CCML

This project is maintained by CCML-UCONN

Getting Started


This page is still under construction.

Running Jobs on TACC

Most of our jobs require dedication of multiple nodes. Therefore jobs cannot be executed on the front node(the node everyone interfaces when logged onto the cluster). The HPC cluster use a job management system to process the resource requirements and its distribution to working/slave nodes. Some more useful information about the batch system on HPC-Storrs can be found here

You can find several different job submission scripts in the following directory: $GRPDIR/share/example/scripts/batch

Pick the one meets your need most and edit the corresponding fields.

Once you account on Stampede has been activated. Follow the instructions and tests to make sure everything is set up properly and functional.

Contents

  1. Installation
  2. Logging On
  3. First Time Logging On
  4. Making Sure Everything Works

Installation

Mac OSX

Download and install:

To prevent X11 from timing out, open the terminal and type:

mkdir -p ~/.ssh
echo $'\nHost *\n ForwardX11Timeout 1000000\n' >>~/.ssh/config

Windows

Download and install:

Linux (Debian-based, e.g. Ubuntu)

From the terminal __

Logging onto the Clusters

For the Stampede cluster, make sure to read through the User Guide here.

login with your Stampede username (most probably start with tgxxxxxx, check your account here)and password (you will aslo be asked for Mulit-factor Authorization).

Follow the instructions below for your system:

Mac OSX

Open “Applications-> Utilities -> Terminal” or “Command+Space” to search Terminal using “spotlight search” In a terminal:

ssh -X username@stampede.tacc.utexas.edu

Windows

Launch Xming. You will always need to have this open in order to forward graphical windows from the external clusters.

Start PuTTY, and:

You can start putty several times, if you need several terminal windows; only one instance of Xming needed.

Linux

In a terminal:

ssh -X username@stampede.tacc.utexas.edu

First Time Logging in

For the first login only, run the following command:

cp /home1/03672/tg829713/vojgroup/bash_script/bashrc_copy ~/.bashrc
source ~/.bashrc

This will enable you to run specific software on the Stampede cluster, including the ASE interface to Quantum ESPRESSO.

There are two file partitions, the home and the work partition. Go ahead and make a symbolic link to the work partition using:

ln -s $WORK work

Change the permission of files:

cd
cd work
mkdir CBE544
chmod g+X $WORK
chgrp -R G-818582 CBE544
chmod g+s CBE544
chmod g+rX CBE544

Perform all your calculations from the work/CBE544

For example, Create a folder hw5 under CBE544 and perform all calculations of HW5 in hw5. If you have already started you jobs somewhere else, you can copy the entire folder to CBE544 once you have done all calculations (e.g. cp -r folderpath ~/work/CBE544/hw5). However, you need to complete this step before sending us the report of HW5, and please include the path of CBE544 folder in your report or email. You can obtain the path by

cdw
cd CBE544
pwd

__

Making Sure Everything Works

Once you are logged into the terminal, run:

ase-gui

and make sure a graphical interface appears. Next, run Python in interactive mode by typing:

python

and make sure the following commands work:

import ase
import numpy