Sign up ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

Is there any way to find out which process is using the GPU for intensive calculations? (OpenCL, CUDA, etc? Primarily OpenCL?)

e.g. if you run an OpenCL-powered video transcoder, you'll see nothing in your task manager that tells your GPU is burning through data at a ridiculous rate - though your fans will have kicked into overdrive.

If the reply is driver-specific, I am interested in nVidia primarily.

share|improve this question

3 Answers 3

up vote 6 down vote accepted

Turns out Sysinternals' Process Explorer can do this (of course it can, there's nothing it can't do, apparently!).

Process Explorer GPU load by process

share|improve this answer

Process Hacker, the open source alternative to Process Explorer, also shows the GPU usage.

Press CTRL+I to get the system information Window and click on the GPU Tab, so get a graph with the GPU usage.

enter image description here

share|improve this answer
By process or just total? – Mahmoud Al-Qudsi Apr 21 '13 at 17:56
hover over the high spikes in the graph and you get a tooltip. But the tool also offers a GPU column. – magicandre1981 Apr 21 '13 at 18:52

AMD Catalyst Control Center has an indicator of GPU activity in the Performance section. I would assume nVidia software has a similar output. It's usually associated with overclocking.

enter image description here

Otherwise you can download the free GPU-Z. It includes a GPU Load output.

enter image description here

share|improve this answer
Thanks, but you misunderstand my question - I need to identify the process responsible for the GPU load. – Mahmoud Al-Qudsi Apr 21 '13 at 1:39
Ah relied more on the title. I'm going to edit it to be more clear. – Brad Patton Apr 21 '13 at 2:10
Thanks, appreciate it. – Mahmoud Al-Qudsi Apr 21 '13 at 2:41

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.