The week of April 12, 2021, marks Nvidia’s showcase Graphics Technology Conference or GTC21 (Press announcement). With the disruptive themes of this virtual event, Nvidia may need to seriously consider a new name for the conference series. Nvidia is still the leader in the graphics market, but the direction of Nvidia’s innovation clearly goes beyond graphics.
Here are some of the event highlights:
Nvidia has created a platform branded Omniverse that allows designers, creators, scientists, engineers, and other developers to create and collaborate in a simulated world. It is such a powerful tool, it can be used to create digital twins of real-world designs to interact, develop and test without ever having to prototype. From creators designing off world animation to automobile manufacturers building assembly lines, Omniverse is a big head start. One of the newest widgets based on Omniverse Nvidia dubs GANverse3D even allows you to use a 2D image to quickly render a high-quality 3D object. If the object is common enough, such as a car, the AI is trained enough to even fill in the features and physics to rapidly accelerate storyboarding, game design, plant automation and much more.
Data center compute
Nvidia is no newcomer when it comes to data center GPU coprocessing, but it is helping to reshape data center computing with all new levels of compute architecture. Up until a few years ago, almost all data center compute was based on raw CPU performance. As shown in the graph below from Yole’s 2021 Q1 Processor Monitor, that strategy is changing rapidly. Nvidia and a host of other processor specialist are driving a coprocessing explosion. Nvidia is not only developing graphics acceleration, but also driving AI acceleration and network acceleration. It is rearchitecting the data center scalability with not only processors, but with systems partners supplying full featured accelerated DGX powerhouse servers and clusters of servers into Superpods. But one of the biggest highlights at GTC21 was the announcement that Nvidia is developing a platform to include its GPUs, Bluefield DPUs and now Nvidia CPUs code named Grace expected in 2023. These CPUs are being architected on Arm processor IP and specially designed to take advantage of Nvidia’s accelerated platform to drastically reduce (an estimated 30x performance) the bottleneck between CPU and memory and parallel processing. Nvidia also continues to expand the applications for accelerated computing in data centers including enterprise, with EGX platforms and expansion of AI for securing networks and providing state of the art language processing with AI acceleration.
Autonomous vehicles on Nvidia drive
Finally, Nvidia continues driving AI to the edge and nowhere is that more evident than Nvidia’s work on autonomous vehicle innovation. Nvidia data center processing is one of the leading platform solutions for training autonomous vehicle software. From Parker at 1 TOP, to Xavier at 30 TOPs to Orin at 256 TOPS, Nvidia has been providing cutting edge platforms for autonomous vehicles and mobility as a service. Many in the Automotive industry have said it is not realistic to build self-driving cars until we can process over a 1000 TOPS of data. At GTC21, Nvidia announced the development of Atlan, expected in 2024 to provide 1000 TOPS performance. Given the development cycle of the automotive industry, it will still likely be another 5 years to see this in mass market vehicles, but the developments of today on Nvidia’s Drive Platform can be applied to this next generation processor. That puts level 5 autonomy ‘on the roadmap’ for this decade.
There was definitely more to see at GTC21. But as this summary highlights, it is definitely going beyond its Graphics Technology Conference roots.
About the author
Tom Hackenberg is a Principal Analyst for Computing and Software in the Semiconductor, Memory and Computing Division at Yole Développement (Yole). Tom is engaged in developing processor market monitors and research into related technology trends. He is currently focused on low and ultralow power solutions such as MCUs.
Tom is an industry leading expert with more than a decade’s experience reporting on markets for semiconductor processors including CPUs, GPUs, MPUs, MCUs, SoC ASICs & ASSPs, FPGAs and configurable processors. Tom is also well-versed in related technology trends including IoT, heterogeneous processing, chiplets, AI and edge computing.
Prior to joining Yole, Tom was a principal analyst at OMDIA, IHS Markit and began processor market research in 2006 for IMS Research. He worked with market-leading processor suppliers developing both syndicated and custom research. Tom holds a BSECE from the University of Texas at Austin specializing in Processors and FPGAs.
This article has been written in collaboration with John Lorenz and Adrien Sanchez, both Technology & Market Analysts and members of the Computing team at Yole Développement.
Related reports and monitors
While finding new application, and sustaining the established ones, CPUs, GPUs, and APUs hold the keys to the next ten years of the processor market.
Popular data-intensive applications and the spread of AI applications are boosting server computing market to reach US$33B in 2025.
Softbank confirmed, in August, that it was either accelerating the IPO or selling its subsidiary, Arm Holdings, Inc., causing rumors of a potential buyer to flourish…
Apple’s successful implementation of in-house processor designs for select new MacBook and Mac Mini computers could open the door to more Arm-based PCs…
Moore’s law has been slowed down but it has not yet reached its limit, semiconductor processor manufactures have rolled out their latest products using the smallest technology node in the market….