NVIDIA GeForce RTX 4090
The GeForce RTX 4090 is NVIDIA's flagship consumer graphics card based on the Ada Lovelace architecture and the AD102 die, launched on 12 October 2022. It features 16384 CUDA cores, 24 GB of GDDR6X memory on a 384-bit bus, plus 3rd-generation RT cores and 4th-generation Tensor cores.
In robotics and AI it is used as a compute module for local inference and model training (including LLMs), image processing and simulation. NVIDIA rates its AI performance at up to 1321 TOPS, with a total graphics power (TGP) of 450 W and a recommended 850 W power supply.
Classification
GPU (discrete graphics processor) · serves as: AI compute.
Which group NVIDIA GeForce RTX 4090 belongs to and how it is built
Subcategory covering discrete graphics cards (GPUs) installed in a PCI Express slot. In robotics and AI they serve as accelerators for local inference, model training and image processing.
A massively parallel graphics processing unit (GPU) used for both graphics rendering and general-purpose compute (GPGPU), including neural network training and inference.
Technical specification
Basic physical properties of NVIDIA GeForce RTX 4090 — dimensions, weight and materials