Hi,
I am trying to set up WIEN2k ver 23.2 to run on a SLURM cluster. I
have gotten it to work with SCALAPACK, runnning with a slurm batch
submission script through w2web by following the examples.
I have two issues.
1. Is it possible to make the “x dstart” button in the initialize web
So it seems to works now.
The last messages are probably because you are using -it with a
perfectly converged calculation.
Remove (temporarely) the -it flag from the runsp_lapw
and remember: -it may only be faster for surfaces and large cells.
On 5/7/19 4:30 PM, webfin...@ukr.net
Dear Prof. Blaha,
I'm using intel mpi 2019.3.199
the scalapack and blacs libs are located in the intel
compilers_and_libraries_2019.3.199 folder
OPTIONS file:
current:FOPT:-O1 -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML -traceback
-assume buffered_io -I$(MKLROOT)/include
current:FPOPT:-O1
The "Permission denied
(publickey,gssapi-keyex,gssapi-with-mic,password)" comes up with
different causes in a Google search. One time, that error seemed to go
away with a user by having them ssh into the nodes and fix the ssh file
permissions following the webpage:
Not enough info. I briefly checked your wiki (I have no idea of French),
but you seem to have Intelmpi (which I would recommend).
What mpi are you loading ?
Did you load all modules also in the batch job
What scalapack ?
What blacs-library ?
Post your OPTION files from $WIENROOT
and also
Dear Prof. BlahaThank you!
The description of script for cluster is here
https://redmine.mcia.univ-bordeaux.fr/projects/cluster-curta/wiki/Slurm
(unfortunately it is in french and I'm not strong in cluster structures)
yes, the cluster uses "module" system. I'v used commands like "module load
So it seems that your cluster forbids to use ssh (even on assigned
nodes). If this is the case. you MUST use USE_REMOTE=0 and with
k-parallel mode you can use only one node (32 cores).
For mpi I do not know. There should be some "userguide" (web-site,
wicki, ...) for your cluster, where all
Dear Prof. Blaha
thank you for the explanation!
Sorry, I should put hostname in quotes. Script I used is based on that in the
WIEN-FAQ and produce .machines based on the nodes provided by the slurm:
for k-points:
#
1:n270
1:n270
1:n270
1:n270
1:n270
granularity:1
extrafine:1
for mpi:
#
When setting USE_REMOTE=0 it means, that you do not use "ssh" in
k-parallel mode.
This has the following consequences:
What you write for "hostname" in .machines is not important, only the
number of lines counts. And it will span as many k-parallel jobs as you
have lines (1:hostname), but they
WIEN2k 18.2 usersguide (pg. 237) has:
USE_REMOTE [0|1] determines whether parallel jobs are run in background
(on shared memory machines) or using ssh.
Since you are utilizing ssh-copy-id for using ssh, you most likely need
USE_REMOTE=1 [
Hi,
WIEN2k has a usersguide, where the different parallelization modes are
extensively described.
On a cluster with a queuing system (like SLURM) it should not even be
possible to access nodes (except the frontend) via ssh without using
SLURM (on our SLURM machine ssh is possible only to
Dear All,
As I am currently trying to get Wien2k running on Stampede (also SLURM),
let me add a little clarification without disagreeing with anything Peter
said.
A typical workflow in Wien2k is (very simplified) an iterative loop
controlled by csh scripts:
1) A single serial multithreaded or
And, at least in an interactive job, none of these work...
Sigh. The man page of srun is also inconsistent with the actual srun
used
On Wed, Nov 11, 2015 at 7:08 AM, Laurence Marks
wrote:
> Thanks. Does it produce a one line/entry list?
>
> I found a variant that
Thanks. Does it produce a one line/entry list?
I found a variant that might also work (need to test)
# Generate Machinefile for mpich such that hosts are in the same order as
if run via srun
srun -l /bin/hostname | sort -n | awk '{print $2}' > MACHINEFILE
# Run using generated Machine file:
On 11/11/2015 03:07 PM, Laurence Marks wrote:
> And, at least in an interactive job, none of these work...
>
> Sigh. The man page of srun is also inconsistent with the actual srun
> used
Not sure if it will be better for your purposes, but what I use is
scontrol show hostnames
The commands
srun hostname -s >slurm.hosts
set proclist=`cat slurm.hosts|sort`
within a slurm-job give you a list of your hosts.
Am 11.11.2015 um 01:04 schrieb Laurence Marks:
Does anyone know the "machines" format for slurm? I want to expand
Machines2W so it can use it (& I can use a slurm
16 matches
Mail list logo