Skip to content

Parallel MPI code for lulesh benchmark optimisation and matrix multiplactions

Notifications You must be signed in to change notification settings

aesir321/supercomputers

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Supercomputers

A Repository to store our work for the Supercomputers course at TUM.

Authors: John Kourdis, Patrick O'Connor and William Parker

Accessing SuperMUC:

First users will need to ssh into the TUM rechnerhalle using the command

~:ssh @halle.in.tum.de

To run programs on the superMUC you will then need to ssh into your provided superMUC account

~:ssh @supermuc.lrz.de

SuperMUC Architectures

To access the Phase one architecture use the following command:

~:ssh [email protected]

To access the Phase two architecture use the following command:

~:ssh [email protected]

Running a batch on the SuperMUC

Once you have logged in to your chosen architecture you need to run a batch in order to get stable results. In order to submit a batch to run a program use the following command:

~:llsubmit ll.sh

To query the status of a job you can use the following

~:llq -u username

To cancel a job run the following

~:llcancel jobid

Recommended workflow:

It is recommended that users clone the git repository found here: https://[email protected]/wparker89/supercomputers.git using the command

~:git clone https://[email protected]/wparker89/supercomputers.git

in their halle.in.tum.de workspace. All development should be carried out here and run locally, any changes should be pushed using a normal git workflow.

Transferring files to and from SuperMUC

Up until now I have not been able to clone a git repo onto the SuperMUC but it is possible (apparently). Until we figure that out recommended to transfer files with rsync. Used as follows:

~:rsync -a ~/dir_to_copy username@remote_host:destination_directory

This will use ssh to copy the specifed directory and all subdirectories to the destination_directory. For more information see: https://www.digitalocean.com/community/tutorials/how-to-use-rsync-to-sync-local-and-remote-directories-on-a-vps

rsync cannot be used on the SuperMUC itself. In order to copy your results from SuperMUC to your local workspace you need to run the following command:

~:rsync -a username@remote_host:dir_to_fetch local_directory_to_copy_to

Report

The report will be written in Latex using the platfrom sharelatex, so that we can all easily collaborate on producing it. The report can be accessed here: https://www.sharelatex.com/project/5a01a12bf091a63ae216d57d

About

Parallel MPI code for lulesh benchmark optimisation and matrix multiplactions

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published