•   
  •   
  •   

Technology NVIDIA's Drive Atlan chip is headed to cars in 2025

20:55  12 april  2021
20:55  12 april  2021 Source:   engadget.com

Nvidia's Grace AI chip leaves Intel processors behind

  Nvidia's Grace AI chip leaves Intel processors behind Also, Nvidia's Atlan chip should boost self-driving cars in 2025.Nvidia's current brainiest chip, the A100, is typically yoked to Intel Xeon processors. Nvidia chips do the grunt work, but Intel chips oversee it. With Grace, named after pioneering programmer Grace Hopper, the company opted to embed several Arm Neoverse processor cores within the chip to speed up processing, said Paresh Kharya, an Nvidia senior director. The chip news arrived at Nvidia's GTC 2021 conference this week.

While we still have to wait until next year to see NVIDA's Orin chip start making its way to consumer cars, the company is already working on its next-generation Drive processor. It's called Atlan, and NVIDIA currently plans to roll it out with model year 2025 cars. Since it's so far out, the company didn't share too many details on the chipset, but what it did say is that it will feature its next-generation GPU architecture and new ARM CPU cores. It will also include a BlueField data processing unit, which NVIDIA says will assist the chip in completing the complex AI workloads necessary to enable autonomous vehicles.

NVIDIA's Grace, its first datacenter CPU, is another major threat to Intel

  NVIDIA's Grace, its first datacenter CPU, is another major threat to Intel Just like Apple did with its M1 chip, NVIDIA is taking on Intel directly with its own ARM-based CPU. But don't expect it in PCs anytime soon. Named after the pioneering computer scientist Grace Hopper, Grace is NVIDIA's first datacenter CPU, targeted at massive workloads like AI supercomputing and natural language processing. It's powered by ARM Neoverse cores and it'll be tightly integrated with the company's latest GPU technology. According toAccording to NVIDIA, a Grace-powered system is ten times faster at training a natural language processing (NLP) model with one trillion parameters compared to its x86 DGXTM machines.

a car parked in a parking lot

Those disparate parts will come together in a chip that will deliver data center-like performance, with NVIDIA claiming Atlan will be capable of more than 1,000 trillion operations per second. To put that in perspective, Orin, the Drive chip that's making its way to consumer cars next year, can complete 254 trillion operations per second. According to NVIDIA, all that power will allow automakers to build "software-defined" vehicles that they'll be able to upgrade in perpetuity with the help of OTA updates. And to help make those updates easier, Atlan will use the same CUDA and TensorRT APIs and libraries as its existing Drive chips.

a car parked in a parking lot: An autonomous-driving Lincoln MKZ equipped with Nvidia technology gives demonstration rides during the 2017 CES in Las Vegas, Nevada, U.S., January 5, 2017. REUTERS/Steve Marcus © Steve Marcus / reuters An autonomous-driving Lincoln MKZ equipped with Nvidia technology gives demonstration rides during the 2017 CES in Las Vegas, Nevada, U.S., January 5, 2017. REUTERS/Steve Marcus

Nvidia warns the great GPU shortage will continue throughout 2021 .
Demand will outstrip supply for the rest of the yearThat means the difficulties of finding a new 30-series GPU will likely continue, possibly into 2022. Nvidia previously warned that supplies would “likely remain lean” through its fiscal Q1, and now that we’re approaching that milestone (end of April) it’s clear that supply hasn’t improved significantly to meet demand.

usr: 3
This is interesting!