Select integrated GPU for Front Panel rendering

[Bruniii]

Hi all,

I have an application running on a system with a discrete Nvidia GPU and an integrated one. The application call a dll to perform data processing on the Nvidia GPU while, in parallel, updating the UI plots in a parallel loop with the data from the previous iteration of data collection. Right now the integrated GPU is not used (at least according to the windows task manager); enabling the plotting results in a small (but visibile) increase in the elaboration time of the data processing.

Hence, I'm assuming that LabVIEW is choosing to use the discrate GPU for the UI/plotting also.

Is it possibile to force LabVIEW to use the integrated GPU for the UI rendering?

Thanks,

Marco.

[Rolf Kalbermatter]

On 3/22/2025 at 10:29 AM, Bruniii said:

Hi all,

I have an application running on a system with a discrete Nvidia GPU and an integrated one. The application call a dll to perform data processing on the Nvidia GPU while, in parallel, updating the UI plots in a parallel loop with the data from the previous iteration of data collection. Right now the integrated GPU is not used (at least according to the windows task manager); enabling the plotting results in a small (but visibile) increase in the elaboration time of the data processing.

Hence, I'm assuming that LabVIEW is choosing to use the discrate GPU for the UI/plotting also.

Is it possibile to force LabVIEW to use the integrated GPU for the UI rendering?

I haven't really looked into this, but AFAIK LabVIEW has no idea on which GPU its output is generated. LabVIEW interfaces to the Windows GDI (Graphic Device Interface), which supposedly abstracts such low level details completely away. This would be something that would have to be configured in Windows itself somehow, however modern Windows 10 and 11 usually try to hide such "user unfriendly" operations away as much as possible.

Edited Monday at 04:11 PM by Rolf Kalbermatter