Thom Group Wiki - User contributions [en]

Getting Started with cerebro

2021-07-27T19:18:57Z

Cak49:

=Overview=
cerebro is a compute cluster. Unlike on the workstations, jobs are not run directly from the command line (e.g. just doing <code>qchem file.in file.out</code>). Instead, jobs are sent to a queue which is managed by <code>SLURM</code>. To submit a job, you need to write a submit file with information about your job then submit it to the queue using the <code>sbatch</code> command. Jobs create a <code>slurm-<jobid>.out</code> file which contains their terminal output. If you think something might have gone wrong with a job (e.g. something crashed, it ran out of time, etc...) the <code>slurm</code> file is usually a good place to start looking for the issue.

General information about cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/cerebro-user-notes here].

=Helpful Commands=
==<code>sbatch</code>==
<code>sbatch</code> is used to submit a job to the queue, e.g.
<pre>sbatch $old $long submit_file</pre>

You can specify whether a job runs on the old nodes (12 CPUs max) or the new nodes (16 CPUs max) using <code>$old</code> and <code>$new</code>, respectively.

You can also specify the partition on which you would like the job to run using one of four options:
{| class="wikitable"
! style="width: 8ex;" |'''Partition''' || style="width: 14ex;" | '''sbatch Option''' || style="width: 10ex;" | '''Time Limit'''|| style="width: 45ex;" | '''Other Notes'''
|-
| TEST || $test || 4 hours || highest priority partition - jobs should run right away
|-
| LONG || $long || 48 hours || default partition
|-
| XLONG || $xlong || 7 days || 96 core limit
|-
| XXLONG || $xxlong || 30 days || 56 core limit
|}
More information about SLURM on cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/slurm-usage here] and [https://www.ch.cam.ac.uk/computing/cerebro-queues-and-parallel-job-configuration here], and general information about SLURM can be found [https://slurm.schedmd.com/overview.html here].

==<code>squeue</code>==
To see all the jobs in the queue, do
<pre>squeue</pre>

To see all the jobs that you have queued or running, do
<pre>squeue -u <your-crsid></pre>

==<code>scancel</code>==
To cancel a job, do
<pre>scancel <JOBID></pre>
'''Make sure you have the right job ID!'''

More information on queuing is available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/Queuing here].

=Using Q-Chem=

If you have not used Q-Chem on cerebro before, first check whether you have access with <code>which qchem</code>. If that does not return anything, you'll need to add a few lines to your <code>.bashrc</code> file. For Q-Chem 5.3, add
<pre># QChem
export QC=trunk
source /home/maf63/code/qcsetup-general.bash
source ~/.slurmrc</pre>
to your <code>.bashrc</code> file, then do <code>source ~/.bashrc</code>. Doing <code>which qchem</code> should give <code>/sharedscratch/maf63/qchem-general/bin/qchem</code>.

Once you have Q-Chem set up, you'll be able to submit jobs. Here is a generic submit file for Q-Chem on cerebro:
<pre>#!/bin/bash

# Set default outfile if not defined
outfile=${outfile:-qchem.out}
scratch=${scratch:-qchem.scratch}
export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
echo "Using Q-Chem: " $QC
rm -rf $QCSCRATCH/$scratch
cp -r $scratch $QCSCRATCH/$scratch
qchem -nt $SLURM_CPUS_ON_NODE -save $infile $outfile $scratch

# Recover the scratch directory (optional)
# cp -r $QCSCRATCH/$scratch/* $scratch/</pre>

To use this submit file, copy it to <code>submit.qchem</code>, then do
# <code>export infile=<your_qchem_input_file></code>
# <code>export outfile=<name_of_qchem_output_file></code>
# <code>export scratch=<name_of_a_scratch_directory></code>
# <code>sbatch <your options> submit.qchem</pre></code>

=Using QCMagic=
If you have not used QCMagic on cerebro before, you will need to install it first - instructions are available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic#Installation here].

Here is a submit file template for a QCMagic job:
<pre>#!/bin/bash

export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
runscanSurface.py -p 12 -L --read-minima=1 etc... </pre>

If you copy this information to a file called <code>submit.qcmagic</code>, you can then submit it with
<pre>sbatch $old $test submit.qcmagic</pre>

You can include multiple commands in a submit file, e.g.:
<pre>#!/bin/bash

export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
runscanSurface.py -p 12 -L --read-minima=4 --read-only output_file.out state_4_read > term_read
runqcSDExtract.py -p 12 -L --reconverge --rem="SCF_CONVERGENCE 10" state_4_read.sd state_4 > term_reconv
runcombineSDXC.py -i state_1.sd state_2.sd state_3.sd state_4.sd -o states_1234 > term_sdxc
runrunSDXC.py -p 12 -L --template=template.in states_1234.sd states_1234 > term_new</pre>

Getting Started with cerebro

2021-07-27T18:27:36Z

Cak49:

=Overview=
cerebro is a compute cluster. Unlike on the workstations, jobs are not run directly from the command line (e.g. just doing <code>qchem file.in file.out</code>). Instead, jobs are sent to a queue which is managed by <code>SLURM</code>. To submit a job, you need to write a submit file with information about your job then submit it to the queue using the <code>sbatch</code> command. Jobs create a <code>slurm-<jobid>.out</code> file which contains their terminal output. If you think something might have gone wrong with a job (e.g. something crashed, it ran out of time, etc...) the <code>slurm</code> file is usually a good place to start looking for the issue.

General information about cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/cerebro-user-notes here].

=Helpful Commands=
==<code>sbatch</code>==
<code>sbatch</code> is used to submit a job to the queue, e.g.
<pre>sbatch $old $long submit_file</pre>

You can specify whether a job runs on the old nodes (12 CPUs max) or the new nodes (16 CPUs max) using <code>$old</code> and <code>$new</code>, respectively.

You can also specify the partition on which you would like the job to run using one of four options:
{| class="wikitable"
! style="width: 8ex;" |'''Partition''' || style="width: 14ex;" | '''sbatch Option''' || style="width: 10ex;" | '''Time Limit'''|| style="width: 45ex;" | '''Other Notes'''
|-
| TEST || $test || 4 hours || highest priority partition - jobs should run right away
|-
| LONG || $long || 48 hours || default partition
|-
| XLONG || $xlong || 7 days || 96 core limit
|-
| XXLONG || $xxlong || 30 days || 56 core limit
|}
More information about SLURM on cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/slurm-usage here] and [https://www.ch.cam.ac.uk/computing/cerebro-queues-and-parallel-job-configuration here], and general information about SLURM can be found [https://slurm.schedmd.com/overview.html here].

==<code>squeue</code>==
To see all the jobs in the queue, do
<pre>squeue</pre>

To see all the jobs that you have queued or running, do
<pre>squeue -u <your-crsid></pre>

==<code>scancel</code>==
To cancel a job, do
<pre>scancel <JOBID></pre>
'''Make sure you have the right job ID!'''

More information on queuing is available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/Queuing here].

=Using Q-Chem=

If you have not used Q-Chem on cerebro before, first check whether you have access with <code>which qchem</code>. If that does not return anything, you'll need to add a few lines to your <code>.bashrc</code> file. For Q-Chem 5.3, add
<pre># QChem
export QC=trunk
source /home/maf63/code/qcsetup-general.bash
source ~/.slurmrc</pre>
to your <code>.bashrc</code> file, then do <code>source ~/.bashrc</code>. Doing <code>which qchem</code> should give <code>/sharedscratch/cbh31/code/qchem5/qc521/bin/qchem</code>.

