Nvidia Drive
{{short description|Computer platform by Nvidia}}
NVIDIA DRIVE is a computer platform by Nvidia, aimed at providing autonomous car and driver assistance functionality powered by deep learning.{{cite journal|last=Umar Zakir Abdul |first=Hamid |title=Current Collision Mitigation Technologies for Advanced Driver Assistance Systems–A Survey |journal=PERINTIS eJournal |date=2016 |volume=6 |issue=2 |url= https://www.researchgate.net/publication/311981545 |display-authors=etal|access-date=14 June 2017}}[https://www.nvidia.com/en-us/self-driving-cars/ NVIDIA DRIVE] The platform was introduced at the Consumer Electronics Show (CES) in Las Vegas in January 2015.{{cite news|url= http://www.cnet.com/roadshow/news/cars-drive-autonomously-with-nvidia-x1-based-computer/ |title=Cars drive autonomously with Nvidia X1-based computer|access-date=29 March 2016|agency=CNET|publisher=CNET|date=5 January 2015}} An enhanced version, the Drive PX 2 was introduced at CES a year later, in January 2016.{{cite news|url= https://blogs.wsj.com/digits/2016/01/04/nvidia-announces-another-car-supercomputer-at-ces/ |title=Nvidia Announces Another Car 'Supercomputer' at CES|access-date=29 March 2016|publisher=The Wall Street Journal|date=4 January 2016}}
The closely platform related software release program at some point in time was branded NVIDIA DRIVE Hyperion along with a revision number helping to match with the generation of hardware it is created for - and also creating ready to order bundles under those term. In former times there were only the terms Nvidia Drive SDK for the developer package and sub-included Nvidia Drive OS for the system software (aka OS) that came with the evaluation platforms or could be downloaded for OS switching and updating later on.
Hardware and semiconductors
= Maxwell based =
The first of Nvidia's autonomous chips was announced at CES 2015, based on the Maxwell GPU microarchitecture.{{Cite web|url= http://www.anandtech.com/show/8811/nvidia-tegra-x1-preview/4 |title=NVIDIA Tegra X1 Preview & Architecture Analysis|last=Smith|first=Joshua Ho, Ryan|access-date=2016-09-18}} The line-up consisted of two platforms:
== Drive CX ==
The Drive CX was based on a single Tegra X1 SoC (System on a Chip) and was marketed as a digital cockpit computer, providing a rich dashboard, navigation and multimedia experience.
Early Nvidia press releases reported that the Drive CX board will be capable of carrying either a Tegra K1 or a Tegra X1.[http://www.nvidia.de/object/nvidia-drive-tegra-X1-jan4-2015-de.html NVIDIA ebnet den Weg für die Autos von Morgen mit den NVIDIA-DRIVE-Automotive-Computern]
== Drive PX ==
File:NVIDIA Drive PX, Computex Taipei 20150601.jpg
The first version of Drive PX is based on two Tegra X1 SoCs, and was an initial development platform targeted at (semi-)autonomous driving cars.
= Pascal based =
Drive PX platforms based on the Pascal GPU microarchitecture were first announced at CES 2016.{{Cite web|url= http://www.anandtech.com/show/9903/nvidia-announces-drive-px-2-pascal-power-for-selfdriving-cars |title=NVIDIA Announces DRIVE PX 2 – Pascal Power For Self-Driving Cars|last=Smith|first=Ryan |access-date=2016-09-18}} This time only a new version of Drive PX was announced, but in multiple configurations.
== Drive PX 2 ==
The Nvidia Drive PX 2 is based on one or two Tegra X2 SoCs where each SoC contains 2 Denver cores, 4 ARM A57 cores and a GPU from the Pascal generation.{{Cite web|url= http://www.nvidia.com/object/drive-px.html |title=Autonomous Car Development Platform from NVIDIA DRIVE PX2|website=www.nvidia.com |access-date=2016-09-18}} There are two real world board configurations:
- for AutoCruise: 1× Tegra X2 + 1 Pascal GPU
- for AutoChauffeur: 2× Tegra X2 + 2 Pascal GPU's
There is further the proposal from Nvidia for fully autonomous driving by means of combining multiple items of the AutoChauffeur board variant and connecting these boards using e.g. UART, CAN, LIN, FlexRay, USB, 1 Gbit Ethernet or 10 Gbit Ethernet. For any derived custom PCB design the option of linking the Tegra X2 Processors via some PCIe bus bridge is further available, according to board block diagrams that can be found on the web.
