Comments (9)
At the top of the message you shared, I see No architecture specified, cannot estimate job costs.
Looks like your server has more than one architecture, it could be that it requires you to specify one of them. Try specifying it in the submit script, and see if it makes any difference (consult your system admin if you're not sure)
Alternatively, the sbatch command might not be defined appropriately. When you type which sbatch
, does it print exactly /usr/bin/sbatch
, or does it have a different path? If different, make sure to update it in settings (along with all other server commands)
from arc.
How to specify my architecture in submit script?
By the way, is that right to change the string 'pharos' to 'local' in line 198 in arc/job/submit.py in my case?
from arc.
Were you able to submit QChem jobs to this server outside ARC? If so, the submit script you specify in arc/job/submit.py should be very similar to the one you previously used. If you do need to specify the architecture, then only someone who knows your server (admin?) can help, I wouldn't know.
I'm not sure replacing 'pharos'
with 'local'
is what you want. The example scripts ARC has under 'pharos'
are for an OGE server, not a Slurm server.
Could you verify whether your server has OGE or Slurm as its cluster software? (or something else?) Also, upload here a submit script you were using for QChem outside ARC, and I'll try to assist you with the modifications.
from arc.
I can submit Qchem job in my server outside ARC. I specify my architecture by #SBATCH -C haswell
. My server is using Slurm as its cluster software, and my Qchem and ARC are built on the same server. so I change the line 52 in arc/settings.py from 'cluster_soft': 'OGE'
to 'cluster_soft': 'Slurm'
. After runing the Qchem job, it shows KeyError: 'local'
, the trace message is below.
then I see my trace message {'c3ddb': ['gaussian', 'orca'], 'rmg': ['molpro', 'gaussian', 'gromacs'], 'pharos': ['onedmin', 'qchem', 'molpro', 'gaussian', 'gaussian03_pharos']}
, I feel so weird that 'local'
is not a key in this dictionary and qchem is in the 'pharos'
,so I change string 'pharos'
to 'local'
.
After running, it shows No architecture specified, cannot estimate job costs.
Below is my submit script using for QChem outside ARC. I think I do something wrong in some parts.
#!/bin/bash -l
#SBATCH -N 1
#SBATCH -t 00:30:00
#SBATCH -q debug
#SBATCH -L SCRATCH
#SBATCH -C haswell
#SBATCH --ntasks-per-node=1
python ~/ARC/ARC.py arc_demo_1.yml
from arc.
Looks like a submit script for ARC, not QChem.
Your submit.py should look something like this (remove all other servers):
#!/usr/bin/env python
# encoding: utf-8
"""
Submit scripts
sorted in a dictionary with server names as keys
"""
##################################################################
submit_scripts = {
'local': {
# QChem
'qchem': """#!/bin/bash -l
#SBATCH -N 1
#SBATCH --time={t_max}
#SBATCH -q debug
#SBATCH -L SCRATCH
#SBATCH -C haswell
#SBATCH -J {name}
#SBATCH -n {cpus}
#SBATCH --mem-per-cpu {mem_per_cpu}
echo "============================================================"
echo "Job ID : $SLURM_JOB_ID"
echo "Job Name : $SLURM_JOB_NAME"
echo "Starting on : $(date)"
echo "Running on node : $SLURMD_NODENAME"
echo "Current directory : $(pwd)"
echo "============================================================"
export QC=/opt/qchem
export QCSCRATCH=/scratch/{un}/{name}
export QCLOCALSCR=/scratch/{un}/{name}/qlscratch
mkdir -p /scratch/{un}/{name}/qlscratch
qchem -nt {cpus} input.in output.out
rm -r /scratch/{un}/{name}
""",
}
}
from arc.
Sorry, the last two lines should be
module load qchem/5.1
qchem -nt 1 hydrogen.in hydrogen_b3lyp_opt.out
I change my submit.py to the one you give. However, it shows sbatch: error: Batch job submission failed: Requested time limit is invalid (missing or exceeds some limit)
. Since my server has different max time for different QOS(ex: debug have 30 min max_time). I am not sure whether #SBATCH --time={t_max}
will cause exceeding or not. Can I just type #SBATCH --time= 00:30:00
in this case? or do something else?
from arc.
Yes - in that case just do #SBATCH -t 00:30:00
as you had in your original script.
(change it if you submit to a different queue than the debug)
from arc.
Thank you so much. The submit.py is Ok for my case. I try to run hydrogen with sp job, but in the file I see that ARC go to run opt job for the molecules I didn't give, and cause submitting too much job in a sudden, so reach this QOS's job submission limitation. Why ARC would want to run these opt job for these molecule and doesn't run the job I submitting in the yml file?
Using Theano backend.
ARC execution initiated on Tue Sep 17 06:29:57 2019
###############################################################
# #
# Automatic Rate Calculator #
# ARC #
# #
# Version: 1.1.0 #
# #
###############################################################
The current git HEAD for ARC is:
ccdb557aa3f369b0de24da9a84ce9f9c3ad6dbb8
Sun Sep 15 21:43:52 2019 -0400
Starting project arc_demo_1
Considering the following job types: ['sp']
Using the following ESS settings:
{'qchem': ['local']}
Warning: Not using a fine grid for geometry optimization jobs
Warning: Not running rotor scans. This might compromise finding the best conformer, as dihedral angles won't be corrected. Also, entropy won't be accurate.
