[mahmoodn]

We have a Quadro M2000 and it is working and GPU-Z detects it. So, I opened Fluent and tried the following scenario:

I set the parallel options to
(processes) number of processors to 1
(GPGPU per machine) to "none"

With nvidia-smi command, I still see that the cortex process is running on the GPU and one CPU is 100%. What does that mean? Also, the run time is slow and I it is reasonable for one CPU runtime.

With this observation, I can not verify that how much GPU speedsup the execution time. I can not verify
1- If GPU is actually running the code.
2- If only one CPU is engaged.
3- If the problem is big.

Any thought?