NVIDIA® GPU Requirements

You do not need a GPU to run In-Sight ViDi, but Cognex recommends to use an NVIDIA® GPU for improved performance. In-Sight ViDi utilizes the GPU during runtime, and during the development of your application, typically with the training of tools. Using a GPU can provide the following benefits:

  • Improved speed and performance while training In-Sight ViDi tools

  • Improved performance of runtime workspaces

Note:
  • A GPU does not affect the performance of the In-Sight D900 series vision system during runtime. With the vision system, the runtime workspace executes solely on the vision system itself.

Minimum GPU Requirements

In-Sight ViDi has the following requirements for minimum GPU capability:

  • In-Sight ViDi 1.9.0 supports all NVIDIA CUDA® enabled GPUs with compute capability of 3.0 or higher. The Low Precision parameter requires CUDA compute capability 6.1 or higher. For more information, visit the NVIDIA website.

  • 4 GB GPU memory is required for runtime operation.

  • 8 GB GPU memory is required for training tools.

  • NVIDIA GeForce® driver version of 456.38 or higher.

Guidelines for GPU Selection

The following table gives you guidance for selecting GPUs, indicating the models Cognex explicitly tested and confirmed to work, as well as the models that do not work. The table does not include all supported GPU models.

Note: Cognex aims to support most common GPU families, but we do not test every model. Cognex customers use most other common GPUs other than very new models, even if the GPUs have not been tested by Cognex.
Note:

The NVIDIA GeForce Titan V, Quadro GV100 and Tesla V100 include NVIDIA Tensor Cores. In-Sight ViDi automatically utilizes Tensor Cores if they are present, provided you have a Standard or Advanced security license. Contact your Cognex Sales Representative for information about enabling the functionality on your Cognex Security Dongle.

GPU Model Tested by Cognex

NVIDIA GeForce®

GTX 1060 6 GB

Untested

GTX 1070
GTX 1080

GTX 1080 Ti

Tested

Note: Recommended for high performance.

RTX 2070

Tested

RTX 2080

RTX 2080 Ti
RTX 3060
RTX 3060 Ti
RTX 3070
RTX 3070 Ti
RTX 3080
RTX 3080 Ti
RTX 3090

Tested

Note: Recommended for high performance.
RTX 4070 Ti Untested
RTX 4080 Tested
RTX 4090

NVIDIA RTX/Quadro®

P2000

Untested

P4000

Tested
P5000 Untested

 

GV100
RTX A3000
RTX A4000 Tested
RTX A5000 Untested

 
RTX 3000

RTX 4000

RTX 5000

RTX 6000

RTX 8000

Titan V

Titan RTX

NVIDIA Tesla®

V100

Models known not to work:

  • GTX 1660 and GTX 1650: Low performance and slow training time for unknown reasons.

  • GeForce 700 series: Not supported.