Selur's Little Message Board

Full Version: Incorrectly recognized GPU
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hello to all. I installed the hybrid frontend in my Arch Linux. When the program initializes i receive the message

Detected the following cards which seem to lack NVIDIA PureVideo support: GeForce GTX 1060 Mobile
No card found that's supported by nvenc, the CUDA encoding library. -> Removing CUDA based encoding support. Card name was compared against: https://developer.nvidia.com/cuda-gpus

But the nvidia-settings recognize my GPU as a GeForce GTX 1060 as shown in the attachment. Any suggestions ?

Btw the system supports Cuda as i used it in deep learning TensorFlow.
It's not about whether the nvidia-settings recognizes the card or not.
It's about whether Hybrids list of nvenc capable cards contains the name of your card. Smile
Sadly NVIDIA doesn't to offer a list of all nvenc capable cards.
Quote:Any suggestions ?
Let me add your 'GeForce GTX 1060 Mobile' to the list of supported cards and wait for the next release of Hybrid.
(may be I get around to compile a Linux dev version over the weekend, if I do, I'll send you a link via PM.)

Cu Selur

Ps.: Only thing you could (in theory) do yourself is to get your system to report your card as a 'GeForce GTX 1060' instead of a 'GeForce GTX 1060 Mobile' when lspci is called. Smile
I confirm that using the dev version fixed the problem after clearing the tool cache. Problem solved.