PVbatch running in a single thread

ParaView Support
1

Hi, I am new to paraview and am struggling with something that seems basic.

I downloaded the full version of paraview 5.9.1 to my desktop machine with Nvidia gpu and am able to script it to run my animation job (generate a surface changing over time) and this runs in parallel fine, leveraging all the cores available.

I do want the same setup to run on an AWS machine with a Nvidia GPU, and it does run and generate results. But the PVbatch command uses a single core only instead of the 8 available.

The only difference is, I am using the mpi egl version of the binary. (The full version doesn’t work presumably because the server is headless)

Don’t really want to build and run on a cloud machine, and using the binary would greatly simplify my workflow.

Any pointers would be greatly appreciated

2

I was able to figure out a solution by having mpiexec invoke paraview batch. Just had to make sure I was giving full paths for everything, and also setting my LD_LIBRARY_PATH appropriately as per the manual.

in my case it was something like this.

LD_LIBRARY_PATH=/opt/ParaView-5.9/lib/
/opt/ParaView-5.9/bin/mpiexec -np 30 /opt/ParaView-5.9/bin/pvbatch /path/to/my/anim_temp.py

Hope this helps someone else in a similar predicament.

1 Like