All Tesla Motors vehicles manufactured from mid-October 2016 include a Drive PX 2, which will be used for neural net processing to enable Enhanced Autopilot and full self-driving functionality.{{cite news|last1=Lambert|first1=Fred |title=All new Teslas are equipped with NVIDIA's new Drive PX 2 AI platform for self-driving|url= https://electrek.co/2016/10/21/all-new-teslas-are-equipped-with-nvidias-new-drive-px-2-ai-platform-for-self-driving/ |access-date=25 January 2017|publisher=Electrek |date=21 October 2016}} Other applications are Roborace.{{cite web|url= https://electrek.co/2017/05/20/roborace-debuts-their-driverless-robocar-on-track-at-the-paris-eprix/ |title=Roborace debuts their driverless "Robocar" on track at the Paris ePrix |last1=Dow|first1=Jameson |date=20 May 2017|publisher=Electrek |access-date=21 May 2017}} Disassembling the Nvidia-based control unit from a recent Tesla car showed that a Tesla was using a modified single-chip Drive PX 2 AutoCruise, with a GP106 GPU added as a MXM Module. The chip markings gave strong hints for the Tegra X2 Parker as the CPU SoC.[https://electrek.co/2017/05/22/tesla-nvidia-supercomputer-self-driving-autopilot/ Look inside Tesla’s onboard Nvidia supercomputer for self-driving][http://www.telegiz.com/articles/23930/20170523/tesla-nvidia-nvidia-drive-px2-supercomputer-drive-px2-for-autocruise-drive-px2-for-autochauffeur-autonomous-cars-artificial-intelligence.htm Why Tesla’s Nvidia Supercomputer for Self-Driving is not as Powerful as Expected]
= Volta and Turing based =
Systems based on the Volta GPU microarchitecture and Turing GPU microarchitecture were first announced at CES 2017.{{cite news|last1=Cutress|first1=Ian|last2=Tallis|first2=Billy|title=CES 2017: Nvidia Keynote Liveblog|url= http://www.anandtech.com/show/10999/ces-2017-nvidia-keynote-live-blog |access-date=9 January 2017|publisher=Anandtech.com |date=4 January 2016}} It was originally named Drive PX, but later changed to DRIVE AGX.
== DRIVE AGX Xavier ==
The first Volta based Drive PX system was announced at CES 2017 as the Xavier AI Car Supercomputer. It was re-presented at CES 2018 as Drive PX Xavier.{{cite news|last1=Baldwin|first1=Roberto|title=NVIDIA unveils its powerful Xavier SOC for self-driving cars|url=https://www.engadget.com/2018/01/07/nvidia-xavier-soc-self-driving-cars/|access-date=8 January 2018|publisher=Engadget|date=8 January 2018}}{{cite web |url=https://web.archive.org/web/20190601192550/https://www.nvidia.com/en-us/self-driving-cars/drive-platform/hardware/ |title=Self-Driving Car Hardware - NVIDIA DRIVE |date= |author= |accessdate= June 2, 2019}} Initial reports of the Xavier SoC suggested a single chip with similar processing power to the Drive PX 2 Autochauffeur system.{{cite news|last1=Smith|first1=Ryan |title=Nvidia Teases Xavier, A High Performance SoC for Drive PX & AI|url= http://www.anandtech.com/show/10714/nvidia-teases-xavier-a-highperformance-arm-soc |access-date=22 June 2017|publisher=Anandtech |date=28 September 2016}} However, in 2017 the performance of the Xavier-based system was later revised upward, to 50% greater than Drive PX 2 Autochauffeur system. Drive PX Xavier is supposed to deliver 30 INT8 TOPS of performance while consuming only 30 watts of power.{{Cite web|url=https://www.nvidia.com/en-us/self-driving-cars/drive-px/|title=Autonomous Car Development Platform from NVIDIA DRIVE PX2|website=NVIDIA|language=en-us|access-date=2017-12-01}} This spreads across two distinct units, the iGPU with 20 INT8 TOPS as published early and the somewhat later on announced, newly introduced DLA that provided an additional 10 INT8 TOPS.
