NVIDIA GPU Architectures Reference
Loading GPU architectures...
No matching architectures found.
Additional Resources
Notes
Compute Capability refers to the version of the CUDA architecture supported by the GPU. Higher compute capabilities generally support more advanced features and better performance.
SM Version (Streaming Multiprocessor) indicates the internal architecture generation of the GPU cores.
GPU Code Names are the internal chip designations used by NVIDIA during development and manufacturing.
GPU Series shows representative consumer (GeForce/RTX), professional (Quadro/RTX), and datacenter (Tesla/A-series/H-series) GPUs for each architecture.
Data Source: The architecture data is loaded from /data/nvidia-architectures.json. Edit that file to update the information displayed on this page.