Once you have Q-Chem set up, you'll be able to submit jobs. Here is a generic submit file for Q-Chem on cerebro:
<pre>#!/bin/bash

# Set default outfile if not defined
outfile=${outfile:-qchem.out}
scratch=${scratch:-qchem.scratch}
export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
echo "Using Q-Chem: " $QC
rm -rf $QCSCRATCH/$scratch
cp -r $scratch $QCSCRATCH/$scratch
qchem -nt $SLURM_CPUS_ON_NODE -save $infile $outfile $scratch

# Recover the scratch directory (optional)
# cp -r $QCSCRATCH/$scratch/* $scratch/</pre>

To use this submit file, copy it to <code>submit.qchem</code>, then do
# <code>export infile=<your_qchem_input_file></code>
# <code>export outfile=<name_of_qchem_output_file></code>
# <code>export scratch=<name_of_a_scratch_directory></code>
# <code>sbatch <your options> submit.qchem</pre></code>

=Using QCMagic=
If you have not used QCMagic on cerebro before, you will need to install it first - instructions are available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic#Installation here].

Here is a submit file template for a QCMagic job:
<pre>#!/bin/bash

export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
runscanSurface.py -p 12 -L --read-minima=1 etc... </pre>

If you copy this information to a file called <code>submit.qcmagic</code>, you can then submit it with
<pre>sbatch $old $test submit.qcmagic</pre>

You can include multiple commands in a submit file, e.g.:
<pre>#!/bin/bash

export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
runscanSurface.py -p 12 -L --read-minima=4 --read-only output_file.out state_4_read > term_read
runqcSDExtract.py -p 12 -L --reconverge --rem="SCF_CONVERGENCE 10" state_4_read.sd state_4 > term_reconv
runcombineSDXC.py -i state_1.sd state_2.sd state_3.sd state_4.sd -o states_1234 > term_sdxc
runrunSDXC.py -p 12 -L --template=template.in states_1234.sd states_1234 > term_new</pre>

QChem

2021-07-27T14:22:02Z

Cak49:

Q-Chem is a commercial quantum chemistry package available for use in our group.

==Fragment Guesses with multiple bases==
To use basis set projection and fragment guesses you have to run the smaller basis set calculation with fragments first, and then read this into a second calculation which does the projection
<pre>$molecule
0 1
--
0 1
H 0.000000 -0.115747 1.133769
H 0.000000 1.109931 -0.113383
O 0.000000 0.005817 -0.020386
--
0 1
He 10000 0 0
$end

$rem
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVDZ
scf_guess fragmo
CORRELATION none
SYMMETRY false
SYM_IGNORE true
$end

@@@

$molecule
read
$end

$rem
scf_guess read
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVTZ
BASIS2 cc-pVDZ
CORRELATION none
MAX_SCF_CYCLES 600
SYMMETRY false
SYM_IGNORE true
basisprojtype ovprojection
$end</pre>

==INTDUMP / FCIDUMP ==
To create files in the FCIDUMP format (used e.g. by Hande) create a QChem input file:
<pre>
$molecule
0 1
H 0 0 0
H 1 0 0
$end

$rem
correlation idump
exchange hf
basis sto-3G
use_abelian_subgroup true
scf_guess_print 9999
$end
</pre>

and use the version of QChem currently (2020-02-20) available from:
<pre>
[you@liminal] $ export QC=public_o
[you@liminal] $ source ~ajwt3/code/qchem/qcsetup.bash
[you@liminal] $ qchem your-input-file.in
</pre>

This should produce a file called <code>INTDUMP</code>
<pre>
&FCI NORB= 2,NELEC= 2,MS2= 0
ORBSYM=1,6,
ISYM=1 UHF=.FALSE.
&END
0.62640249948715 1 1 1 1
0.19679058349422 1 2 1 2
0.6217067630807 2 2 1 1
0.19679058349422 2 1 1 2
0.19679058349422 2 1 2 1
0.65307074688898 2 2 2 2
-1.1108441795661 1 1 0 0
-0.58912100326925 2 2 0 0
-0.39980670964291 1 0 0 0
0.37823811728909 2 0 0 0
0.52917721026434 0 0 0 0
</pre>

==Troubleshooting SCF Metadynamics Calculations==

A full example for SCF Metadynamics can be found at [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic QCMagic SymmetryTools Tutorial (Section 7.1)]. This section provides more information on how to choose parameters in different situations.

If a calculation repeatedly converges on the same solution, adjusting one or more of these keywords can help to find different solutions.
* Increasing <code>SCF_MINFIND_INITNORM</code>, <code>SCF_MINFIND_INITLAMBDA</code> and <code>SCF_MINFIND_INCREASEFACTOR</code> can help get out of the minimum. However, setting <code>SCF_MINFIND_INITLAMBDA</code> too high can result in SCF solutions with very high energies, and <code>SCF_MINFIND_INITNORM</code> can get in the way of convergence, particularly when reading in a solution (in that case, set it to zero).
* Increasing <code>SCF_MINFIND_RESTARTSTEPS</code> helps if a solution is close to another minima but does not converge before the calculation restarts with new orbitals
* Changing <code>SCF_MINFIND_RANDOMMIXING</code> is generally helpful. +/- 15708 and 31416 are sensible values to try, and 07854 (which corresponds to halfway between swapping the orbitals) can also help.

If the resulting solutions are too high in energy,
* decreasing <code>SCF_MINFIND_INITNORM</code>
* decreasing <code>SCF_MINFIND_INITLAMBDA</code>
* decreasing <code>SCF_MINFIND_MIXENERGY</code> (even to 00005)
* adjusting <code>MOM_START</code>
* and using <code>SCF_MINFIND_MIXMETHOD</code> 1 or 2
can help to find lower-energy excited states.

==NOCI Tutorial==
This tutorial includes two possible methods for doing NOCI. The examples below show how to do NOCI at a single geometry, but these methods can easily be extended to cover NOCI at multiple points along a path.

===Method #1===
In this approach, we find minima using SCF Metadynamics and then do NOCI on them in the same Q-Chem output file. A sample Q-Chem input file for formaldehyde is shown below. In this example, we have Q-Chem generate four initial reference determinants using SCF Metadynamics by selecting <code>NOCI_REGEN 0</code>, <code>NOCI_DETGEN 3</code>, and <code>SCF_SAVEMINIMA 4</code>.

