Computational Chemistry Pre-Compiled Binaries for Linux

Pre-compiled binaries of computational chemistry programs can save the time and trouble of obtaining and compiling the program source code.

When using a pre-compiled binary, be aware that one must use a binary that is appropriate for one's particular hardware (e.g., x86 or ia32, Intel x86_64 or EM64T or Intel64, AMD x86_64 or Opteron, SGI MIPS R10000, etc.), operating system, (e.g., Linux, OS X, Windows, IRIX, AIX, Solaris, etc.), bitsize (32-bit or i4, 64-bit or i8), distribution/vendor (e.g., RHEL, CentOS, Fedora Core, RedHat, SuSE, SGI, Solaris, etc.), and version (4, 6.2, 7.2, 11, etc.).

GAMESS

GAMESS is currently no longer available for Linux as a pre-compiled binary, although it is available for other OS's. You must compile the source code on linux using a system compiler, e.g., f77, gfortran, or another popular compiler.

GAUSSIAN 16

  1. Obtain Gaussian 16
    Gaussian 16 is a commercial software product and must be purchased from Gaussian, Inc. Both source code and pre-compiled binary versions of Gaussian 16 are available.
  2. Setup gaussian group and add authorized users to this group
    $ su
    # cd /etc
    # cp -p group group.bak
    # cp -p gshadow gshadow.bak
    # grep 499 group
    # groupadd -g 499 gaussian
    # usermod -a -G gaussian webmo
    # usermod -a -G gaussian smith
    # usermod -a -G gaussian jones
  3. Determine which capabilities are supported by your cpu
    # cat /proc/cpuinfo | grep avx2
    # cat /proc/cpuinfo | grep avx
    # cat /proc/cpuinfo | grep sse
  4. Copy the appropriate binary distribution from CD-ROM
    # mount /mnt/cdrom
    # cd /usr/local
    # cp -p /mnt/cdom/tar/<distribution>.tb? .
    # umount /mnt/cdrom
  5. Uncompress and extract files
    if the distribution is a *.tbz file:
    # yum install bzip2
    # bunzip2 -k <distribution>.tbz
    # tar xvf <distribution>.tar
    if the distribution is a *.tbJ file:
    # tar xvJf <distribution>.tbJ
  6. Change ownership and permissions
    # chown -R root:gaussian g16
    # chmod -R o-rwx g16
    # exit
  7. Run a test job from the command line
    Note: you may need to log out and back in for new group memberships to be recognized
    $ cd ~
    $ mkdir g16
    $ cd g16
    $ vi g16setup
          export g16root=/usr/local
          export GAUSS_SCRDIR=/tmp
          source $g16root/g16/bsd/g16.profile
    $ cp -p /usr/local/g16/tests/com/test0001.com test0001.com
    $ cp -p /usr/local/g16/tests/amd64/test0001.log test0001.log.amd64
    $ source g16setup
    $ cat test0001.com
    $ g16 < test0001.com > test0001.log.linux
    $ tail test0001.log.amd64
    $ tail test0001.log.linux
  8. Configure WebMO (version 18 or greater) to use Gaussian16 as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for Gaussian
    • Click 'Edit' to configure the Gaussian interface
    • Verify that the first two entries are correct:
      • Gaussian Version: Gaussian 16
      • Gaussian root directory: /usr/local
    • Click the 'Suggest' button to fill the remining entries
    • Click the 'Submit' button for the changes to take effect
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  9. Login as a WebMO user, and run a test job using Gaussian as the computational engine

