I have a server that is using a Quada P2000 graphics card and I keep having issues with a program crashing

Accelerated Computing CUDA CUDA Programming and Performance
1

I am a broadcast engineer at television station and I have server that is running a program called Encompass. We are trying to use this server and program to record the same off-air signals that viewers would watch at home. That way we have a record of what was broadcasted in the event that we get complaints from viewers about the audio being to loud or a show not having closed captioning.

The encompass program is playing back 5 streams via IP, 5 Streams via an antenna so I am sure the GPU is has a large load on it.

Unfortunately, the encompass program keeps crashing. In my troubleshooting I keep coming across some errors in Windows Event Viewer that are making me think the problem may be related to the GPU.

The first error that I am seeing is :

Faulting module name: ucrtbase.dll, version: 10.0.18362.1110

Now every time I see this error in the application log, I see this error at the same time in the system log:

"The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

Another error that I keep seeing in the application log is :

Faulting module name: nvcuvid.dll, version: 7.17.15.1640,

The research that I’ve done suggests to install the latest drivers, but I updated the drivers last week. I was going to try to do a clean install of the drivers when I noticed there a lot of NVIDA CUDA programs installed in addition to the drivers for the GPU. I’ve attached a screen shot of what I am seeing.

NVIDIA CUDA question
NVIDIA CUDA question918×218 23.1 KB

At this point I have three questions:

  1. Do I need to have all of the NVIDIA CUDA programs installed to make the GPU functional?
  2. Are the nvcuvid.dll and nvlddmkm issues related to the CUDA programs?
  3. Should I do a clean install of the CUDA programs as well as the drivers?

Here are the specs for the server:

GPU: Quadro P2000 Diver Version 516.40 Driver Type DHC

CPU: Intel XEON(R) ES-2630v4 @ 2.20GHz 2 Processors
Installed Ram: 64.0 GB.

Any help would be greatly appreciated

2

Not necessarily, but its difficult to say what you might need. You could start by just installing the driver itself. That picture depiction includes components from the CUDA toolkit, not only the GPU driver itself. The GPU driver itself is represented by the item “NVIDIA Graphics Driver 516.40”, and that certainly appears to be a recent driver, if not the latest. It might very well be the latest.

nvcuvid.dll is related to video decode. It’s not really connected with CUDA programs. nvlddmkm is the kernel mode GPU driver. (nvidia longhorn display driver model kernel mode) Everything that uses the GPU (CUDA, display, video, whatever) requires that.

I don’t normally find that necessary, but since I don’t know the underlying issue here, I cannot say.