<pre>$molecule
0 1
H -0.0000000 0.9275885 1.1766889
C -0.0000000 0.0000000 0.6019825
H -0.0000000 -0.9275885 1.1766889
O 0.0000000 -0.0000000 -0.6001772
$end

$rem
EXCHANGE HF
CORRELATION NOCI
BASIS STO-3G
UNRESTRICTED true
MAX_SCF_CYCLES 1000
SCF_CONVERGENCE 10
MOM_START 1
SCF_SAVEMINIMA 4
SCF_MINFIND_INITNORM 05000
SCF_MINFIND_INITLAMBDA 00300
SCF_MINFIND_RANDOMMIXING 30000
SCF_MINFIND_MIXMETHOD 1
USE_LIBNOCI true
NOCI_REFGEN 0
NOCI_DETGEN 3
NOCI_NEIGVAL 4
NOCI_PRINT 5
$end</pre>

We then run

<pre>qchem formaldehyde.noci.in formaldehyde.noci.out</pre>

and look for the energies of the SCF solutions with

<pre>grep “Saving Minimum” formaldehyde.noci.out</pre>

which gives

<pre> Saving Minimum 1: -112.3535105301
Saving Minimum 2: -111.9177589859
Saving Minimum 3: -111.7116823375
Saving Minimum 4: -112.3571993219</pre>

Information about the resulting NOCI states is printed at the end of the output file. In this example, the energies and S^2 values can be examined with

<pre>grep -A5 "NOCI Energy" formaldehyde.noci.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.36268 -112.16945 -111.91685 -111.71168
NOCI <S^2>:
1 2 3 4
1 0.0835515 1.8508569 0.0321459 1.0273785</pre>

===Method #2===
In this method, we use QCMagic to read in SCF Metadynamics states in <code>.sd</code> files into a Q-Chem NOCI calculation.

In this example, we are starting with HF SCF solutions for formaldehyde in four separate <code>.sd</code> files (<code>form.s1.sd</code>, <code>form.s2.sd</code>, <code>form.s3.sd</code> and <code>form.s4.sd</code>) with the following energies:
<pre>-112.3535105299 Minimum 1
-111.9177348081 Minimum 2
-111.4909067597 Minimum 3
-111.8200842261 Minimum 4</pre>

First, we combine all <code>.sd</code> files for the states at a particular geometry using <code>runcombineSDXC.py</code>, e.g.:

<pre>runcombineSDXC.py -i form.s1.sd\
form.s2.sd\
form.s3.sd\
form.s4.sd\
-o form_1234 > terminal_sdxc</pre>

This creates an <code>.sd</code> file, <code>form_1234.sd</code>, which contains all the minima at that geometry. We then use <code>runrunSDXC.py</code> to apply a template for NOCI and call Q-Chem:

<pre>runrunSDXC.py -p 12 -L --template=noci.template form_1234.sd noci_form_1234 > terminal_noci</pre>

An example <code>noci.template</code> file for LIBNOCI is shown below:

<pre>$molecule
read
$end

$rem
EXCHANGE HF
CORRELATION NOCI
UNRESTRICTED TRUE
BASIS STO-3G
SCF_SAVEMINIMA 4
SCF_READMINIMA -4
NOCI_PRINT 10
USE_LIBNOCI TRUE
NOCI_REFGEN 1
NOCI_DETGEN 0
SKIP_SCFMAN TRUE
SYMMETRY OFF
$end

$noci
1 2 3 4
$end</pre>

In this template, we specify <code>NOCI_REFGEN 1</code> and <code>NOCI_DETGEN 0</code> since we are reading in our reference determinants and do not wish to generate any more. We use the <code>$noci</code> section to specify which states to read in: here, we read in all the states, though we could read in a subset of the states in the <code>.sd</code> file if we wanted to. In this example, we use <code>SKIP_SCFMAN TRUE</code> to avoid reconverging the solutions. If we wanted to reconverge them, we would omit this option and include any missing SCF options in the template.

We can now examine our output file, <code>noci_form_1234.g0.xc0.out</code>. First, we check the energies of the HF states we read in with

<pre>grep "Total Energy =" noci_form_1234.g0.xc0.out</pre>

which gives

<pre> Total Energy = -112.3535105266
Total Energy = -111.9177347312
Total Energy = -111.4909067030
Total Energy = -111.8200841138</pre>

and examine the energies and S^2 values of the NOCI solutions with

<pre>grep -A5 "NOCI Energy" noci_form_1234.g0.xc0.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.35436 -111.91688 -111.82008 -111.49091
NOCI <S^2>:
1 2 3 4
1 0.0000000 0.0000000 1.0107832 1.9977364</pre>

(N.B. If we compare the NOCI states from method #1 and method #2, we notice that they are different because we used different HF states in both methods. If we had used the same states in both examples, we would get the same NOCI results.)

===Further Information===
Finding SCF solutions with Metadynamics:
* [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic QCMagic SymmetryTools Tutorial (Section 7.1)]
* Q-Chem section on Troubleshooting Metadynamics Calculations: see above
* [https://manual.q-chem.com/5.3/sec_SCF-meta-dyn.html Q-Chem documentation on SCF Metadynamics (Section 4.9.2)]
* [https://manual.q-chem.com/5.3/subsubsec_SCFMetadynNOCI.html Q-Chem documentation on the LIBNOCI implementation of SCF Metadynamics (Section 4.9.3)]
NOCI
* [https://manual.q-chem.com/5.3/subsec_NOCI.html Q-Chem documentation for NOCI (Section 7.4)]
* Burton H. G. A.; Thom, A. J. W. General Approach for Multireference Ground and Excited States Using Nonorthogonal Configuration Interaction. J Chem. Theory Comput. 2019, 15, 4851.
Keywords for different implementations of NOCI in Q-Chem
* LIBNOCI (most recent implementation): <code>USE_LIBNOCI TRUE</code>
* older implementation: <code>USE_LIBNOCI FALSE</code> and <code>GEN_SCFMAN OFF</code>

Getting Started with cerebro

2021-07-25T00:03:06Z

Cak49: Added information about getting started with cerebro

=Overview=
cerebro is a compute cluster. Unlike on the workstations, jobs are not run directly from the command line (e.g. just doing <code>qchem file.in file.out</code>). Instead, jobs are sent to a queue which is managed by <code>SLURM</code>. To submit a job, you need to write a submit file with information about your job then submit it to the queue using the <code>sbatch</code> command. Jobs create a <code>slurm-<jobid>.out</code> file which contains their terminal output. If you think something might have gone wrong with a job (e.g. something crashed, it ran out of time, etc...) the <code>slurm</code> file is usually a good place to start looking for the issue.

General information about cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/cerebro-user-notes here].

=Helpful Commands=
==<code>sbatch</code>==
<code>sbatch</code> is used to submit a job to the queue, e.g.
<pre>sbatch $old $long submit_file</pre>

You can specify whether a job runs on the old nodes (12 CPUs max) or the new nodes (16 CPUs max) using <code>$old</code> and <code>$new</code>, respectively.

You can also specify the partition on which you would like the job to run using one of four options:
{| class="wikitable"
! style="width: 8ex;" |'''Partition''' || style="width: 14ex;" | '''sbatch Option''' || style="width: 10ex;" | '''Time Limit'''|| style="width: 45ex;" | '''Other Notes'''
|-
| TEST || $test || 4 hours || highest priority partition - jobs should run right away
|-
| LONG || $long || 48 hours || default partition
|-
| XLONG || $xlong || 7 days || 96 core limit
|-
| XXLONG || $xxlong || 30 days || 56 core limit
|}
More information about SLURM on cerebro is available on the department website [https://www.ch.cam.ac.uk/computing/slurm-usage here] and [https://www.ch.cam.ac.uk/computing/cerebro-queues-and-parallel-job-configuration here], and general information about SLURM can be found [https://slurm.schedmd.com/overview.html here].