MOLPRO 2015

  1. Obtain a Molpro license
    MolPro is a commercial software product, and a license must be purchased from www.molpro.net. Registered licensees will receive a Molpro username and password, which grant access to the download area of the Molpro website.
  2. Download the Molpro binary and license token
    At the Molpro website, follow the "Log in" link, and supply your Molpro username and password. Follow the "Account information and ordering" link. Download a copy of the license token. Follow the "Download" link. Download the latest binary version appropriate for your hardware architecture and operating system, for example, molpro-mpp-2015.1.18.linux_x86_64_openmp.sh.gz for 64-bit Linux. (Note that mpp versions will run on a single processor as well as in parallel.)
  3. Install the molpro binary
    $ su -
    # mkdir /tmp/molpro
    # cd /tmp/molpro
    # cp -p {download_dir}/molpro-mpp-2015.1.18.linux_x86_64_openmp.sh.gz .
    # gunzip molpro-mpp-2015.1.18.linux_x86_64_openmp.sh.gz
    # chmod 755 molpro-mpp-2015.1.18.linux_x86_64_openmp.sh
    # mkdir /usr/local/molpro        
    # ./molpro-mpp-2015.1.18.linux_x86_64_openmp.sh
    Enter bin directory to link Molpro (optional) []
    /usr/local/molpro
    Enter installation directory for Molpro files [/usr/local/molpro/molprop_2015_1_linux_x86_64_i8]
    Installation of Molpro files complete
    Please give your username for accessing molpro
    ******
    Please give your password for accessing molpro
    ******
    Token sucessfully downloaded and installed
  4. If Molpro license token was not automatically installed in last step, install license token manually
    # cp {download_dir}/token.txt /usr/local/molpro/molprop_2015_1_linux_x86_64_i8/lib/.token
    # chmod 644 /usr/local/molpro/molprop_2015_1_linux_x86_64_i8/lib/.token
  5. If necessary, fix file ownerships
    # chown -R root:root /usr/local/molpro
  6. Clean up
    # rm -rf /tmp/molpro/
    # exit
  7. Run a test job from the command line
    $ cd ~
    $ mkdir test
    $ cd test
    $ cp -p /usr/local/molpro/molprop_2015_1_linux_x86_64_i8/examples/procedures .
    $ cp -p /usr/local/molpro/molprop_2015_1_linux_x86_64_i8/examples/h2o_optmp2.com .
    $ /usr/local/molpro/molpro h2o_optmp2.com
    $ ls -alF h2o*
  8. Configure WebMO to use Molpro as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable interface' icon for Molpro
    • Click the 'Edit interface' icon to configure the Molpro interface
    • Verify that the entries are correct; if necessary, edit entries and click 'Submit'
      • Molpro version: 2015.1
      • Molpro path: /usr/local/molpro/molpro
      • LD library directory:
      • Cores: Min=1, Max=1, Default=1 (Pro only; more according to architecture)
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  9. Login as a WebMO user, and run a test job using Molpro as the computational engine
  10. Notes:

MOPAC 7

  1. Obtain the MOPAC 7 binary
    MOPAC 7 is public domain and the pre-compiled binary code mopac7.tar.gz can be obtained from the WebMO website
  2. Uncompress the archive
    $ su
    # cd /usr/local
    # tar xzvf /{download_dir}/mopac7.exe.tar.gz
    # chown -R root:root mopac7
    # exit
  3. Configure WebMO to use MOPAC7 as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for MOPAC
    • Click 'Edit' to configure the MOPAC interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • Mopac version: 7
      • Mopac path: /usr/local/mopac7/mopac.exe
      • External parameter dir:
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  4. Login as a WebMO user, and run a test job using MOPAC as the computational engine

MOPAC 2016

  1. Visit the MOPAC homepage at openmopac.net and request an academic license for MOPAC. After a few days, you should receive a license key of the form 12345678a12345678
  2. Download the 64-bit linux version of MOPAC2016
  3. Make sure you have unzip installed
    $ sudo yum install unzip (for CentOS and Amazon Linux)
    $ sudo apt-get install unzip (for Debian, Ubuntu)
  4. Unzip the archive
    $ sudo su -
    # mkdir /usr/local/mopac2016/
    # cd /usr/local/mopac2016/
    # unzip /{download_dir}/MOPAC2016_for_Linux_64_bit.zip
  5. Install the library
    # cp -p libiomp5.so /usr/lib/
    # sudo ldconfig
  6. Change permissions on mopac
    # chmod 755 MOPAC2016.exe
    # sudo chmod 777 .
  7. Run MOPAC2016.exe, passing in the license key you recieved from Mr MOPAC. Follow on-screen prompts to accept the license.
    # export MOPAC_LICENSE=/usr/local/mopac2016
    # ./MOPAC2016.exe license
    # sudo chmod 755 .
    # exit
  8. Run a test job from the command line
    $ cd ~
    $ mkdir -p test/mopac2016/
    $ cd test/mopac2016/
    $ cp "/usr/local/mopac2016/Example data set.mop" .
    $ export MOPAC_LICENSE=/usr/local/mopac2016
    $ /usr/local/mopac2016/MOPAC2016.exe "Example data set.mop"
    $ more "Example data set.out"
  9. Configure WebMO to use MOPAC as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for MOPAC
    • Click 'Edit' to configure the MOPAC interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • Mopac version: 2016
      • Mopac path: /usr/local/mopac2016/MOPAC2016.exe
      • External parameter dir:
    • Click 'Return to Admin', followed by 'Logout' to exit the WebMO administration page
  10. Login as a WebMO user, and run a test job using MOPAC as the computational engine
  11. Documentation can be found at openmopac.net/manual.

