The example you showed us, was a k-parallel job on only one node.
To fix this, just set USE_REMOTE to zero (either in $WIENROOT
permanently, or temporarily in your submitted job script.
Another test would be to make a new wien2k-installation using
"ifort+slurm" in siteconfig. It may work
With apologies to Lukasz and Miro, there are some inaccurate statements
being made about how to use Wien2k in parallel -- the code is more complex
(and smarter). Please read carefully section 5.5 in detail, the read it
again. Google what commands such as ssh, srun, rsh, mpirun do.
If your cluster
Dear Miro,
On my cluster it works by a command
salloc -p cluster_name -N6 sleep infinity &
This particular command allocates 6 nodes. You can find which ones by
squeue command. Then passworless to these nodes is allowed in my
cluster. Then in .machines I include the names of these nodes and
Dear all,
ad: https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg22588.html
" In order to use multiple nodes, you need to be able to do passwordless ssh to
the allocated nodes (or any other command substituting ssh). "
According to our cluster admin, one can use (maybe) 'srun' to
it crashed with the message "Host key verification failed. "
Seems that your cluster does not allow ssh to an allocated node.(Ask
your sys admin).
In $WIENROOT/WIEN2k_parallel_options there are variables like
USE_REMOTE. If set to zero, ssh is not used and you can run in
parallel,
The "Host key verification failed" is an error from ssh [1].
Thus, it seems like you need to fix your ssh so that WIEN2k can connect
to your remote node (in your case below, it looks like the remote node
is lxbk1177).
It looks like there is an ssh example on slide 10 of the WIEN2k
Dear Professor Blaha,
thanks, I used PATH variable extension instead of linking;
it crashed with the message "Host key verification failed. "
Here the content of file
/lustre/ukt/milias/scratch/Wien2k_23.2_job.main.N1.n4.jid3009460/LvO2onQg/.machines:
1:lxbk1177
1:lxbk1177
1:lxbk1177
Dear Miro,
It is hard to give your a meaningful answer with little info, but I will
try my best guess because I needed to set this up recently. I assume
that you want to use k-parallel and you don't have mpi.
With a serial job you automatically run on a single node. Single node is
a
Well, the important files (output and error log files from slurm, the
dayfile and all error files, .machines) are not present.
Also, in this way it is nearly impossible to see which files are recent
and which are older ones.
The scripts are so complicated, that I cannot follow them Why
Hello,
I am able to run serial SCF via SLURM
https://github.com/miroi/open-collection/blob/master/theoretical_chemistry/software/wien2k/runs/LvO2_on_small_quartz/wien2k/LvO2onQg/virgo_slurm_wien2kgnupar_fromdstart.01
but when trying parallel
10 matches
Mail list logo