==<code>squeue</code>==
To see all the jobs in the queue, do
<pre>squeue</pre>

To see all the jobs that you have queued or running, do
<pre>squeue -u <your-crsid></pre>

==<code>scancel</code>==
To cancel a job, do
<pre>scancel <JOBID></pre>
'''Make sure you have the right job ID!'''

More information on queuing is available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/Queuing here].

=Using Q-Chem=
If you have not used Q-Chem on cerebro before, first check if you have access with <code>which qchem</code>.

Once you have Q-Chem set up, you'll be able to submit jobs. Here is a generic submit file for Q-Chem on cerebro:
<pre>#!/bin/bash

# Set default outfile if not defined
outfile=${outfile:-qchem.out}
scratch=${scratch:-qchem.scratch}
export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
echo "Using Q-Chem: " $QC
rm -rf $QCSCRATCH/$scratch
cp -r $scratch $QCSCRATCH/$scratch
qchem -nt $SLURM_CPUS_ON_NODE -save $infile $outfile $scratch

# Recover the scratch directory (optional)
# cp -r $QCSCRATCH/$scratch/* $scratch/</pre>

To use this submit file, copy it to <code>submit.qchem</code>, then do
# <code>export infile=<your_qchem_input_file></code>
# <code>export outfile=<name_of_qchem_output_file></code>
# <code>export scratch=<name_of_a_scratch_directory></code>
# <code>sbatch <your options> submit.qchem</pre></code>

=Using QCMagic=
If you have not used QCMagic on cerebro before, you will need to install it first - instructions are available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic#Installation here].

Here is a submit file template for a QCMagic job:
<pre>#!/bin/bash

export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

# Run the calculation
runscanSurface.py -p 12 -L --read-minima=1 etc... </pre>

If you copy this information to a file called <code>submit.qcmagic</code>, you can then submit it with
<pre>sbatch $old $test submit.qcmagic</pre>

You can include multiple commands in a submit file.

Main Page

2021-07-24T23:44:40Z

Cak49: Added page on getting started with cerebro

''
# _________ # ___ ___ # ______ # ___ __ __ # # _______ # ______ # ______ # __ __ # ______ #
#/________/\ #/__/\ /__/\ #/_____/\ #/__//_//_/\ # #/______/\ #/_____/\ #/_____/\ #/_/\/_/\ #/_____/\ #
#\__.::.__\/ #\::\ \\ \ \ #\:::_ \ \ #\::\| \| \ \ # #\::::__\/__ #\:::_ \ \ #\:::_ \ \ #\:\ \:\ \ #\:::_ \ \ #
# \::\ \ # \::\/_\ .\ \ # \:\ \ \ \ # \:. \ \ # # \:\ /____/\ # \:(_) ) )_ # \:\ \ \ \ # \:\ \:\ \ # \:(_) \ \ #
# \::\ \ # \:: ___::\ \ # \:\ \ \ \ # \:.\-/\ \ \ # # \:\\_ _\/ # \: __ `\ \ # \:\ \ \ \ # \:\ \:\ \ # \: ___\/ #
# \::\ \ # \: \ \\::\ \# \:\_\ \ \ # \. \ \ \ \# # \:\_\ \ \ # \ \ `\ \ \ # \:\_\ \ \ # \:\_\:\ \ # \ \ \ #
# \__\/ # \__\/ \::\/# \_____\/ # \__\/ \__\/# # \_____\/ # \_\/ \_\/ # \_____\/ # \_____\/ # \_\/ #
# # # # # # # # # # #
''

Hello, Guten Abend, and Bienvenue to the new Thom Group Wiki. (Apparently evenings are good times to read the Wiki.)

In this Wiki you can find various pieces of useful information, such as when group meetings are, who is going to bring cake, how to run a particular calculation or perform a certain computer trick, who's using which computer in the group, and so on. You can also see how fun we are as a group by looking at our various photos.

=Group Calendar=
https://calendar.google.com/calendar/render?mode=day&date=20160601T153539#main_7%7Cday-1+23745+23745+23745

{{Special:IframePage}}

=Group Meetings=

Past Group Meetings [[Past Group Meetings|here]].

===Summer 2021===

{| class="wikitable"
! style="width: 20ex;" |'''Date''' || style="width: 18ex;" | '''Talk''' || style="width: 18ex;" | '''Cake'''
|-
|12th July 2021 || INFORMAL ||
|-
|20th July 2021 || INFORMAL ||
|-
|27th July 2021 || Cesar & Tarik ||
|-
|3rd August 2021 || Constance ||
|-
|9th August 2021 || INFORMAL ||
|-
|16th August 2021 || INFORMAL ||
|-
|23rd August 2021|| Benjamin ||
|-
|30th August 2021|| INFORMAL ||
|-
|6th September 2021 || Arta ||
|-
|13th September 2021 || ||
|}

=Summer 'Interns'=
Past Summer Interns [[Past Summer Interns|here]].
==2021==
{| class="wikitable"
! style="width: 18ex;" |'''Name''' || style="width: 18ex;" | '''Dates''' || style="width: 25ex;" | '''Project'''|| style="width: 18ex;" | '''Machine'''
|-
| Anna Bui || 19th Jul - 17th Sep || NOCI Carbon Capture || hypatia
|-
| Zian Wang || 19th Jul - 17th Sep || NOCI Photochemistry || obsidian
|-
| Brian Zhao || 19th Jul - 17th Sep || Stochastic Coupled Cluster || gritstone
|-
| Nick Lee || 19th Jul - 17th Sep || NOCI Singlet Fission || moonraker
|}

= [[Group_List | Group List]] =
The full timeline of all current and past group members is available [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/Group_List here].

An up to date list of group members is also available [http://www.ch.cam.ac.uk/group/thom here].

[[File:AJWT_group_photo_2_5_19-4926_taken_by_Nathan_Pitt_University_of_Cambridge-tighter.jpg|500px|thumb|center|Thom Group photo - taken by Nathan Pitt, ©University of Cambridge, May 2019]]
[[File:AJWT_Group_photo-5470_taken_by_Nathan_Pitt_University_of_Cambridge.jpg|500px|thumb|center|Thom Group photo - taken by Nathan Pitt, ©University of Cambridge, November 2017]]

= [[Computer_List | Computer List]] =
- Group computers available.

= [[Introduction | Introduction to basic shell commands]] =
- There are a number of tutorials available which document basic operations that are useful including:

# General bash commands
# Setting up cygwin
# Setting up [https://git-scm.com/book/en/v2/Git-on-the-Server-Generating-Your-SSH-Public-Key ssh keys] and general bash commands what are helpful
# Setting up ssh config files
# using [[GIT]]
# Useful cerebro queue commands
# Some [[vim]] tidbits

Currently undergoing construction.

