Quick reference

The tutorial assumes some familiarity with the Linux command-line environment. If you're unsure there is a quick reference guide available. Tutors will be happy to help if you need it!

'Workflow' machine

workflow.iu.xsede.org is the general-purpose machine that we will use for this tutorial. It is a multi-core Linux server and is shared by all participants at the tutorial. You will use this machine for the all of the practical sessions, both for running CoCo and LSDMap directly, and also for launching ExTASY workflows on the remote HPC machines. It is also possible to do this from your own workstation or laptop, but we have tested the software environment on this machine to avoid any software installation problems during the tutorial.

You can log in using your XSEDE guest account as follows:

ssh -X USERNAME@workflow.iu.xsede.org

• Your username is trainYY where YY is a two-digit identified in the range 40-65. Conveniently, the IDs for your XSEDE guest account your ARCHER account are related by YY+80 = XXX

• Your home directory is /home/USERNAME

• Shared data files needed for the tutorial are in ~merzky/accounts/, but copies and links should also already exist in your home directories on that machine.

ARCHER

ARCHER is a Cray XC30 supercomputer, the UK's National HPC service. It has 4900 nodes, each with 24 Intel Sandy Bridge cores. Extensive documentation is available on the ARCHER website. For this tutorial you will not normally need to log in to ARCHER directly using SSH, instead you will use the ExTASY tools to submit jobs from workflow.iu.xsede.org.

To do this you will need to following information:

• Your username is guestXXX where XXX is a three-digit identifier in the range 120-145 which you can find on the ARCHER guest account slip. This is your UNAME for ExTASY workflows. The username is also printed on screen when you login to the workflow machine.

• You are part of the project y14-ExTASY. Use this as your ALLOCATION for ExTASY workflows.

• The guest accounts have access to the normal ARCHER queuing system QUEUE='standard' which is available 24/7, however jobs may take a long time to run if the machine is busy. During the tutorial, a special reserved queue is available, which contains 20 nodes and is available 9:30 - 17:00 on each day of the tutorial. To submit jobs to this queue use QUEUE='course1'.

At some point, you may wish to log in to ARCHER to check the status of your jobs in the queue. SSH keys have been prepared that will allow you to log in from the workflows machine to ARCHER without requiring a password. To log in do:

ssh login.archer.ac.uk

or in short:

ssh archer

Then you can view the status of your jobs:

qstat -u guestXXX

XSEDE Stampede

Stampede is a Dell PowerEdge cluster, part of the XSEDE environment. It has 6400 nodes, each with 16 Intel Sandy Bridge Cores. Documentation is available on the XSEDE Stampede website. For this tutorial you will not normally need to log in to Stampede directly using SSH, instead you will use the ExTASY tools to submit jobs from workflow.iu.xsede.org.

To do this you will need to following information:

• Your username is tg45ZZZZ where ZZZZ is a four-digit identifier which you can find on the guest account slip. This is your UNAME for ExTASY workflows. The username is also printed on screen when you login to the workflow machine.

• You are part of the project TG-TRA100001S. Use this as your ALLOCATION for ExTASY workflows.

• The guest accounts have access to the normal Stampede queuing system which is available 24/7, however jobs may take a long time to run if the machine is busy. For a better job turnaround, submit jobs to the development queue QUEUE='development'. During the tutorial, a special reservation is available, which contains 50 nodes and is available 9:00 - 17:00 on each day of the tutorial.

At some point, you may wish to log in to Stampede to check the status of your jobs in the queue. SSH keys have been prepared that will allow you to log in from the workflow machine to Stampede without requiring a password. To log in do:

gsissh -p 2222 stampede.tacc.utexas.edu

or in short:

gsissh stampede

Then you can view the status of your jobs:

showq -u tg45ZZZZ