NWCHEM 7.0

For Centos

  1. Get and install NWchem binary
    sudo yum -y install epel-release
    sudo yum -y install nwchem
    sudo yum -y install nwchem-openmpi
  2. Organize libraries
    (copy paste these three lines all together into the command prompt and press >enter<)
    echo '/usr/lib64/openmpi/lib /usr/lib64/compat-openmpi16/lib' | sudo tee /etc/ld.so.conf.d/nwchem_openmpi.conf
  3. Refresh library cache
    sudo ldconfig
  4. Make necessary edits to nwchemrc
    cat /usr/share/nwchem/nwchemrc | tr [:upper:] [:lower:] | sudo tee /etc/nwchemrc
  5. Make symbolic link
    sudo ln -s /usr/lib64/openmpi/bin/nwchem_openmpi /usr/bin/nwchem
  6. Configure WebMO to use NWChem as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for NWChem
    • Click 'Edit' to configure the NWChem interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • version: get at url- http://www.nwchem-sw.org/index.php/Main_Page
      • NWChem binary path: /usr/bin/nwchem
      • NWChem config path: /etc/nwchemrc
    • Click 'Return to Admin', followed by 'Logout' to exit the WebMO administration page
  7. Login as a WebMO user, and run a test job using NWChem as the computational engine

For Ubuntu/Debian8:

  1. Install NWchem
    sudo apt-get -y install nwchem
  2. Make nwchemrc file
    sudo touch /etc/nwchemrc
  3. Configure WebMO to use NWChem as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for NWChem
    • Click 'Edit' to configure the NWChem interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • version: get at url- http://www.nwchem-sw.org/index.php/Main_Page
      • NWChem binary path: /usr/bin/nwchem
      • NWChem config path: /etc/nwchemrc
    • Click 'Return to Admin', followed by 'Logout' to exit the WebMO administration page
  4. Login as a WebMO user, and run a test job using NWChem as the computational engine

ORCA 4

  1. Obtain Orca
    Orca is available free of charge for academic use. Obtain by making an account at Orca Forums and downloading most recent version of orca 4 for linux in *.zst format
  2. Install zstd
    (for CentOS7 and earlier/Amazon Linux)
    # sudo yum install zstd
    (for Centos8)
    # curl https://mirror.rackspace.com/elrepo/testing/el8/x86_64/RPMS/zstd-1.3.4-1.el8.elrepo.x86_64.rpm -O
    # sudo yum localinstall zstd-1.3.4-1.el8.elrepo.x86_64.rpm
    (for Ubuntu/Debian)
    # sudo apt-get -y install bzip2 # sudo apt-get -y install zstd
  3. Extract files (replace X_X_X with current version number)
    # cd /usr/local
    # tar -I zstd -xvf {download_dir}/orca_X_X_X_X_linux_x86-64_openmpiXXX.tar.zst
    # chown -R root:root orca_X_X_X_X_linux_x86-64_openmpiXXX/
    # cd orca_X_X_X_X_linux_x86-64_openmpiXXX/ # sudo cp ./liborca_tools_4_1_2.so /lib/ # sudo cp ./liborca_tools_4_1_2_mpi.so /lib/ # sudo /sbin/ldconfig # exit
  4. Run a test job from the command line
    • Make a directory to hold test files
      $ cd ~
      $ mkdir orca
      $ cd orca/
    • Make a new file called h2o.inp containing the following:
      ! B3LYP def2-SVP Opt
      *xyz 0 1
      O 0.000000000 0.000000000 0.000000000
      H 0.000000000 0.759337000 0.596043000
      H 0.000000000 -0.759337000 0.596043000
      *
    • Run Orca using the file as input
      $ /usr/local/orca_X_X_X_X_linux_x86-64_openmpiXXX/orca h2o.inp >h2o.out
      $ more h2o.out
  5. Configure WebMO to use Orca as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for ORCA
    • Click 'Edit' to configure the Orca interface
    • Verify that the first two entries are correct (replace X's with current version):
      • ORCA version: X.X.X.X
      • ORCA directory: /usr/local/orca_X_X_X_X_linux_x86-64_openmpiXXX
    • Click the 'Submit' button for the changes to take effect
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  6. Login as a WebMO user, and run a test job using Orca as the computational engine

