Under Construction

On this page you will find useful tools that are used by various people in the group.

top

FTNCHECK - Fortran Checker

More information: http://www.dsm.fordham.edu/~ftnchek/
More information for GAMESS application can probably be found in checkgms.info file that is included with GAMESS

1) Get ftnchek from:
http://www.dsm.fordham.edu/~ftnchek/
2) In terminal:
where ftnchek
3) In terminal:
untar ftnchek in Applications folder
4) In terminal:
vi README
vi INSTALL
---They also have some Makfile s.
5) In terminal:
ftnchek --help
6) Use ftnchek via checkgms

top

NWChem Installation

Notes about linking against Intel's MKL
- Intel has created the MKL Link Line Advisor to help you determine how to link your applications against the Math Kernel Library.
- This web-based tool can be found at http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor .
- While you can easily identify your version of ifort (or icc) with the command "ifort -V", it's not so easy to identify the MKL version.
Once you've identified your compiler version, you can identify the corresponding version of MKL using the table available at http://software.intel.com/en-us/articles/which-version-of-ipp--mkl--tbb-is-installed-with-intel-compiler-professional-edition/ .
- For NWChem on TheBunny, I enter the following in the Link Line Advisor:

Intel Product	Intel MKL 10.2
OS	Linux
processor architecture	Intel 64
compiler	Intel Fortran
dynamic or static linking	Static
interface layer	ILP64 (64-bit integer)
sequential or multi-threaded layer	Sequential **Multi-threaded might be useful if you don't fill nodes with MPI processes
all remaining options	Don't select anything

Instructions for Installing NWChem on TheBunny

1. Get the tar-ball

http://www.nwchem-sw.org -> Download Code

2. Move tar-ball to TheBunny
a. sftp or scp to move file to Gateway
b. Move file from Gateway to TheBunny

3. Extract all files from tar-ball
tar -xzf Nwchem-6.1-2012-Feb-10.tar.gz

4. Set environment variables for build settings
** good idea to put them in a shell script for future reference
for example: /home/nick/public/Build_Env_NWChem_6.1_02102012.zsh

5. Configure NWChem build for your module selection
a. cd ${NWCHEM_TOP}/src
b. make nwchem_config >& nwchem_config_log &

6. Build NWChem
a. make FC=ifort CC=icc >& makelog.MMDDYYYY.X &
b. Wait about 75 minutes for build to finish
c. Check that no errors are reported at bottom of log. (Warnings are okay)
d. ls -lh ${NWCHEM_TOP}/bin/LINUX64/nwchem
Check that timestamp corresponds to end of step 6a

7. Create an installation directory
a. mkdir ~/nwchem-6.1-02102012
** Name and location of your installation are up to you.
b. cd ~/nwchem-6.1-02102012
c. mkdir bin data
d. cp ${NWCHEM_TOP}/bin/LINUX64/nwchem ~/nwchem-6.1-02102012/bin
e. cp -r ${NWCHEM_TOP}/src/basis/libraries ~/nwchem-6.1-02102012/data
f. cp -r ${NWCHEM_TOP}/src/data ~/nwchem-6.1-02102012
g. cp -r ${NWCHEM_TOP}/src/nwpw/libraryps ~/nwchem-6.1-02102012/data

8. Create .nwchemrc file to tell NWChem about installation paths
a. mkdir ~/nwchem-6.1-02102012/etc
b. Create file ~/nwchem-6.1-02102012/etc/default.nwchemrc containing next 10 lines
nwchem_basis_library /home/nick/nwchem-6.1-02102012/data/libraries/
nwchem_nwpw_library /home/nick/nwchem-6.1-02102012/data/libraryps/
ffield amber
amber_1 /home/nick/nwchem-6.1-02102012/data/amber_s/
amber_2 /home/nick/nwchem-6.1-02102012/data/amber_q/
amber_3 /home/nick/nwchem-6.1-02102012/data/amber_x/
amber_4 /home/nick/nwchem-6.1-02102012/data/amber_u/
spce/home/nick/nwchem-6.1-02102012/data/solvents/spce.rst
charmm_s /home/nick/nwchem-6.1-02102012/data/charmm_s/
charmm_x /home/nick/nwchem-6.1-02102012/data/charmm_x/

9. Create a symbolic link to default.nwchemrc in your home directory
a. mv ~/.nwchemrc ~/.old_nwchemrc
** save old .nwchemrc in case you want to go back to using the site-installed NWChem
b. ln -s ~/nwchem-6.1-02102012/etc/default.nwchemrc ~/.nwchemrc

10. Modify existing submission scripts to use your NWChem installation
a. mkdir ~/nwchem-6.1-02102012/scripts
b. cp /home/thebunny/NWChem-5.1.1-r19297/scripts/subnw ~/nwchem-6.1-02102012/scripts/subnw
** If you want to adapt the site version of nwque
c. cp /home/thebunny/NWChem-5.1.1-r19297/scripts/nwque ~/nwchem-6.1-02102012/scripts/nwque
** If you want to try my revised script (no warranty)
c. cp /home/nick/public/nwque-02102012 ~/nwchem-6.1-02102012/scripts/nwque
d. Near top of subnw script, change line with set NWCHEM_DIR = ... to the following:
set NWCHEM_DIR = "/home/nick/nwchem-6.1-02102012"
e. create an alias for your modified subnw script for example:
alias nw61='/home/nick/nwchem-6.1-02102012/scripts/./subnw'

