Post-processing on HPC with CFX-Post
Hi everyone,
I have a large quantity of results files which I need to post process in the same way (extract data, no visuals rendering). For this I have written a session file which iterates through the different timesteps extracting some data and writing it to a file. Since I have many results files and that process takes quite a while I want to do this on the HPC running slurm. I do manage to run the session file on the login node (which I really should not ) but I cannot create a job, submit the job and successfully run the script.Does anyone have experience with this? I am using PuTTy and Xming. Though I use Xming for X11 forwarding, this does not work when submitting it as a job. I am getting the following error message: "viewer_ogl.exe: cannot connect to X server localhost:10.0" Many thanks for your help. |
I'm not answering your question. But I believe the drive type is the most important factor for this. An SSD or better yet, an NVME SSD would be much faster than a spinning disk hard drive, so use one of those as your working directory to speed this up.
|
Thank you Erik for your response.
I am currently using an SSD. However, it is not necessary the speed which makes me move to the HPC, but the amount of simulation I have to extract data from. Being able to do this on the HPC frees up my workstation for other work. |
All times are GMT -4. The time now is 10:26. |