PQS 3.3

  1. Download PQS binary
    Visit the PQS homepage and download a copy of PQS from their software page. Choose an appropriate implementation and architecture, or choose the free 32-bit trial version, e.g., pqs_mpi2-3.3-19.i386.tar.gz. The following instructions are based on a TAR file download, not the RPM download.
  2. Untar the binary files
    $ su
    # cd /tmp
    # tar xzf {download_dir}/pqs_mpi2-3.3-19.i386.tar.gz
    # cd pqs_trial-3.3-19.i386
    # ./install.sh
    Please choose the installation type: 1 Multi-user install (requires root privileges)
    Enter the target directory for PQS install [/usr/local/share/PQS]: /usr/local/PQS
    Enter the scratch directory for PQS jobs [/scr]: /tmp
    Please choose one of the following actions: 1 Skip this step leaving PQS_SCRDIR = /tmp
    Now you can choose to create links to the PQS scripts. Do you want to proceed [y]: n
  3. Fix file ownerships
    # chown -R root:root /usr/local/PQS
  4. Clean up
    # cd /tmp
    # rm -rf pqs_trial-3.3-19.i386
    # exit
  5. Configure WebMO to use PQS as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for PQS
    • Click 'Edit' to configure the PQS interface
    • Verify that the entries are correct; if necessary, edit entries and click 'Submit'
      • PQS version: 3.3
      • PQS directory: /usr/local/PQS-3.3trial
      • PQS binary: /usr/local/PQS-3.3trial/pqs.x
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  6. Login as a WebMO user, and run a test job using PQS as the computational engine

PSI 4

  1. Make sure you have bzip2 installed
    $ sudo yum install bzip2 (for CentOS, Amazon Linux)
    $ sudo apt-get install bzip2 (for Debian, Ubuntu)
  2. Download the miniconda installer
    $ sudo su -
    # cd /tmp/
    # sudo curl -o Miniconda-latest.sh "https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh"
  3. Install miniconda
    # bash ./Miniconda-latest.sh -b -p /usr/local/miniconda
  4. Include miniconda in your path
    # export PATH=/usr/local/miniconda/bin:$PATH
  5. Update miniconda
    # conda update --yes --all
  6. Install psi4
    # conda install --yes psi4 -c psi4 (version 1.3.2 has a bug calculating vibrational frequencies)
    # conda install --yes psi4 dftd3 -c psi4/label/dev
    # exit
  7. Run a test job from the command line
    $ cd ~
    $ mkdir -p test/psi4/
    $ cd test/psi4/
    $ cp /usr/local/miniconda/share/psi4/samples/stability1/input.dat .
    $ export PATH=/usr/local/miniconda/bin:$PATH
    $ psi4 input.dat
    $ more output.dat
  8. Configure WebMO to use psi4 as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for PSI4
    • Click 'Edit' to configure the PSI4 interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • Psi version: 4.0
      • Psi directory: /usr/local/miniconda/
    • Click 'Return to Admin', followed by 'Logout' to exit the WebMO administration page
  9. Login as a WebMO user, and run a test job using Psi4 as the computational engine

