Getting hardware accelerated graphics on a headless Windows Server 2008 R2

Question

I have a series of command line applications which use accelerated hardware contexts to perform logic on the GPU, in a CUDA like fashion.

The applications are Windows specific, so they can not run on other operating systems. I need to integrate these into a web service, which is running in a Windows Server 2008 R2 cluster.

Currently, when I run them via remote desktop or from the web service, they fail to acquire an accelerated hardware context. I understand the failure via remote desktop, but that is non-essential.

The real key is get the accelerated context when invoked by the web service (a node.js application). I thought that a dongle might do the trick, but I'm not sure.

The server has an Nvidia Quadro FX 3700 - a fairly beefy, but dated card (with the latest drivers installed).

Anyone have any advice how to achieve this?

Did you write these applications? If not, you should name them. — Michael Hampton, Jan 09 '14 at 23:57
They are proprietary neural nets written by a business partner. They do computer vision and 3D reconstruction operations. — Blake Senftner, Jan 10 '14 at 01:41

score 0 · Answer 1 · edited May 23 '17 at 12:41

0

Seems to be a common problem with things like CUDA. See this question https://stackoverflow.com/questions/2464182/windows-2008-renderfarm-service-createprocessasuser-session-0-isolation-and-o

edited May 23 '17 at 12:41

Community

1

answered Jan 10 '14 at 10:58

JamesRyan

8,138
2
24
36

Getting hardware accelerated graphics on a headless Windows Server 2008 R2

1 Answers1