== DRIVE AGX Pegasus ==
In October 2017 Nvidia and partner development companies announced the Drive PX Pegasus system, based upon two Xavier CPU/iGPU devices and two Turing generation dGPUs. The companies stated the third generation Drive PX system would be capable of Level 5 autonomous driving, with a total of 320 INT8 TOPS of AI computational power and a 500 Watts TDP.{{Cite news|url=https://www.anandtech.com/show/11913/nvidia-announces-drive-px-pegasus-at-gtc-europe-2017-feat-nextgen-gpus |website=AnandTech |date=October 10, 2017 |title=NVIDIA Announces Drive PX Pegasus at GTC Europe 2017: Level 5 Self-Driving Hardware, Feat. Post-Volta GPUs|last=Oh|first=Nate|access-date=2017-10-10}}{{cite news|last1=Auchard|first1=Eric|title=NVIDIA unveils next-generation platform for fully autonomous cars|url=https://ca.reuters.com/article/technologyNews/idCAKBN1CF13F-OCATC|access-date=17 October 2017|publisher=Reuters|date=10 October 2017 |url-status=dead |archive-url=https://web.archive.org/web/20171018013530/https://ca.reuters.com/article/technologyNews/idCAKBN1CF13F-OCATC |archive-date= Oct 18, 2017 }}
= Ampere based =
== DRIVE AGX Orin ==
The Drive AGX Orin board family was announced on December 18, 2019, at GTC China 2019.{{Cite web|url=https://www.anandtech.com/show/15245/nvidia-details-drive-agx-orin-a-herculean-arm-automotive-soc-for-2022|title=NVIDIA Details DRIVE AGX Orin: A Herculean Arm Automotive SoC For 2022|last=Smith|first=Ryan|website=AnandTech |date=December 18, 2019 |access-date=2019-12-21}} On May 14, 2020, Nvidia announced that Orin would be utilizing the new Ampere GPU microarchitecture and would begin sampling for manufacturers in 2021 and be available for production in 2022.{{cite news|url=https://www.slashgear.com/nvidia-drive-upshifts-to-ampere-gpus-for-a-smoother-self-driving-roadmap-14620569/|title=NVIDIA Drive upshifts to Ampere GPUs for a smoother self-driving roadmap|author=Eric Abent|date=May 14, 2020|publisher=Slash Gear}} Follow up variants are expected to be further equipped with chip models and/or modules from the Tegra Orin SoC.
= Ada Lovelace based =
== DRIVE Atlan (Cancelled) ==
Nvidia announced the SoC codenamed Atlan on April 12, 2021 at GTC 2021.{{cite web |last1=Tomazin |first1=Tom |title=A Data Center on Wheels: NVIDIA Unveils DRIVE Atlan Autonomous Vehicle Platform |date=12 April 2021 |url=https://blogs.nvidia.com/blog/nvidia-drive-atlan-autonomous-vehicle-platform/|publisher=NVIDIA |access-date=18 March 2024}}
Nvidia announced the cancellation of Atlan on September 20, 2022, which was supposed to be equipped with a Grace-Next CPU, and an Ada Lovelace based GPU, and Nvidia announced that their next SoC was called Thor.{{cite web |last1=Smith |first1=Ryan |title=NVIDIA Drops DRIVE Atlan SoC, Introduces 2 PFLOPS DRIVE Thor for 2025 Autos |date=20 September 2022 |url=https://www.anandtech.com/show/17582/nvidia-drops-drive-atlan-soc-introduces-2-pflops-drive-thor-for-2025-autos|publisher=Anandtech |access-date=18 March 2024}}
= Blackwell based =
== DRIVE AGX Thor ==
Announced on September 20, 2022,{{cite web |last1=Kani |first1=Ali |title=DRIVE Thor Unites AV and Cockpit on a Single SoC |date=20 September 2022 |url=https://blogs.nvidia.com/blog/2022/09/20/drive-thor/|publisher=NVIDIA |access-date=6 January 2023}} Nvidia DRIVE AGX Thor comes equipped with an Arm Neoverse V3AE CPU,{{cite web |last1=Kani |first1=Ali |title=DRIVE Thor Unites AV and Cockpit on a Single SoC |date=20 September 2022 |url=https://blogs.nvidia.com/blog/2022/09/20/drive-thor/|publisher=NVIDIA |access-date=6 January 2023}} and a Blackwell based GPU, which was announced on March 18, 2024.{{cite web |last1=Labrie |first1=Marie |title=NVIDIA DRIVE Powers Next Generation of Transportation — From Cars and Trucks to Robotaxis and Autonomous Delivery Vehicles |date=18 March 2024 |url=https://nvidianews.nvidia.com/news/nvidia-drive-powers-next-generation-transportation|publisher=NVIDIA |access-date=18 March 2024}} It features 8-bit floating point support (FP8) and delivers 1000 Sparse INT8 TOPS, 1000 Sparse FP8 TFLOPS or 500 Sparse FP16 TFLOPS of performance.{{Cite web |title=NVIDIA DRIVE In-Vehicle Computing for Autonomous Vehicles |url=https://www.nvidia.com/en-us/self-driving-cars/in-vehicle-computing/ |access-date=2024-06-05 |website=NVIDIA |language=en-us}} Two Thor SoCs can be connected via NVLink-C2C.