= How to connect to department machines =

'''On Mac/WSL:'''

# Check you have an Admitto account and collect your password from https://www.ch.cam.ac.uk/computing/admitto-service
# In the terminal run the command ''ssh -X crsid@citadel.ch.cam.ac.uk''
# When prompted input your Admitto username (crsid) and password
# Once this has worked run the command ''ssh -X crsid@machinename'' in the terminal to log into a particular machine

Set up logging in without a password:

# Generate an ssh key by running the command ''ssh-keygen'' in the terminal, as described in detail in the section "Introduction to basic shell commands"
# Use ''cd ~/.ssh'' to navigate to the directory holding the keys, and copy the text from the file "id_rsa.pub" beginning with ssh-rsa. This is your public ssh key
# Log into the department citadel machine, and create / navigate to a directory called .ssh
# Use ''chmod 700 .ssh'' to set permissions for the directory
# Input ''vi authorized_keys'' to open up the vi text editor. Press i, then paste in your public ssh key. Press escape, then type '':wq'' and press enter to write and quit the editor
# Log into the particular machine you want to ssh to, and repeat steps 3 to 5
# Exit back to your machine and navigate to the home directory
# Input ''vi .ssh/config'' to open the vi text editor. Press i, then paste in the following, with your crsid and machine name in the places given:
Host citadel
User crsid
Hostname citadel.ch.cam.ac.uk
ForwardAgent yes
ProxyCommand none
Host machinename
Hostname machinename
ProxyCommand ssh citadel -W %h:%p
User crsid
ForwardAgent yes
ServerAliveInterval 60
ServerAliveCountMax 10

Press escape, then input '':wq'' followed by enter to exit the editor.
You can now log straight into citadel with ''ssh citadel'' or straight into your machine with ''ssh machinename''. The last two commands stop the ssh being killed if you are idle for too long. The ''ServerAliveInterval'' is how many seconds to ping a null packet, and the ''ServerAliveCountMax'' are how many consecutive times it needs to fail for the ssh to be killed.

= Useful Software =
# Using [[QChem]]
# Using [[QCMagic]]
# [[SimpleDMC]]
# [[MRCC]]

= Useful Information =
# Guidelines on [[Code Review]]
# How to do things relating to [[HANDE]]
# How to run PySCF or other [[Python software on Archer]]
# How to run QChem on [[darwin]]
# Where to get [[Travel Money]]
# Backed-up [[Storage]]
# [[How to get IQMol to run a local version of Q-Chem via SSH]]
# [[Slow ubuntu dash]]
# [[Persistent X]] sessions for remote working
# [[Paper submission]]
# [[Getting Started with cerebro]]

= [[Archiving_data | Archiving data for the university repository]] =

= [[Website_to_do | To-do list for the Thom Group Website]] =

= [[Group_activities | Group Activities]] =

= To-do list for the Wiki=

- A pretty picture

- A "How to:" page on setting up cygwin, ssh keys and general bash commands what are helpful

- A "How to:" page on using qchem

- A Pretty picture for the $wgLogo

QChem

2021-07-24T21:19:18Z

Cak49: Fixed formatting, other small changes

Q-Chem is a commercial quantum chemistry package available for use in our group.

==Fragment Guesses with multiple bases==
To use basis set projection and fragment guesses you have to run the smaller basis set calculation with fragments first, and then read this into a second calculation which does the projection
<pre>$molecule
0 1
--
0 1
H 0.000000 -0.115747 1.133769
H 0.000000 1.109931 -0.113383
O 0.000000 0.005817 -0.020386
--
0 1
He 10000 0 0
$end

$rem
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVDZ
scf_guess fragmo
CORRELATION none
SYMMETRY false
SYM_IGNORE true
$end

@@@

$molecule
read
$end

$rem
scf_guess read
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVTZ
BASIS2 cc-pVDZ
CORRELATION none
MAX_SCF_CYCLES 600
SYMMETRY false
SYM_IGNORE true
basisprojtype ovprojection
$end</pre>

==INTDUMP / FCIDUMP ==
To create files in the FCIDUMP format (used e.g. by Hande) create a QChem input file:
<pre>
$molecule
0 1
H 0 0 0
H 1 0 0
$end

$rem
correlation idump
exchange hf
basis sto-3G
use_abelian_subgroup true
scf_guess_print 9999
$end
</pre>

and use the version of QChem currently (2020-02-20) available from:
<pre>
[you@liminal] $ export QC=public_o
[you@liminal] $ source ~ajwt3/code/qchem/qcsetup.bash
[you@liminal] $ qchem your-input-file.in
</pre>

This should produce a file called <code>INTDUMP</code>
<pre>
&FCI NORB= 2,NELEC= 2,MS2= 0
ORBSYM=1,6,
ISYM=1 UHF=.FALSE.
&END
0.62640249948715 1 1 1 1
0.19679058349422 1 2 1 2
0.6217067630807 2 2 1 1
0.19679058349422 2 1 1 2
0.19679058349422 2 1 2 1
0.65307074688898 2 2 2 2
-1.1108441795661 1 1 0 0
-0.58912100326925 2 2 0 0
-0.39980670964291 1 0 0 0
0.37823811728909 2 0 0 0
0.52917721026434 0 0 0 0
</pre>

==Troubleshooting SCF Metadynamics Calculations==

A full example for SCF Metadynamics can be found at [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic QCMagic SymmetryTools Tutorial (Section 7.1)]. This section provides more information on how to choose parameters in different situations.

If a calculation repeatedly converges on the same solution, adjusting one or more of these keywords can help to find different solutions.
* Increasing <code>SCF_MINFIND_INITNORM</code>, <code>SCF_MINFIND_INITLAMBDA</code> and <code>SCF_MINFIND_INCREASEFACTOR</code> can help get out of the minimum. However, setting <code>SCF_MINFIND_INITLAMBDA</code> too high can result in SCF solutions with very high energies, and <code>SCF_MINFIND_INITNORM</code> can get in the way of convergence, particularly when reading in a solution (in that case, set it to zero).
* Increasing <code>SCF_MINFIND_RESTARTSTEPS</code> helps if a solution is close to another minima but does not converge before the calculation restarts with new orbitals
* Changing <code>SCF_MINFIND_RANDOMMIXING</code> is generally helpful. +/- 15708 and 31416 are sensible values to try, and 07854 (which corresponds to halfway between swapping the orbitals) can also help.

If the resulting solutions are too high in energy,
* decreasing <code>SCF_MINFIND_INITNORM</code>
* decreasing <code>SCF_MINFIND_INITLAMBDA</code>
* decreasing <code>SCF_MINFIND_MIXENERGY</code> (even to 00005)
* adjusting <code>MOM_START</code>
* and using <code>SCF_MINFIND_MIXMETHOD</code> 1 or 2
can help to find lower-energy excited states.

==NOCI Tutorial==
This tutorial includes two possible methods for doing NOCI. The examples below show how to do NOCI at a single geometry, but these methods can easily be extended to cover NOCI at multiple points along a path.

===Method #1===
In this approach, we find minima using SCF Metadynamics and then do NOCI on them in the same Q-Chem output file. A sample Q-Chem input file for formaldehyde is shown below. In this example, we have Q-Chem generate four initial reference determinants using SCF Metadynamics by selecting <code>NOCI_REGEN 0</code>, <code>NOCI_DETGEN 3</code>, and <code>SCF_SAVEMINIMA 4</code>.

