Guidelines for GPU Selection

VisionPro Deep Learning4.1 supports all NVIDIA® CUDA®-enabled GPUs with compute capability of 6.1 and higher. For more information about GPU compute capability, see the NVIDIA website.

Cognex recommends selecting an NVIDIA® GPU with a higher core clock frequency, higher memory, and a higher number of CUDA cores and Tensor cores because these models provide faster computational speeds.

Note: To use GPU driver versions higher than 546.01, you need to disable the Sysmem Fallback setting in the NVIDIA Control Center. See Disable Shared GPU Memory.

Tested GPU Models

The following list includes the GPU models that Cognex has tested. Cognex aims to support most common GPU families, but we do not test every model A specific spatial arrangement of a set of features (Blue Locate and Blue Read tools only.) During a post-processing step, the Blue Locate and Blue Read tools can fit all of the features detected in an image to the models defined for the tool. The overall pose and identity of the model is then returned.. If the GPU model meets the compute capability requirements, you can use it. Cognex customers use most common GPUs other than very new models, even if the GPUs have not been tested by Cognex.

  • NVIDIA® GeForce® RTX 3080

  • NVIDIA® GeForce® RTX 3080 Ti

  • NVIDIA® GeForce® RTX 3090

  • NVIDIA® GeForce® RTX 4080

  • NVIDIA® GeForce® RTX 4090

  • NVIDIA® Quadro® P4000

Not Recommended GPU Models

Cognex does not recommend the following GPU models because they provide low performance and slow training Training is the process that your tool, which is a neural network, is learning about the features (pixels) based on the labels you made. For example, a tool will learn the defect/normal pixels in each image based on the defect/normal labels you drew. The goal of the tool Training is learning enough to give the correct inspection results of whether an unseen image is defective or not. The key to training is to ensure that you include all possible variations within your training set, and that your images are accurately labeled. Training times vary by the application, tool setup and the GPU in the PC being used to train the network. time for unknown reasons:

  • NVIDIA® GeForce® GTX 1650

  • NVIDIA® GeForce® GTX 1660