NVIDIA’s latest compact generative AI supercomputer is also its cheapest

Date:

Jeremy Gan

NVIDIA has just revealed the Jetson Orin Nano Super Developer Kit, which is the successor to its Jetson Orin Nano kit from 2022. This new compact generative AI supercomputer can fit into the palm of your hand. Included in the developer kit is an 8GB Jetson Orin Nano system-on-module and a reference carrier board.

In terms of performance, the Jetson Orin Nano Super can reach 68 trillion operations per second (TOPS), a 70 percent increase from its predecessor. NVIDIA also claims a 1.7 times improvement in generative AI inference performance and a 50 percent bandwidth increase to 102GB per second. As for CPU frequency, it’s now 1.7GHz, a modest bump compared to its predecessor’s 1.5GHz.

NVIDIA also claims that it “delivers gains” for all popular generative AI models. These include large language models (LLMs), vision language models and vision transformers.

The Jetson Orin Nano Super’s software updates can also boost its predecessor, which is great news if you have the original Jetson Orin Nano. All you need to do is to download the software and install the update. That said, NVIDIA didn’t give any specifics on how the older device’s performance would change.

The Jetson Orin Nano Super Developer Kit is currently available from NVIDIA-approved resellers. It’s currently just $249, down from its predecessor’s $499, making it the most affordable product of its kind right now.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Share post:

Subscribe

Popular

More like this
Related

Players rebuke clumsy ad strategies, even in popular games | Mobile Premier League

December 20, 2024 5:54 PM Candy Crush Saga has $20...

Arm lawsuit against Qualcomm ends in mistrial and favorable ruling for Qualcomm

December 20, 2024 4:17 PM Arm Total Design is aimed...

Players invested 8.34B hours into Blizzard titles in 2024, says studio

December 20, 2024 2:10 PM Image Credit: Blizzard Entertainment Blizzard this...

Hugging Face shows how test-time scaling helps small language models punch above their weight

December 20, 2024 12:46 PM Image credit: VentureBeat with Ideogram Join...