<pre>$molecule
0 1
H -0.0000000 0.9275885 1.1766889
C -0.0000000 0.0000000 0.6019825
H -0.0000000 -0.9275885 1.1766889
O 0.0000000 -0.0000000 -0.6001772
$end

$rem
EXCHANGE HF
CORRELATION NOCI
BASIS STO-3G
UNRESTRICTED true
MAX_SCF_CYCLES 1000
SCF_CONVERGENCE 10
MOM_START 1
SCF_SAVEMINIMA 4
SCF_MINFIND_INITNORM 05000
SCF_MINFIND_INITLAMBDA 00300
SCF_MINFIND_RANDOMMIXING 30000
SCF_MINFIND_MIXMETHOD 1
USE_LIBNOCI true
NOCI_REFGEN 0
NOCI_DETGEN 3
NOCI_NEIGVAL 4
NOCI_PRINT 5
$end</pre>

We then run

<pre>qchem formaldehyde.noci.in formaldehyde.noci.out</pre>

and look for the energies of the SCF solutions with

<pre>grep “Saving Minimum” formaldehyde.noci.out</pre>

which gives

<pre> Saving Minimum 1: -112.3535105301
Saving Minimum 2: -111.9177589859
Saving Minimum 3: -111.7116823375
Saving Minimum 4: -112.3571993219</pre>

Information about the resulting NOCI states is printed at the end of the output file. In this example, the energies and S2 values can be examined with

<pre>grep -A5 "NOCI Energy" formaldehyde.noci.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.36268 -112.16945 -111.91685 -111.71168
NOCI <S^2>:
1 2 3 4
1 0.0835515 1.8508569 0.0321459 1.0273785</pre>

===Method #2===
In this method, we use QCMagic to read in SCF Metadynamics states in <code>.sd</code> files into a Q-Chem NOCI calculation.

In this example, we are starting with HF SCF solutions for formaldehyde in four separate <code>.sd</code> files (<code>form.s1.sd</code>, <code>form.s2.sd</code>, <code>form.s3.sd</code> and <code>form.s4.sd</code>) with the following energies:
<pre>-112.3535105299 Minimum 1
-111.9177348081 Minimum 2
-111.4909067597 Minimum 3
-111.8200842261 Minimum 4</pre>

First, we combine all <code>.sd</code> files for the states at a particular geometry using <code>runcombineSDXC.py</code>, e.g.:

<pre>runcombineSDXC.py -i form.s1.sd\
form.s2.sd\
form.s3.sd\
form.s4.sd\
-o form_1234 > terminal_sdxc</pre>

This creates an <code>.sd</code> file, <code>form_1234.sd</code>, which contains all the minima at that geometry. We then use <code>runrunSDXC.py</code> to apply a template for NOCI and call Q-Chem:

<pre>runrunSDXC.py -p 12 -L --template=noci.template form_1234.sd noci_form_1234 > terminal_noci</pre>

An example <code>noci.template</code> file for LIBNOCI is shown below:

<pre>$molecule
read
$end

$rem
EXCHANGE HF
CORRELATION NOCI
UNRESTRICTED TRUE
BASIS STO-3G
SCF_SAVEMINIMA 4
SCF_READMINIMA -4
NOCI_PRINT 10
USE_LIBNOCI TRUE
NOCI_REFGEN 1
NOCI_DETGEN 0
SKIP_SCFMAN TRUE
$end

$noci
1 2 3 4
$end</pre>

In this template, we specify <code>NOCI_REFGEN 1</code> and <code>NOCI_DETGEN 0</code> since we are reading in our reference determinants and do not wish to generate any more. We use the <code>$noci</code> section to specify which states to read in: here, we read in all the states, though we could read in a subset of the states in the <code>.sd</code> file if we wanted to. In this example, we use <code>SKIP_SCFMAN TRUE</code> to avoid reconverging the solutions. If we wanted to reconverge them, we would omit this option and include any missing SCF options in the template.

We can now examine our output file, <code>noci_form_1234.g0.xc0.out</code>. First, we check the energies of the HF states we read in with

<pre>grep "Total Energy =" noci_form_1234.g0.xc0.out</pre>

which gives

<pre> Total Energy = -112.3535105266
Total Energy = -111.9177347312
Total Energy = -111.4909067030
Total Energy = -111.8200841138</pre>

and examine the energies and S2 values of the NOCI solutions with

<pre>grep -A5 "NOCI Energy" noci_form_1234.g0.xc0.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.35436 -111.91688 -111.82008 -111.49091
NOCI <S^2>:
1 2 3 4
1 0.0000000 0.0000000 1.0107832 1.9977364</pre>

(N.B. If we compare the NOCI states from method #1 and method #2, we notice that they are different because weused different HF states in both methods. If we had used the same states in both examples, we would get the same NOCI results.)

===Further Information===
Finding SCF solutions with Metadynamics:
* [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic QCMagic SymmetryTools Tutorial (Section 7.1)]
* Q-Chem section on Troubleshooting Metadynamics Calculations: see above
* [https://manual.q-chem.com/5.3/sec_SCF-meta-dyn.html Q-Chem documentation on SCF Metadynamics (Section 4.9.2)]
* [https://manual.q-chem.com/5.3/subsubsec_SCFMetadynNOCI.html Q-Chem documentation on the LIBNOCI implementation of SCF Metadynamics (Section 4.9.3)]
NOCI
* [https://manual.q-chem.com/5.3/subsec_NOCI.html Q-Chem documentation for NOCI (Section 7.4)]
* Burton H. G. A.; Thom, A. J. W. General Approach for Multireference Ground and Excited States Using Nonorthogonal Configuration Interaction. J Chem. Theory Comput. 2019, 15, 4851.
Keywords for different implementations of NOCI in Q-Chem
* LIBNOCI (most recent implementation): <code>USE_LIBNOCI TRUE</code>
* older implementation: <code>USE_LIBNOCI FALSE</code> and <code>GEN_SCFMAN OFF</code>

QChem

2021-07-24T21:12:59Z

Cak49: Added NOCI Tutorial

Q-Chem is a commercial quantum chemistry package available for use in our group.

= Fragment Guesses with multiple bases =
To uses basis set projection and fragment guesses you have to run the smaller basis set calculation with fragments first, and then read this into a second calculation which does the projection
<code>
$molecule
0 1
--
0 1
H 0.000000 -0.115747 1.133769
H 0.000000 1.109931 -0.113383
O 0.000000 0.005817 -0.020386
--
0 1
He 10000 0 0
$end
$rem
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVDZ
scf_guess fragmo
CORRELATION none
SYMMETRY false
SYM_IGNORE true
$end
@@@
$molecule
read
$end
$rem
scf_guess read
JOBTYPE sp
EXCHANGE b3lyp
BASIS cc-pVTZ
BASIS2 cc-pVDZ
CORRELATION none
MAX_SCF_CYCLES 600
SYMMETRY false
SYM_IGNORE true
basisprojtype ovprojection
$end
</code>

==INTDUMP / FCIDUMP ==

To create files in the FCIDUMP format (used e.g. by Hande) create a QChem input file:

<pre>
$molecule
0 1
H 0 0 0
H 1 0 0
$end

$rem
correlation idump
exchange hf
basis sto-3G
use_abelian_subgroup true
scf_guess_print 9999
$end
</pre>

and use the version of QChem currently (2020-02-20) available from:
<pre>
[you@liminal] $ export QC=public_o
[you@liminal] $ source ~ajwt3/code/qchem/qcsetup.bash
[you@liminal] $ qchem your-input-file.in
</pre>

This should produce a file called <code>INTDUMP</code>
<pre>
&FCI NORB= 2,NELEC= 2,MS2= 0
ORBSYM=1,6,
ISYM=1 UHF=.FALSE.
&END
0.62640249948715 1 1 1 1
0.19679058349422 1 2 1 2
0.6217067630807 2 2 1 1
0.19679058349422 2 1 1 2
0.19679058349422 2 1 2 1
0.65307074688898 2 2 2 2
-1.1108441795661 1 1 0 0
-0.58912100326925 2 2 0 0
-0.39980670964291 1 0 0 0
0.37823811728909 2 0 0 0
0.52917721026434 0 0 0 0
</pre>

==Troubleshooting SCF Metadynamics Calculations==

If a calculation repeatedly converges on the same solution, adjusting one or more of these keywords can help to find different solutions.

