From: Gerald Keller (gerald.keller_at_uni-wuerzburg.de)
Date: Sun Jul 31 2022 - 03:39:13 CDT
I successfully was using namd3-alpha10 on a CentOS7 system for plain MD simulations.
This worked fine so far, but now I am getting an error: Low global CUDA exclusion count.
This is only observed for one prepared protein-ligand system on a RTX 2080 TI.
When looking for this error, I found that this could be a bug in the namd-version (https://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2020-2021/0466.html).
Not sure if this would cover my issue, but I then tried to use namd3-alpha13 to see if the error still remains, but now the library libmvec.so.1 is missing. I found out, that this is a glibc-2.2x and higher library. CentOS7 provides glibc-2.17, only. Compiling a newer branch of glibc-2.17 would need to upgrade the provided make-3.82 program as well.
Getting this dependencies to work on my machine could be arbitrary difficult. So I wanted to ask if you could put the libmvec.so.1 library inside the lib-folder of the namd3-alpha release? Of course I do not now if other missing packages would follow.
Or do you have any other idea how to get the alpha13 running on CentOS7 or get rid of the error messeage posted above?
This archive was generated by hypermail 2.1.6 : Tue Dec 13 2022 - 14:32:44 CST