Page 34 - EE Times Europe Magazine – June 2024
P. 34

34 EE|Times EUROPE



         AUTONOMOUS VEHICLES | ARTIFICIAL INTELLIGENCE
        What Is Nvidia Doing in Automotive?


        By Egil Juliussen

        Nvidia’s lucrative detour into AI, where it has established dominance, has yielded
        technologies that are steering its automotive business back into growth mode.



                he Nvidia GPU Technology Conference (GTC) has been the most       The first generations, Drive CX and Drive
                important GPU conference for over a decade. Now, it may also be the   PX, used the Maxwell microarchitecture and
                most important AI hardware-software-deployment conference, as   focused on digital cockpits and ADAS applica-
                                                                                tions. Even in 2015, there were 256 or
        T GPUs are driving much of AI training and inferencing activities.      512 CUDA cores available that could be used
          At past GTCs, autonomous vehicles played an important role, as they were   for parallel compute operations.
                                                                                  The second-generation Drive PX2 tar-
        expected to be an early deployment segment for AI. The 2024 GTC presented   geted ADAS functions and was introduced
        some automotive content, but less than previous conferences. This year’s   in January 2016. Drive PX2 used the Pascal
        automotive content was mostly rolled into the presentations on Nvidia’s AI   GPU architecture, and the number of Arm
                                                                                processors increased to 12 CPUs, all using
        strategy, and in the keynote presentation, automotive AI was part of the robot-  64-bit processors. Tesla used the Drive PX for
        ics strategy. The most important automotive-related announcement was that   its battery electric vehicle (BEV) Autopilot for
        the Blackwell GPU would be part of Nvidia’s Drive Thor centralized computer   several years.
                                                                                  Drive PX Xavier was introduced in
        for safe and secure AVs.                                                January 2017 and used the Volta microarchi-
          This article offers perspectives on what Nvidia is doing in the automotive   tecture. Nvidia positioned it for use in L3 and
        industry.                                                               L4 vehicles.
                                                                                  Drive PX Pegasus arrived in
                                                                                September 2017, was based on the Tur-
        THE EVOLUTION OF NVIDIA DRIVE                         ing architecture and was Nvidia’s first automotive product with AI
        Drive is Nvidia’s computer platform for developing advanced driver-   functionality. The Pegasus platform provided a performance increase
        assistance systems (ADAS) and AVs. Nvidia Drive was introduced at    of about 10× over Drive PX2. Availability was scheduled for mid-2018,
        CES 2015 and has grown through multiple generations. Table 1 sum-  and as of October 2017, Nvidia had more than 200 partners developing
        marizes Nvidia Drive’s evolution over the years.      hardware and/or software products for its Drive platform.






























                                                                                                                   IMAGE: ADOBE STOCK









        JUNE 2024 | www.eetimes.eu
   29   30   31   32   33   34   35   36   37   38   39