2 GPU's Machine Learning.

[MikeFish12345]

I have this idea in theory and was wondering if it will actually work. 

 

I want to buy 2 GPU's. 1 is an rtx 3080/3090 that will be used for machine learning tasks (training neural networks etc...). The other will be for video output and more of an integrated graphics replacement. I had the 1030 in mind for this card. 

 

Will this setup work in practice as I want to have the more powerful GPU dedicated to Machine learning tasks and not be bogged down by me doing other low-end tasks simultaneously like web-browsing, YouTube, etc..?

 

Thanks in advance. 

[GuiltySpark_]

If you connect the display to the low end card, I don't see an issue.

[SorryBella]

9 minutes ago, MikeFish12345 said:

I want to buy 2 GPU's. 1 is an rtx 3080/3090 that will be used for machine learning tasks (training neural networks etc...). The other will be for video output and more of an integrated graphics replacement. I had the 1030 in mind for this card. 

Personally if you can run igpu equipped CPU, thatll be better for your power consumption and pcie management. But if you cant spare that, any low end video output should be okay.

[igormp]

22 minutes ago, MikeFish12345 said:

I have this idea in theory and was wondering if it will actually work. 

 

I want to buy 2 GPU's. 1 is an rtx 3080/3090 that will be used for machine learning tasks (training neural networks etc...). The other will be for video output and more of an integrated graphics replacement. I had the 1030 in mind for this card. 

 

Will this setup work in practice as I want to have the more powerful GPU dedicated to Machine learning tasks and not be bogged down by me doing other low-end tasks simultaneously like web-browsing, YouTube, etc..?

 

Thanks in advance. 

Yes, that's perfectly fine. You can even throw your 3080/3090 onto a secondary PCIe slot (even if it's an x4, you won't notice any performance downsides).