Installation of DALTON 2016 & 2017

Installation Instruction

1. Go to https://gitlab.com/dalton/dalton/tree/release/2016. There you can find the source code of DALTON, which is free of charge.

2. On your Linux machine, open new terminal (suppose that you are installing with your normal account.)

3. Clone the repository of DALTON from Gitlab using command

git clone --recursive git@gitlab.com:dalton/dalton.git

Wait for 3 - 10 minutes (download speed depends on your internet speed).

Otherwise, download the source code of program from Gitlab directly and upload to your Linux.

4. Move to "dalton" directory.

cd dalton

5. Setup configuration for DALTON compilation (serial) using command

./setup 

If you want to compile DALTON with MPI or OpenMP, use following flags

./setup --mpi --omp 

If you want to use ScaLAPACK, use "--scalapack" flag.

./setup --mpi --omp --scalapack

if your machine has GPU, you can compile DALTON with GP-GPU by using "--gpu" flag.

./setup --mpi --omp --gpu

Otherwise, use "--help" flag to open setup manual.

./setup --help

6. Wait for 5 - 10 seconds. Now you are ready to install the DALTON.

7. Go to build directory and compile using parallel make, for example, use 4 CPU cores.

cd ./build

make -j 4

wait for 3 - 5 minutes until installation is completed. You should also see message like this

[100%] Built target dalton

Scanning dependencies of target dalton.x

[100%] Building Fortran object CMakeFiles/dalton.x.dir/DALTON/main/dalton.F.o

Linking Fortran executable dalton.x

[100%] Built target dalton.x

DALTON executable file is at /folder/of/dalton/build/dalton.x.

Running DALTON

1. Create DALTON input file and molecule file. For example, foo.dal and foo.mol.

1.1 Copy the following input deck to foo.dal file.

**DALTON INPUT

.RUN WAVEFUNCTION

.DIRECT

.OPTIMIZE

**WAVE FUNCTIONS

.DFT

 BLYP

*DFT INPUT

.DFTD3

.GRID TYPE

BECKE

*SCF INPUT

.THRESH

 1.0D-06

**END OF INPUT

1.2 Copy following basis set and molecule data to foo.mol file.

BASIS

STO-3G

CH4 Dimer

----------

Atomtypes=2 Angstrom

Atoms=2  Charge=6.0

C   0.000000  -0.000140   1.859161

C   0.000000   0.000140  -1.859161

Atoms=8  Charge=1.0

H  -0.888551   0.513060   1.494685

H   0.888551   0.513060   1.494685

H   0.000000  -1.026339   1.494868

H   0.000000   0.000089   2.948284

H   0.000000  -0.000089  -2.948284

H  -0.888551  -0.513060  -1.494685

H   0.888551  -0.513060  -1.494685

H   0.000000   1.026339  -1.494868

2. Run calculation using either command

nohup dalton foo &

or command

nohup dalton foo.dal foo.mol &

Wait until you get following message

[1]+  Done                    dalton foo

DALTON output will be printed to foo.out and foo.tar.gz.

Caveats

1. Running DALTON using the 1st command both foo.dal and foo.mol must be the same location. DALTON can detect these files automatically.

2. Another useful command is nohup. This command can force DALTON to run in background.

Optional

1. Test DALTON in parallel using command

ctest -j 4

This test can take a long time.

2. Set path of directory where DALTON will store the scratch files. For connected cluster, it should be a local temporary directory of compute node.

export DALTON_TMPDIR=/full/path/of/scratch/folder

3. Run DALTON using its executable

/folder/of/dalton/build/dalton

Each user should also add DALTON directory to PATH environment variable, e.g.

export PATH=/folder/of/dalton/build/:$PATH

4. If you want to build DALTON with Intel/MKL, please consult the manual.

http://dalton-installation.readthedocs.io/en/latest/doc/math.html

Rangsiman Ketkaew