GPU support
The OCR SDK supports computation acceleration based on NVIDIA GPUs. To use the GPU smoothly, the following specifications and settings must be met.
Note: As of August 2025, the Windows version does not support GPUs.
CUDA and driver requirements
- CUDA version: 11.6.x or later
- NVIDIA driver version:
Linux (x86_64): 450.80.02 or later
Supported architectures
The OCR SDK supports the following NVIDIA GPU architectures.
- Maxwell
- Pascal
- Volta
- Turing
- Ampere
- Ada
- Hopper
- Blackwell
- Rubin
Note: You can check your GPU architecture based on the manufacturer or model name.
How to check compute capabilities
Use the following command to check whether the GPU in your system is suitable for using the OCR SDK.
nvidia-smi --query-gpu=compute_cap format=csv
Supported graphics cards
| NVIDIA Architecture Name | Maxwell | Pascal | Volta | Turing | Ampere | Ada | Hopper | Blackwell | Rubin |
|---|---|---|---|---|---|---|---|---|---|
| Compute Capabilities | sm_50 | sm_60 | sm_70 | sm_75 | sm_80 | sm_89 | sm_90 | sm_100 | sm_110? |
| sm_52 | sm_61 | sm_72 (Xavier) | sm_86 | sm_90a (Thor) | sm_100a | ||||
| sm_53 | sm_62 | sm_87 (Orin) |
| Available | |
| Available (check version) | |
| Contact us if usage is abnormal | |
| Not available |
Maxwell
- SM52 or SM_52, compute_52 – Quadro M6000 , GeForce 900, GTX-970, GTX-980, GTX Titan X
- SM53 or SM_53, compute_53 – Tegra (Jetson) TX1 / Tegra X1, Drive CX, Drive PX, Jetson Nano
Pascal
- SM60 or SM_60, compute_60 – Quadro GP100, Tesla P100, DGX-1 (Generic Pascal)
- SM61 or SM_61, compute_61 – GTX 1080, GTX 1070, GTX 1060, GTX 1050, GTX 1030 (GP108), GT 1010 (GP108) Titan Xp, Tesla P40, Tesla P4, NVIDIA Drive PX2 discrete GPU
- SM62 or SM_62, compute_62 – NVIDIA Drive PX2 integrated GPU, Tegra (Jetson) TX2
Volta
- SM70 or SM_70, compute_70 – DGX-1 with Volta, Tesla V100, GTX 1180 (GV104), Titan V, Quadro GV100
- SM72 or SM_72, compute_72 – Jetson AGX Xavier, Drive AGX Pegasus, Xavier NX
Turing
- SM75 or SM_75, compute_75 – GTX/RTX Turing – GTX 1660 Ti, RTX 2060, RTX 2070, RTX 2080, Titan RTX, Quadro RTX 4000, Quadro RTX 5000, Quadro RTX 6000, Quadro RTX 8000, Quadro T1000/T2000, Tesla T4
Ampere
- SM80 or SM_80, compute_80 – NVIDIA A100 ("Tesla" naming removed – GA100), NVIDIA DGX-A100
- SM86 or SM_86, compute_86 – (Since CUDA 11.1) Tesla GA10x GPUs, RTX Ampere – RTX 3080, GA102 – RTX 3090, RTX A2000, A3000, RTX A4000, A5000, A6000, NVIDIA A40, GA106 – RTX 3060, GA104 – RTX 3070, GA107 – RTX 3050, RTX A10, RTX A16, RTX A40, A2 Tensor Core GPU, A800 40GB
Ada Lovelace
- SM89 or SM_89, compute_89 – NVIDIA GeForce RTX 4090, RTX 4080, RTX 6000 Ada, Tesla L40, L40s Ada, L4 Ada
Notes
- For GPUs not listed, please contact us after testing.
- For CUDA 12.x or later, please contact us after testing.