* Increasing <code>SCF_MINFIND_INITNORM</code>, <code>SCF_MINFIND_INITLAMBDA</code> and <code>SCF_MINFIND_INCREASEFACTOR</code> can help get out of the minimum. However, setting <code>SCF_MINFIND_INITLAMBDA</code> too high can result in SCF solutions with very high energies, and <code>SCF_MINFIND_INITNORM</code> can get in the way of convergence, particularly when reading in a solution (in that case, set it to zero).
* Increasing <code>SCF_MINFIND_RESTARTSTEPS</code> helps if a solution is close to another minima but does not converge before the calculation restarts with new orbitals
* Changing <code>SCF_MINFIND_RANDOMMIXING</code> is generally helpful. +/- 15708 and 31416 are sensible values to try, and 07854 (which corresponds to halfway between swapping the orbitals) can also help.

If the resulting solutions are too high in energy,
* decreasing <code>SCF_MINFIND_INITNORM</code>
* decreasing <code>SCF_MINFIND_INITLAMBDA</code>
* decreasing <code>SCF_MINFIND_MIXENERGY</code> (even to 00005)
* adjusting <code>MOM_START</code>
* and using <code>SCF_MINFIND_MIXMETHOD</code> 1 or 2
can help to find lower-energy excited states.

==NOCI Tutorial==
This tutorial includes two possible methods for doing NOCI. The examples below show how to do NOCI at a single geometry, but these methods can easily be extended to cover NOCI at multiple points along a path.

===Method #1===
In this approach, we find minima using SCF Metadynamics and then do NOCI on them in the same Q-Chem output file. A sample Q-Chem input file for formaldehyde is shown below. In this example, we have Q-Chem generate four initial reference determinants using SCF Metadynamics by selecting <code>NOCI_REGEN 0</code>, <code>NOCI_DETGEN 3</code>, and <code>SCF_SAVEMINIMA 4</code>.

<pre>$molecule
0 1
H -0.0000000 0.9275885 1.1766889
C -0.0000000 0.0000000 0.6019825
H -0.0000000 -0.9275885 1.1766889
O 0.0000000 -0.0000000 -0.6001772
$end

$rem
EXCHANGE HF
CORRELATION NOCI
BASIS STO-3G
UNRESTRICTED true
MAX_SCF_CYCLES 1000
SCF_CONVERGENCE 10
MOM_START 1
SCF_SAVEMINIMA 4
SCF_MINFIND_INITNORM 05000
SCF_MINFIND_INITLAMBDA 00300
SCF_MINFIND_RANDOMMIXING 30000
SCF_MINFIND_MIXMETHOD 1
USE_LIBNOCI true
NOCI_REFGEN 0
NOCI_DETGEN 3
NOCI_NEIGVAL 4
NOCI_PRINT 5
$end</pre>

We then run

<pre>qchem formaldehyde.noci.in formaldehyde.noci.out</pre>

and look for the energies of the SCF solutions with

<pre>grep “Saving Minimum” formaldehyde.noci.out</pre>

which gives

<pre> Saving Minimum 1: -112.3535105301
Saving Minimum 2: -111.9177589859
Saving Minimum 3: -111.7116823375
Saving Minimum 4: -112.3571993219</pre>

Information about the resulting NOCI states is printed at the end of the output file. In this example, the energies and S2 values can be examined with

<pre>grep -A5 "NOCI Energy" formaldehyde.noci.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.36268 -112.16945 -111.91685 -111.71168
NOCI <S^2>:
1 2 3 4
1 0.0835515 1.8508569 0.0321459 1.0273785</pre>

===Method #2===
In this method, we use QCMagic to read in SCF Metadynamics states in <code>.sd</code> files into a Q-Chem NOCI calculation.

In this example, we are starting with HF SCF solutions for formaldehyde in four separate <code>.sd</code> files (<code>form.s1.sd</code>, <code>form.s2.sd</code>, <code>form.s3.sd</code> and <code>form.s4.sd</code>) with the following energies:
<pre>-112.3535105299 Minimum 1
-111.9177348081 Minimum 2
-111.4909067597 Minimum 3
-111.8200842261 Minimum 4</pre>

First, we combine all <code>.sd</code> files for the states at a particular geometry using <code>runcombineSDXC.py</code>, e.g.:

<pre>runcombineSDXC.py -i form.s1.sd\
form.s2.sd\
form.s3.sd\
form.s4.sd\
-o form_1234 > terminal_sdxc</pre>

This creates an <code>.sd</code> file, <code>form_1234.sd</code>, which contains all the minima at that geometry. We then use <code>runrunSDXC.py</code> to apply a template for NOCI and call Q-Chem:

<pre>runrunSDXC.py -p 12 -L --template=noci.template form_1234.sd noci_form_1234 > terminal_noci</pre>

An example <code>noci.template</code> file for LIBNOCI is shown below:

<pre>$molecule
read
$end

$rem
EXCHANGE HF
CORRELATION NOCI
UNRESTRICTED TRUE
BASIS STO-3G
SCF_SAVEMINIMA 4
SCF_READMINIMA -4
NOCI_PRINT 10
USE_LIBNOCI TRUE
NOCI_REFGEN 1
NOCI_DETGEN 0
SKIP_SCFMAN TRUE
$end

$noci
1 2 3 4
$end</pre>

In this template, we specify <code>NOCI_REFGEN 1</code> and <code>NOCI_DETGEN 0</code> since we are reading in our reference determinants and do not wish to generate any more. We use the <code>$noci</code> section to specify which states to read in: here, we read in all the states, though we could read in a subset of the states in the <code>.sd</code> file if we wanted to. In this example, we use <code>SKIP_SCFMAN TRUE</code> to avoid reconverging the solutions. If we wanted to reconverge them, we would omit this option and include any missing SCF options in the template.

We can now examine our output file, <code>noci_form_1234.g0.xc0.out</code>. First, we check the energies of the HF states we read in with

<pre>grep "Total Energy =" noci_form_1234.g0.xc0.out</pre>

which gives

<pre> Total Energy = -112.3535105266
Total Energy = -111.9177347312
Total Energy = -111.4909067030
Total Energy = -111.8200841138</pre>

and examine the energies and S2 values of the NOCI solutions with

<pre>grep -A5 "NOCI Energy" noci_form_1234.g0.xc0.out</pre>

giving

<pre> NOCI Energy:
1 2 3 4
1 -112.35436 -111.91688 -111.82008 -111.49091
NOCI <S^2>:
1 2 3 4
1 0.0000000 0.0000000 1.0107832 1.9977364</pre>

(N.B. If we compare the NOCI states from method #1 and method #2, we notice that they are different because weused different HF states in both methods. If we had used the same states in both examples, we would get the same NOCI results.)