BYD, Hyper, XPENG, Li Auto and ZEEKR have said to be use DRIVE AGX Thor in their vehicles.{{Cite web |last=Shapiro |first=Danny |date=2024-03-18 |title=All Eyes on AI: Automotive Tech on Full Display at GTC 2024 |url=https://blogs.nvidia.com/blog/automotive-tech-gtc-2024/ |access-date=2024-06-05 |website=NVIDIA Blog |language=en-US}} The Lynk & Co 900 is the first production vehicle to feature the DRIVE AGX Thor SoC.{{Cite web |last=Bobylev |first=Denis |date=2025-02-19 |title=Lynk & Co 900 is the first car with Nvidia Thor chip for 1,000 TOPS |url=https://carnewschina.com/2025/02/19/lynk-co-900-is-the-first-car-with-nvidia-thor-chip-4-times-more-powerful-than-orin/ |access-date=2025-03-13 |website=CarNewsChina.com |language=en-US}}
Software and bundling
With the label Hyperion{{cite web | url=https://developer.nvidia.com/drive-faq | title=Drive-faq | date=3 November 2021 }} added to their reference platformpresentation from December 2019: https://on-demand.gputechconf.com/gtc-cn/2019/pdf/CN9618/presentation.pdf series Nvidia promotes their mass products so that others can easily test drive and then create their own automotive grade products on top. Especially the feature rich software part of the base system is meant to be a big help for these others to quickly go ahead into developing their application specific solutions. Third-party companies, such as DeepRoute.ai, have publicly indicated using these software platform as their base of choice.{{cite web | url=https://pandaily.com/deeproute-ai-to-integrate-nvidia-drive-hyperion-into-l4-autonomous-driving-solutions/ | title=DeepRoute.ai to Integrate NVIDIA DRIVE Hyperion into L4 Autonomous Driving Solutions | date=22 March 2022 }} The whole design is concentrating on UNIX/Posix compatible or derived runtime environments (Linux,{{cite web | url=https://developer.nvidia.com/drive-archive | title=Nvidia Drive Archives | date=27 July 2017 }} Android,{{cite web | url=https://forums.developer.nvidia.com/t/android-installation-on-drive-agx/188412 | title=Android installation on DRIVE AGX | date=3 September 2021 }} QNX - aka the DRIVE OS variants) with special support for the semiconductors mentioned before in form of internal (CUDA, Vulkan) and external support (special interfaces and drivers for camera, lidar, CAN and many more) of the respective reference boards. For clearness Nvidia bundles the core of the developer needed software as Drive SDK that is sub-divided into DRIVE OS, DriveWorks, DRIVE AV, and DRIVE IX components.{{cite web | url=https://developer.nvidia.com/drive/drive-sdk | title=Nvidia Drive SDK | date=20 February 2019 }}
Note: As of now the above table is still 'fresh' and thus might be incomplete.
Developer kit comparison
Note: dGPU and memory are stand-alone semiconductors; all other components, especially ARM CPU cores, iGPU, DLA and PVA are integrated components of the listed main computing device(s). Tesla 2.0 and 2.5 are products, not developer kits.
References
{{Reflist}}
{{Linux devices}}
{{Single-board computer}}
{{Nvidia}}
Category:Neural processing units