Use of Pyxaid2

[Kamal Dhungana]

I have encountered a problem while using Pyxaid2 for step2. I use the same batch script for Pyxaid and Pyxaid2. It is always work for Pyxaid, but does not work for Pyxaid2.

It gives me the following error massage:

srun: error: hpc-3-39: tasks 5,7-10,13: Exited with exit code 1

The job submission command  "os.system( "%s -n %s %s < %s.%d.in > %s.%d.out" % (BATCH_SYSTEM,NP,EXE,prefix0,t,prefix0,t) )"  in runMD function looks good. The only difference in  Pyxaid and Pyxaid2 in this regard is mpirun vs srun.  Did anyone encounter this issue before? please kindly let me know. Here is my batch script (qe module is already available) :

 #!/bin/sh

#SBATCH -J "job-submit"

#SBATCH -n 16

#SBATCH -t 4:00:00

#SBATCH -o test.out

#SBATCH -p backfill

#SBATCH --mail-type=ALL

module load gnu-openmpi

cd $SLURM_SUBMIT_DIR

# Setup all manual parameters here

# Must be ABSOLUTE paths

NP=16

exe_qespresso=/opt/hpc/gnu/bin/qe_pw.x

exe_export=/opt/hpc/gnu/bin/qe_pw_export.x

res=/kbdhungana/project1/Lib-PYx2-test/Pyxaid2/Tut1_basics/step2/res

# These will be assigned automatically, leave them as they are

param1=

param2=

# This is invocation of the scripts which will further handle NA-MD calclculations

# on the NAC calculation step

python -c "from PYXAID2 import *

params = { }

params[\"NP\"]=$NP

params[\"EXE\"]=\"$exe_qespresso\"

params[\"EXE_EXPORT\"]=\"$exe_export\"

params[\"start_indx\"]=$param1

params[\"stop_indx\"]=$param2

params[\"wd\"]=\"wd_test\"

params[\"rd\"]=\"$res\"

params[\"minband\"]=1

params[\"nocc\"]=102

params[\"maxband\"]=122

params[\"nac_method\"]=1

params[\"wfc_preprocess\"]=\"restore\"

params[\"do_complete\"]=1

params[\"prefix0\"]=\"x0.scf\"

params[\"prefix1\"]=\"x1.scf\"

params[\"compute_Hprime\"]=1

print params

runMD.runMD(params)

"

Best Regards,

kamal

[Alexey Akimov]

Hi Kamal,

It looks like you are using somewhat outdated submit file. Have you tried any of the examples found here: https://github.com/Quantum-Dynamics-Hub/pyxaid2/tree/dev/example ? Especially 1 and 2 (the other two are more of experimental). 

Only, you are better set the environment variables in your .bash_profile file, not in the submit file, like in the example. 

Also, the variable MPIRUN should be set "srun" if you use SLURM. 

Perhaps, Wei Li could provide more guidance on the use of Pyxaid2

Best,

Alexey

[Wei Li]

Hi Kamal,

Yes, as Alexey commented, please try the examples in the dev branch. Please noted that the submit file pyxaid2 is different from pyxaid. Also, pay attention to the parameters need be set in the submit file of pyxaid2. 

Some of the functionality in pyxaid2 is still under development.

Please let us know if you have more concerns.

Best,

Wei

[Shudong Wang]

Dear Wei Li,

I ran Pyxaid2 in step2 ans then swithch to step3, but I got a error that compliant cannot find the file:

params["Haa_re_prefix"] = rt+"/res/0_Hvib_aa_"
params["Haa_re_suffix"] = "_re"
params["Haa_im_prefix"] = rt+"/res/0_Hvib_aa_"
params["Haa_im_suffix"] = "_im"
params["Hab_re_prefix"] = rt+"/res/0_Hvib_ab_"
params["Hab_re_suffix"] = "_re"
params["Hab_im_prefix"] = rt+"/res/0_Hvib_ab_"
params["Hab_im_suffix"] = "_im"
params["Hbb_re_prefix"] = rt+"/res/0_Hvib_bb_"
params["Hbb_re_suffix"] = "_re"
params["Hbb_im_prefix"] = rt+"/res/0_Hvib_bb_"
params["Hbb_im_suffix"] = "_im"

I have checked the dir of res and did not find them. How could we got these file when running step2?

Thanks!

Shudong

[Wei Li]

Hi Shudong,

Please wait for sometimes we are in the progress for step3. We will let you all know once it is ready to be used.

Wei

[Alexey Akimov]

Dear Shudong,

As Wei suggested, some of the functionality in Pyxaid2 is still under construction. Nonetheless, you should be able to use the Pyxaid1-type functionality in Pyxaid2 (that is the dynamics without SOC). In that sense, the filenames that you generate in step2 may be different from the example setup, but step3 calculations should be able to take the different set of filenames that is generated in step2. I think, the examples put forward by Wei should give the idea on how the Pyxaid2 calculations in steps 2 and 3 work.

Best,

Alexey

[Wei Li]

Hi Shudong and Alexey,

I think we need to change the step 3 code a bit, remove and add some stuff. So maybe need sometime to finish it.
But in principle if we want to use the files generated from the step2 in pyxaid2, it is better use the step3 calculation in Pyxaid1. The step3 calculation in pyxaid2 is actually similar as pyxaid1 right now if we don’t want to use the new functionality, I.e., multiple kpoint, spin relaxation with soc, demagnetizations, etc. . Just need to change the prefix for these files in py-scr3.py. And also the 0_Ham_ files contained vibronic Ham will be what needed.