===Further Information===
Finding SCF solutions with Metadynamics:
* [https://wikis.ch.cam.ac.uk/thom/wiki/index.php/QCMagic QCMagic SymmetryTools Tutorial (Section 7.1)]
* Q-Chem section on Troubleshooting Metadynamics Calculations: see above
* [https://manual.q-chem.com/5.3/sec_SCF-meta-dyn.html Q-Chem documentation on SCF Metadynamics (Section 4.9.2)]
* [https://manual.q-chem.com/5.3/subsubsec_SCFMetadynNOCI.html Q-Chem documentation on the LIBNOCI implementation of SCF Metadynamics (Section 4.9.3)]
NOCI
* [https://manual.q-chem.com/5.3/subsec_NOCI.html Q-Chem documentation for NOCI (Section 7.4)]
* Burton H. G. A.; Thom, A. J. W. General Approach for Multireference Ground and Excited States Using Nonorthogonal Configuration Interaction. J Chem. Theory Comput. 2019, 15, 4851.
Keywords for different implementations of NOCI in Q-Chem
* LIBNOCI (most recent implementation): <code>USE_LIBNOCI TRUE</code>
* older implementation: <code>USE_LIBNOCI FALSE</code> and <code>GEN_SCFMAN OFF</code>

QChem

2021-07-12T21:16:00Z

Cak49: Added section on troubleshooting SCF metadynamics

Computer and Storage List

2020-10-08T16:45:07Z

Cak49:

{| class="wikitable"
!name
!office
!user
!cores
!processor
!RAM
!OS
!Video Ports
!Displays
!Hardware requests
!Working from home
|-
|carpathia
|379
|Alex
|6
|Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
|32GB
|Ubuntu 18.04.5
|-
|liminal
|379
|Alex
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 18.04.5
|
|
|-
|hypatia
|
|
|6
|Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
|32GB
|Ubuntu 18.04.5
|-
|serenity
|378
|Andreea
|6
|Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz
|64GB
|Ubuntu 16.04
|
|
|-
|sandstone
|378
|Kripa, Tiger
|6
|Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
|32GB
|Ubuntu 16.04.7
|
|
|-
|gritstone
|
|Hang
|6
|Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
|32GB
|Ubuntu 16.04.4
|-
|moonraker
|
|Nick Lee
|4
|Intel(R) Xeon(R) CPU E3-1270 v5 @ 3.60GHz
|64GB
|Ubuntu 16.04.7
|-
|obsidian
|
|Isha
|6
|Intel(R) Core(TM) i7-6800K CPU @ 3.40GHz
|32GB
|Ubuntu 18.04
|-
|hylas
|378
|Fabio
|6
|Intel(R) Core(TM) i7-6800K CPU @ 3.40GHz
|32GB
|Ubuntu 16.04.7
|
|-
|cerberus
|
|Alex
|6
|Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz
|32GB
|CentOS 7 [FPGA development board host]
|-
|chucksty
|
|David
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 16.04.3
|-
|chesterian
|
|Bang
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 16.04.3
|-
|behemoth
|380
|Arta
|8
|Intel(R) Xeon(R) Silver 4208 CPU @ 2.10GHz
|256GB
|Ubuntu 18.04.5
|-
|nemesis
|378
|Constance
|6
|Intel(R) Core(TM) i7-4930X CPU @ 3.40GHz
|16GB
|Ubuntu 18.04.5
|-
|cerebro
|
|Alavi & Thom Groups
|12 x 20
16 x 8
|2x Intel(R) Xeon(R) CPU X5650 @ 2.67GHz
2x Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz
|48GB
64GB
|Rocks 6.2 (CentOS 6.9)
|-
|}

=Notes=

To find out your OS version, run

lsb_release -a

To determine the RAM, run

head -1 /proc/meminfo

To find out core counts, run

cat /proc/cpuinfo

NB the number of 'processors' may be different from the number of cores owing to hyperthreading. The 'cpu cores' value is the one to take for single CPU machines.

[http://hobbit.ch.cam.ac.uk/xymon/workstations/workstationsThom/workstationsThomLinux/ Hobbit] may also have some useful information.

[https://www.ch.cam.ac.uk/computing/group-computer-representatives Group computer reps] can manage group entries in the department [https://chemdb.ch.cam.ac.uk/hotwire3/chemistry/ database] and there's a [https://apps.ch.cam.ac.uk/computer-reps/group-computers.php hardware inventory] and a [https://apps.ch.cam.ac.uk/space-management/space-report.php space report] too.

Computer and Storage List

2020-08-31T14:16:26Z

Cak49:

{| class="wikitable"
!name
!office
!user
!cores
!processor
!RAM
!OS
!Video Ports
!Displays
!Hardware requests
|-
|carpathia
|379
|Alex
|6
|Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
|32GB
|Ubuntu 18.04.5
|-
|liminal
|379
|Alex
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 18.04.5
|-
|hypatia
|
|
|6
|Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz
|32GB
|Ubuntu 16.04
|-
|serenity
|
|Andreea
|6
|Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz
|64GB
|Ubuntu 16.04
|-
|sandstone
|
|Kripa
|6
|Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
|32GB
|Ubuntu 16.04.7
|-
|gritstone
|
|
|6
|Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
|32GB
|Ubuntu 16.04.4
|-
|moonraker
|
|Nick Lee
|4
|Intel(R) Xeon(R) CPU E3-1270 v5 @ 3.60GHz
|64GB
|Ubuntu 16.04.7
|-
|obsidian
|
|Hugh
|6
|Intel(R) Core(TM) i7-6800K CPU @ 3.40GHz
|32GB
|Ubuntu 16.04
|-
|hylas
|
|Fabio
|6
|Intel(R) Core(TM) i7-6800K CPU @ 3.40GHz
|32GB
|Ubuntu 16.04
|-
|cerberus
|
|Alex
|6
|Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz
|32GB
|CentOS 7 [FPGA development board host]
|-
|chucksty
|
|David
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 16.04.3
|-
|chesterian
|
|Bang
|6
|Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz
|64GB
|Ubuntu 16.04.3
|-
|nemesis
|
|Constance
|6
|Intel(R) Core(TM) i7-4930X CPU @ 3.40GHz
|16GB
|Ubuntu 16.04.7
|-
|cerebro
|
|Alavi & Thom Groups
|12 x 20
16 x 8
|2x Intel(R) Xeon(R) CPU X5650 @ 2.67GHz
2x Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz
|48GB
64GB
|Rocks 6.2 (CentOS 6.9)
|}

=Notes=

To find out your OS version, run

lsb_release -a

To determine the RAM, run

head -1 /proc/meminfo

To find out core counts, run

cat /proc/cpuinfo

NB the number of 'processors' may be different from the number of cores owing to hyperthreading. The 'cpu cores' value is the one to take for single CPU machines.

[http://hobbit.ch.cam.ac.uk/xymon/workstations/workstationsThom/workstationsThomLinux/ Hobbit] may also have some useful information.

[https://www.ch.cam.ac.uk/computing/group-computer-representatives Group computer reps] can manage group entries in the department [https://chemdb.ch.cam.ac.uk/hotwire3/chemistry/ database] and there's a [https://apps.ch.cam.ac.uk/computer-reps/group-computers.php hardware inventory] and a [https://apps.ch.cam.ac.uk/space-management/space-report.php space report] too.