QCHEM 5.X

  1. Download the QChem installer
    # cd /usr/local
    # wget -N http://www.q-chem.com/download/qcinstall/qcinstall.sh
    # chmod +x qcinstall.sh
  2. Run the installer, specifying the following options when prompted
    # ./qcinstall.sh
          Please specify the path of the new Q-Chem directory: /usr/local/qchem
          Available packages for LINUX_Ix86_64:
              1 -- Linux 64-bit shared-memory version
              2 -- Linux 64-bit cluster version (requires MPICH3)
              3 -- Linux 64-bit cluster version (requires OpenMPI 1.10)
              4 -- Linux 64-bit version with NVIDIA GPU support (installs BrianQC)
          Please choose package to install or press 'x' to exit [1-4] 1
          Please specify the full path for a scratch directory: /scratch
          Please specify a default value for MEM_TOTAL in megabytes [8000] 8000
          Would you like to view the licensing agreement? <y/n>...<y> n
          Please enter your order number: 0 {or order number}
          Q-Chem User Registration: {supply requested information}
          Is the above information correct?<y/n>...<y> y
          Q-Chem installation/update has been completed.
          Q-Chem has two options for licensing:
              1-- Node-locked licensing
              2-- FlexNet based remote licensing
          Please specify the licensing scheme from above. (<1>/2) {make choice}
          The licenses data has been submitted to Q-Chem office. You should receive a confirmation email and subsequently the license file
          Q-Chem installation/update has been completed.
  3. Install license
    To regenerate license data, run
    # /usr/local/qchem/qcinstall.sh --update-lic
    To install a node-locked license file, save it as /usr/local/qchem/qcaux/license/qchem.license.dat
    To install a FlexNet license, consult the License Administration Guide under /usr/local/qchem/doc/flexnet and the Q-Chem specific instructions in /usr/local/qchem/doc/README.FlexNet
  4. Documentation is available at
    # wget -N https://manual.q-chem.com/5.3/qchem_manual_5.3.pdf
  5. Run a test job from the command line
    • Make a directory to hold test files and set up the environment
      $ cd ~
      $ mkdir qchem
      $ cd qchem/
      $ source /usr/local/qchem/qcenv.sh
    • Copy in a sample job
      $ cp -p /usr/local/qchem/samples/freq/FREQ_water.in .
    • Run Q-Chem using the file as input
      $ /usr/local/qchem/bin/qchem FREQ_water.in >FREQ_water.out
      $ more FREQ_water.out
  6. Configure WebMO to use Q-Chem as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for QChem
    • Click 'Edit' to configure the QChem interface
    • Verify that the first two entries are correct (replace X's with current version):
      • QChem version: X.X
      • QChem directory: /usr/local/qchem
    • Click the 'Submit' button for the changes to take effect
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
  7. Login as a WebMO user, and run a test job using QChem as the computational engine

QUANTUM ESPRESSO

  1. Visit the quantum espresso download page and download the latest espresso-X.X.X.tar.gz file
  2. Install gfortran and make
    $ sudo yum install gcc-gfortran make (for CentOS)
    $ sudo apt-get install gfortran make (for Debian, Ubuntu)
  3. Unzip the source code to /usr/local/
    $ sudo su -
    # cd /usr/local/
    # tar xzf /path/to/espresso-X.X.X.tar.gz
  4. Compile quantum espresso
    # cd espresso-X.X.X/
    # ./configure
    # make all
  5. Run a test job from the command line
    # cd PW/examples/example01/
    # ./run_example
    # cd results/
    # less al.band.cg.out
  6. Enable quantum espresso in WebMO
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for Quantum Espresso
    • Click 'Edit' to configure the Quantum Espresso interface
    • Verify that the entries are correct; if necessary, edit entries and click Submit
      • PWSCF Version: X.X.X
      • PWSCF bin directory: /usr/local/espresso-X.X.X/bin
      • PWSCF pseudopotential directory: /usr/local/espresso-X.X.X/pseudo
      • Use MPI: (unchecked)
    • Click 'Return to Admin', followed by 'Logout' to exit the WebMO administration page
  7. Quantum espresso jobs use pseudopotentials, which must be downloaded separately. These should be downloaded from www.quantum-espresso.org/pseudopotentials/ and placed in /usr/local/espresso-X.X.X/pseudo/

