As announced in the group meeting on Monday, I will be giving a short (15-20 min) interactive introduction to working with the ETHZ Euler Cluster this Friday, February 15th, 2pm, HIT E 41.1. Please feel welcome to attend but not obliged if you do not feel like this session is necessary for you. For anybody who is interested but can not make it this Friday, I can also quickly go over the basics with you in person. 

The ETH clusters can be very useful for anybody doing computationally intense simulations or data processing. The cluster is essentially a second, powerful computer where you can "outsource" jobs from your machine to. This means that once you started a program on the cluster, you may turn off your own computer or work on something else and check back at a later time if the job has finished. The cluster supports all sorts of programming environments (Python, Matlab, Mathematica, ...) and has a lot of packages pre-installed for direct use.

In order for us to run a very basic exemplary code on the cluster, I have attached a Python file to this email. Please download it and save it to a location that you will be able to find using the terminal during the tutorial, for example the Desktop. There is no need to install python and run this code on your own computer.

For Windows users: Windows does not provide an included ssh client. In order to access the cluster, you will therefore need to download third-party software for this. ETH recommends Cygwin for this (http://www.cygwin.com). There are other options that you might prefer, but with Cygwin the syntax should be the same as it is for Mac and Linux users, so I would go for this solution for the tutorial. Please download and install Cygwin or some alternative before the tutorial, then we can all follow along without many delays.

If you want to follow along in this interactive session, make sure to bring your prepared Laptop with the files listed above. Other than this, we do not need anything else.

Yes. Using the cluster is actually not that hard and ETH has done a very good job in documenting the first steps of using the cluster. The aim of our tutorial is more to show what working with the cluster looks like and to set it up directly for everybody who thinks might work with it in the future. 

You can also check out https://scicomp.ethz.ch/wiki/Getting_started_with_clusters if you prefer to have a look at it by yourself. Those who are attending the tutorial do not need to read up on anything, we will look at it together. 

Unfortunately I will not provide anything other than this somewhat prepared tutorial, but of course I highly encourage anybody who is motivated to bring food or drinks to lighten up this Friday afternoon. Otherwise an ITP coffee will probably do.

If there is anything else that I forgot or that you would like to know, don't hesitate to write me an email. 

Looking forward to the Friday session, 

see you all then, 

Lennart
--