To begin with, I'm a system's admin and not a well versed programmer. I've got enough understanding of the basics but not enough of what I need to program an answer to my question.

I'm interested in GPU performance per running process/application. I know there's various tools that can help with this, but I'm working on making my own. I've been able to get most of the "asset" information from WMI calls (make, model of the GPU(s), specs, etc.), but what I cannot track down is how to transform performance data per process to the GPU. When you look at Windows Performance Counters, I can backtrack PID to running process, but I cannot figure out how to backtrack LUID to GPU.

There were examples in other forums that worked pre-DXGI 1.4, but a forum posting here pointed out that those commands are no longer valid and it requires different methods. Posting: https://www.gamedev.net/forums/topic/677964-d3d12-acquiring-idxgiadapter3-interface-from-d3d12device/

I was barely able to hack together the commands from the pre-1.4 change over thanks to the example, but I have no real idea where to start to get the new commands to work.

Are the LUIDs from the DXGI methods the same LUIDs shown in PerfMon's GPU Performance Counters?

Is this something trivial enough someone might be able to spoon feed me a bit to reverse the LUID to the GPU and I can take it from there?

Thanks everyone,

-Gwain