Using b3lyp/6-31g* for refined conformer searches (after filtering via force fields)
Using b3lyp/6-31g* for TS guesses comparison of different methods
Using b3lyp/6-31g* for geometry optimizations
Using b3lyp/6-31g* for frequency calculations
Using b3lyp/6-31g* for single point calculations
Using b3lyp/6-31g* for rotor scans
Warning: Could not determine appropriate Model Chemistry to be used in Arkane for thermochemical parameter calculations.
Not using atom energy corrections and bond additivity corrections!
Using b3lyp/6-31g*//b3lyp/6-31g* as a model chemistry in Arkane
Warning: :root:No frequency scaling factor found for model chemistry b3lyp/6-31g*. Assuming a value of unity. This will affect the partition function and all quantities derived from it (thermo quantities and rate coefficients).
Could not determine the harmonic frequencies scaling factor for b3lyp/6-31g* from Arkane.
Calculating it using Truhlar's method:
FREQ: A PROGRAM FOR OPTIMIZING SCALE FACTORS (Version 1)
written by
Haoyu S. Yu, Lucas J. Fiedler, I.M. Alecu, and Donald G. Truhlar
Department of Chemistry and Supercomputing Institute
University of Minnesota, Minnesota 55455-0431
CITATIONS:
1. I.M., Alecu, J. Zheng, Y. Zhao, D.G. Truhlar, J. Chem. Theory Comput. 2010, 6, 9, 2872-2887,
DOI: 10.1021/ct100326h
2. H.S. Yu, L.J. Fiedler, I.M. Alecu,, D.G. Truhlar, Computer Physics Communications 2017, 210, 132-138,
DOI: 10.1016/j.cpc.2016.09.004
starting ARC...
Considering the following job types: [u'opt', u'conformers', u'sp', u'1d_rotors', u'freq', u'fine']
Computing scaling factors at the b3lyp/6-31g* level of theory...
Using the following ESS settings:
{'qchem': ['local']}
Running job opt_a45 for C2H2
Running job opt_a46 for CH4
Running job opt_a47 for CO2
Running job opt_a48 for CO
Running job opt_a49 for F2
Running job opt_a50 for CH2O
sbatch: error: QOSMaxSubmitJobPerUserLimit
sbatch: error: Batch job submission failed: Job violates accounting/QOS policy (job submit limit, user's size and/or time limits)
Error: The following command is erroneous:
Traceback (most recent call last):
File "/global/homes/s/shcheng/ARC/ARC.py", line 76, in <module>
main()
File "/global/homes/s/shcheng/ARC/ARC.py", line 69, in main
arc_object = ARC(input_dict=input_dict, project_directory=project_directory)
File "/global/u2/s/shcheng/ARC/arc/main.py", line 405, in __init__
self.check_freq_scaling_factor()
File "/global/u2/s/shcheng/ARC/arc/main.py", line 1015, in check_freq_scaling_factor
level, ess_settings=self.ess_settings, init_log=False)[0]
File "/global/u2/s/shcheng/ARC/arc/utils/scale.py", line 108, in determine_scaling_factors
ess_settings=ess_settings, job_types=job_types, allow_nonisomorphic_2d=True)
File "/global/u2/s/shcheng/ARC/arc/scheduler.py", line 363, in __init__
self.run_opt_job(species.label)
File "/global/u2/s/shcheng/ARC/arc/scheduler.py", line 805, in run_opt_job
job_type='opt', fine=False)
File "/global/u2/s/shcheng/ARC/arc/scheduler.py", line 659, in run_job
self.job_dict[label][job_type][job.job_name].run()
File "/global/u2/s/shcheng/ARC/arc/job/job.py", line 805, in run
self.job_status[0], self.job_id = submit_job(path=self.local_path)
File "/global/u2/s/shcheng/ARC/arc/job/local.py", line 98, in submit_job
stdout = execute_command(cmd)[0]
File "/global/u2/s/shcheng/ARC/arc/job/local.py", line 43, in execute_command
' sbatch path required in the submit_command dictionary.'.format(e.message))
arc.arc_exceptions.SettingsError: The following command is erroneous:
To correct the command, modify settings.py
Tips: use "which" command to locate cluster software commands on server.
Example: type "which sbatch" on a server running Slurm to find the correct sbatch path required in the submit_command dictionary.
from arc.
Glad to hear that it works! I addressed your question in #197.
Good luck!
from arc.
Related Issues (20)
- ARC uses a fine grid although it deduced that it shouldn't
- Using bruteforce torsional scans blocks up servers storange and queue
- Trsh Orca auxilliary basis demand HOT 1
- Test
- " Q-Chem fatal error occurred in module qparser/new_read_rem.C, line 591:".
- ARCSpecies does not show atom labels for the mol object if it was generated from a `species_dict`
- ValueError in the "coordinates conversions" tutoial HOT 4
- Bad Conformer Id error HOT 1
- ARC's report of currently running jobs
- Implement the HEAT protocol
- ARC didn't find a correct conformer for a singlet carbene species
- rmgdb object does not load families if they are already loaded HOT 2
- Fail to copy reaction
- We should implement the HEAT protocol in ARC HOT 1
- ARC conda environment is too stiff to be created HOT 2
- Unblocking Python 3.11
- Consider implementing MRCC
- Could not determine reaction family for reaction C3H6-2 + O <=> C3H6O HOT 1
- parse_frequencies doesn't parse imaginary frequencies correctly.
- trsh: scan_qualitiy_check
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from arc.