Highlighted
ceejay1989
1 Copper

PE M610x with nVIDIA Tesla M2075

Hello,

We recently got a PE M610x second hand as well as an NVIDIA Tesla M2075 card.  I have placed the card in and installed Server 2012 R2.  I have installed the drivers as recommended from NVIDIA website (340.84) and the Display Adapter shows correctly in Device Manager.  

I have two issues:

1. The nVIDIA Control Panel will not open.  It says the settings are not available because I am not attached to an NVIDIA GPU.  However, even if I disable to embedded Display adapter in Device Management this behavior persists.  I have also tried to disable the embedded video card in the BIOS but it doesn't allow me to make that change.  

2. I have installed Hyper-V and RDS Virtualization Host role but Hyper-V manager does not recognize the NVIDIA card to allow me to use register is as a RemoteFX GPU.  I have checked and the M2075 is supposed to be compatible with RemoteFX.  

 

Any one with experience installing a video card in these blades?

 

Cheers

0 Kudos
2 Replies

Re: PE M610x with nVIDIA Tesla M2075

Someone in the community may be able to speak to the specific card, but looking at my documentation, I see that the M1060, the M2050-204, and the M2070-205 were supported. You might use the iDRAC and CMC to see if the hardware sees the card, and also check Windows to see if the device manager sees it. I'd start by making sure that it shows up the way it needs to, before focusing on the software side. 

Regarding software, I found a Microsoft article that may be helpful: https://social.technet.microsoft.com/wiki/contents/articles/16652.remotefx-vgpu-setup-and-configurat...

#Iwork4Dell
0 Kudos
ceejay1989
1 Copper

Re: PE M610x with nVIDIA Tesla M2075

I was able to get the card to register in Hyper-V RemoteFX.  Figured I would let everyone know if case someone else stumbles across this.

 

Nvidia Tesla cards default to a driver mode called TCC (Tesla Compute Cluster). Nvidia states “The TCC driver is a Windows driver that supports CUDA C/C++ applications. The driver enables remote desktop services and reduces the CUDA kernel launch overhead on Windows. Note that the TCC driver disables graphics on the Tesla products. The main purpose of TCC and the Tesla products is to aid applications that use CUDA to perform simulations, and large-scale calculations (especially floating-point calculations), such as image generation for professional use and scientific fields of study.”

Since I am trying to run RemoteFX on the card I was required to change the driver mode to WDDM (Windows Display Driver Mode) in order to get the card to register as a 3D adapter in Hyper-V. To change the mode you must use the Nvidia SMI utility in cmd.

C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe -dm 1|0

Where 0 = WDDM and 1 = TCC. A reboot is required after a successful prompt.

After the reboot and checking Hyper-V, the GPU shows up as a RemoteFX accepted GPU and the nVidia Control Panel launches successfully.

Here is an article on the Nvidia website about TCC:

https://docs.nvidia.com/gameworks/content/developertools/desktop/nsight/tesla_compute_cluster.htm

0 Kudos