Co-existence of Tesla with two different Geforce cards

Virtual GPU Forums More vGPU Forums Tesla Boards
1

Hi all,

I have a Tesla C2050 in a HP Z620 workstation and I’ve been using it fine in TCC mode for 4+ years with a Geforce 210 as the graphics card. The Tesla is only used for computations and is not used for any graphics-related tasks.

I’ve recently upgraded my Geforce graphics card to Geforce GT730 for two reasons: 1) to free up a PCI x16 slot; 2) upgrade the computer’s graphics performance.

The two Nvidia products co-exist fine as long as the Tesla GPU is NOT in TCC mode. When I put it in TCC mode using nvidia-smi command line utility, the Geforce stops working completely. It ceases to show up on the "nvidia-smi -L" query. However, it still shows up as in working order on the Device Manager. This didn’t happen before and I’m wondering if it’s a driver conflict issue or these two cards are just not compatible.

The drivers are:
Geforce driver version 399.24 (Sep 2018)
Tesla driver version 391.29 (Mar 2018)

Any help is greatly appreciated. TCC mode is not a must-have but it potentially affects my simulation runtimes and I’d like to have it in TCC mode than not.

Thanks.

2

Not sure if this helps but I’ve noticed that the computer is actually starting up with the Tesla as the graphics card. I can tell because I can only see the "press ESC to enter bios settings" message when the monitor is connected to the Tesla. The Geforce only kicks in at the windows log-in screen.

I couldn’t find a way to tell the BIOS to use a certain slot for graphics output.

Any help is appreciated.