How to run Paraview on multiple GPUs in a single node

Hello,

I compiled 5.6 superbuild version with EGL and MPI support on my linux machine. I have an instance with 4 GPUs and 24 cores but since I have EGL capabilities, I want to use all 4 GPUs. The following is a command I tried:

pvbatch --egl-device-index=0,1,2,3 post_scalar_V3.py

The issue is that upon reviewing the GPU usage, only the GPU with index 0 is being utilized. What’s the correct command to use all the GPUs in my machine? See the nvidia-smi output. As you see at very bottom, only one GPU is being utlized.

| NVIDIA-SMI 470.82.01    Driver Version: 470.82.01    CUDA Version: 11.4     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA A10G         On   | 00000000:00:1B.0 Off |                    0 |
|  0%   29C    P0    73W / 300W |   1717MiB / 22731MiB |     24%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  NVIDIA A10G         On   | 00000000:00:1C.0 Off |                    0 |
|  0%   25C    P8    18W / 300W |      0MiB / 22731MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   2  NVIDIA A10G         On   | 00000000:00:1D.0 Off |                    0 |
|  0%   25C    P8    10W / 300W |      0MiB / 22731MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   3  NVIDIA A10G         On   | 00000000:00:1E.0 Off |                    0 |
|  0%   26C    P8    16W / 300W |      0MiB / 22731MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A     40294      G   ...aview-egl-5.6/lib/pvbatch     1717MiB |
+-----------------------------------------------------------------------------+

Hi @Bruno_Sotelo ,

You need to use MPI for that, and specify different device for each mpi process.

Best,