11. Test your build
** probably easiest to run a few short jobs with the functionality you're interested in and compare with results from site version.
If you want to run NWChem's QA battery (a pain):
a. copy /home/nick/public/nwque_QA_02102012 to a directory of your choice
b. modify a subnw script to call nwque_QA... (instead of nwque)
c. in nwque_QA... , modify all paths that include /home/nick to those that correspond to your build/installation
d. create an empty file with any name ending in .nw (we'll call it "input.nw")
e. run the subnw script from b with arguments "-N 1 -np 1 input.nw"
f. tests will run for many hours and produce a file named "log"
Check log for failed tests and then check the outputs of tests that failed
Many failures will be small numerical differences or differences in the number of iterations
==== Procedure for identifying problem with failed test ====
** assume name of failed test is "test"
i. cd testoutputs
ii. diff --side-by-side --suppress-common-lines test.ok.out.nwparse test.out.nwparse
Small numerical differences and differing numbers of iterations should be apparent
iii. for closer investigation, look at the output files
test.ok.out = reference output file
test.out = output file produced with your NWChem binary
g. parallel execution of tests
i. modify nwque_QA... to replace two instances of "doqmtests.mpi 1" with "doqmtests.mpi 8"
** Ideally, all tests would run across 2 nodes without problems, but this doesn't seem to be the case
ii. modify ${NWCHEM_TOP}/QA/doqmtests.mpi to vary the number of processes being used for different tests
add "set np = X" before the test(s) you want to use X processes

top

Avogadro

To Download: http://avogadro.openmolecules.net/wiki/Main_Pag
More information: http://avogadro.openmolecules.net/wiki/Tutorials
http://showmedo.com/videotutorials/video?name=3770000&fromSeriesID=377
Optional useful tool, Povray: http://www.povray.org/

----------
Upon opening Avogadro:

left side:
Display Types
Tool Settings

right side:
View tab
Messages tab (click tab to toggle)

----------
Draw and Manipulate atoms/molecules:
*hover over tool to display explanation

Draw tool (F8): select atom
click or click and drag to add connected atom
Navigation tool (F9)
Bond Centric Manipulation tool (have never used)
Select tool (F11): Choose mode (molecule, atom/bond or residue)
Manipulation tool (F10): left mouse button drag ->translate selection
right mouse button drag ->rotate selection
Auto Optimization tool: Choose appropriate options
(for explanation of forcefields, http://avogadro.openmolecules.net/wiki/Tutorials:Force_fields)

----------
Generate GAMESS input file:
Extensions>GAMESS>Input Generator...

Go through dialogue and select various options.

EFPs:
Select molecule
Extensions>GAMESS>EFP Selection...
Enter name (name that will follow in input file after "FRAGNAME=")
QM molecule:
all atoms default unless using EFPs
Select molecule
Extensions>GAMESS>QM Selection...
Name molecule (does not affect input file generated)

----------
Generate NWChem input file:
Extensions>NWChem...
Go through dialogue and select various options.

----------
To view multiple frames:
Extensions>Animation...>Load File...

top

Gnuplot

General information regarding Gnuplot: http://www.gnuplot.info/
Downloaded Gnuplot for Windows from: http://sourceforge.net/projects/gnuplot/files/

----

Using Gnuplot:

the 'plot' command: which file to get data from
the 'using' command: tells gnuplot which column to use for x and y axis

> plot "filename.txt" using 1:2 **for this example use test1_freq.txt

the 'help' command can be used to get more information on the various commands and options.

----

**The following parameters are some that I find useful to beautify your plot

the 'set' directive: sets a parameter (a list of various parameters are below)
> set xlabel "frequency"
> set ylabel "intensity"

the 'unset' command allows you to remove a set parameter
> unset xlabel

the 'set key' is on by default and the the legend will display whatever you
typed after the plot command. To remove, use:
> unset key

the 'with' directive: sets how the data is plotted
the 'using' directive: which columns the data is in
> plot "./test1_frequencies.txt" using 1:2 with impulses

the 'set size ratio [number]': set the size aspect ratio. The number
refers to the y-axis : x-axis ratio

> set size ratio 2.0
> set size noratio * plot fills the terminal

set ranges using:
> set xrange [500:3000]
or for auto scale:
> set xrange [*:*]

----

**The following commands are recommended to be performed always!
**If you ever decide to modify your graph you can do it easily by loading
** your .gp file.

the "save" command: saves all parameters that were set including
the final plot command
> save "filename.gp"

the "load" command: opens the saved .gp file and plots the graph.
> load "filename.gp"

----

the terminal is where your plot will be displayed/printed.

!!! Before you change the terminal, save your file using 'save' command.
!!! This will allow you to easily change any parameters or typos in your plot.

if you type the following command, you will get a list of allowable terminals.
> set terminal
Choose one and retype set terminal command
> set terminal jpeg

the 'set output' command: creates a file with the previously designated format.
> set output "filename"

then use the plot command again
> plot "file"

alternatively you can copy the .gp file and change the terminal and output file names.
I recommend you do this so you can easily load and edit the original .gp file.
----