AI, Maxine can reduce the bandwidth consumed by video calls by a factor of 10. \u201cAI can do magic for video calls,\u201d Huang said.<\/p>\n

\u201cWith Jarvis and Maxine, we have the opportunity to revolutionize video conferencing of today and invent the virtual presence of tomorrow,\u201d Huang said.<\/p>\n

\"\"<\/a><\/p>\n

Healthcare\u00a0<\/b><\/h2>\n

When it comes to drug discovery amidst the global COVID-19 pandemic, lives are on the line.<\/p>\n

Yet for years the costs of new drug discovery for the $1.5 trillion pharmaceutical industry have risen. New drugs take over a decade to develop, cost over $2.5 billion in research and development \u2014 doubling every nine years \u2014 and 90 percent of efforts fail.<\/p>\n

New tools are needed. \u201cCOVID-19 hits home this urgency,\u201d Huang said.<\/p>\n

Using breakthroughs in computer science, we can begin to use simulation and in-silico methods to understand the biological machinery of the proteins that affect disease and search for new drug candidates, Huang explained.<\/p>\n

To accelerate this, Huang announced NVIDIA Clara Discovery \u2014 a state-of-the-art suite of tools for scientists to discover life-saving drugs.<\/p>\n

\u201cWhere there are popular industry tools, our computer scientists accelerate them,\u201d Huang said. \u201cWhere no tools exist, we develop them \u2014 like NVIDIA Parabricks, Clara Imaging, BioMegatron, BioBERT, NVIDIA RAPIDS.\u201d<\/p>\n

Huang also outlined an effort to build the U.K.\u2019s fastest supercomputer, Cambridge-1<\/a>, bringing state-of-the-art computing infrastructure to \u201can epicenter of healthcare research.\u201d<\/p>\n

\"\"<\/a><\/p>\n

Cambridge-1 will boast 400 petaflops of AI performance, making it among the world\u2019s top 30 fastest supercomputers. It will host NVIDIA\u2019s U.K. AI and healthcare collaborations with academia, industry and startups.<\/p>\n

NVIDIA\u2019s first partners are AstraZeneca, GSK, King\u2019s College London, the Guy\u2019s and St Thomas\u2019 NHS Foundation Trust<\/a> and startup Oxford Nanopore.<\/p>\n

NVIDIA also announced a partnership with GSK to build the world\u2019s first AI drug discovery lab<\/a>.<\/p>\n

Arm<\/h2>\n

Huang wrapped up his keynote with an update on NVIDIA\u2019s partnership with Arm<\/a>, whose power-efficient designs run the world\u2019s smart devices.<\/p>\n

NVIDIA agreed to acquire<\/a> the U.K. semiconductor designer last month for $40 billion.<\/p>\n

\u201cArm is the most popular CPU in the world,\u201d Huang said. \u201cTogether, we will offer NVIDIA accelerated and AI computing technologies to the Arm ecosystem.\u201d<\/p>\n

Last year, Huang said, NVIDIA announced it would port CUDA and our scientific computing stack to Arm. Today, Huang announced a major initiative to advance the Arm platform \u2014 we\u2019re making investments across three dimensions:<\/p>\n