TeraChem

  1. Install a Graphics Processing Unit (GPU) card compatible with TeraChem. Consult the TeraChem User Guide for CUDA and/or Driver requirements. Consult Nvidia for the capabilities of their GPU cards.
  2. Install CUDA
    There are three options for installing CUDA for TeraChem:
    • Install the latest Nvidia video driver. TeraChem includes the CUDA libraries necessary for it to run. While this is the quickest and easiest approach, it is compatible only with TeraChem, and other GPU software vendors require the CUDA libraries to be installed via the toolkit.
    • Install the Nvidia CUDA toolkit from a RPM. This approach uses the the operating system software installation tools (rpm, yum, dnf, apt) to install the necessary software.
    • Install the Nvidia CUDA toolkit from a runfile.
    The following instructions install the CUDA toolkit from a local RPM on a CentOS 7 system. The Nvidia CUDA tookit website provides specific instuctions for other operating systems.
    Determine whether the epel repository is enabled
    $ su
    # yum repoinfo epel
    ...
    Repo-status : enabled
    If not, then enable access to the epel repository
    # yum install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm (for CentOS 7)
    Download the CUDA 11.0 local rpm from Nvidia
    # wget http://developer.download.nvidia.com/compute/cuda/11.0.1/local_installers/cuda-repo-rhel7-11-0-local-11.0.1_450.36.06-1.x86_64.rpm
    Install the CUDA toolkit
    # rpm -i cuda-repo-rhel7-11-0-local-11.0.1_450.36.06-1.x86_64.rpm
    (note: installing 3:kmod-nvidia-latest-dkms took a very long time)
    # yum clean all
    # yum -y install nvidia-driver-latest-dkms cuda
    # yum -y install cuda-drivers
    # exit
    Verify that the video driver is present:
    $ cat /proc/driver/nvidia/version
    NVRM version: NVIDIA UNIX x86_64 Kernel Module 450.36.06 Mon Jun 1 23:19:54 UTC 2020
    Optionally , compile and run a CUDA program
    $ mkdir -p ~/test/cuda
    $ cd ~/test/cuda
    $ export PATH=/usr/local/cuda-11.0/bin${PATH:+:${PATH}}
    $ export LD_LIBRARY_PATH=/usr/local/cuda-11.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
    $ cuda-install-samples-11.0.sh ~/test/cuda
    $ cd NVIDIA_CUDA-11.0_Samples/1_Utilities/deviceQuery/
    $ make
    $ ./deviceQuery
    ...
    Detected 1 CUDA Capable device(s)
    Device 0: "GeForce GTX 1060 6GB"
    CUDA Driver Version / Runtime Version 11.0 / 11.0
    CUDA Capability Major/Minor version number: 6.1
    Result = PASS
  3. Download TeraChem
    Contact TeraChem for a trial license. They will send you a link that allows you to download the latest version of TeraChem
    $ su
    # cd <download_directory>
    # wget --content-disposition http://download.petachem.com\?id=####\&dist=####
    ...'tc1.93P.tar' saved [588298240/588298240]
    # tar xvf tc1.93P.tar
  4. Install TeraChem
    # cd TCInstaller
    # ./install
    Install script for TeraChem 1.93P
    CUDA Driver Version 450 detected
    You have 1 TeraChem-compatible GPUs in your system.
    Do you promise to read the following license carefully? [yes/NO] yes
    Do you accept? [yes/no] yes
    Prefix of directory to install TeraChem[ /root ]: /usr/local
    TeraChem 1.93P has been installed in /usr/local/TeraChem
    To run TeraChem:
      source /usr/local/TeraChem/SetTCVars.sh
      /usr/local/TeraChem/bin/terachem inpfile
    where inpfile is a valid TeraChem input file
    To obtain your permanent license:
      Please fill in the information below
      and send to help@petachem.com.
  5. Fix file ownership
    # chown -R root:root /usr/local/TeraChem
    # exit
  6. Run a job from the Command line
    $ mkdir -p ~/test/terachem
    $ cd ~/test/terachem
    $ source /usr/local/TeraChem/SetTCVars.sh
    $ cp -p /usr/local/TeraChem/tests/caffeine/start.631.sp .
    $ cp -p /usr/local/TeraChem/tests/caffeine/caffeine.xyz .
    $ terachem start.631.sp >output.631.sp
    $ cat output.631.sp
    $ ls -alF scr/
  7. Configure WebMO to use TeraChem as a computational engine
    • Login to WebMO as user 'admin'
    • Click 'Interface Manager' to enable the interfaces to any computational chemistry packages that you have installed on your system
    • Click the 'Enable' button for TeraChem
    • Click 'Edit' to configure the TeraChem interface
    • Verify that the entries are correct; if necessary, edit entries (replace X's with current version) and click 'Submit':
      • TeraChem version: X.XXX
      • TeraChem directory: /usr/local/TeraChem
      • TeraChem environment setup script: /usr/local/TeraChem/SetTCVars.sh
    • Click the 'Submit' button for the changes to take effect
    • Click 'Return', 'Return to Admin', and 'Logout' to exit the WebMO administration page
    • Login as a WebMO user, and run a test job using TeraChem as the computational engine