Although it isn't officially Spring at Harvard, MLB pitchers report in February for Spring Training. The following list of 30-45 min sessions will be offered each Fall and Spring, both at FAS and HSPH.
Registration : Registration is not required, but seating is limited, so come early.
Intro to Odyssey: Slides
Come learn the basics about the Odyssey Supercomputer and other Research Computing services. This session will cover cluster access, file transfers, storage options, and basic compute-job submission using SLURM.
FAS: Thursday, February 2nd 11:00AM in NWL 426
HSPH: Thursday, February 9th 12:30PM in Kresge 205
Intro to Unix
This will be a hands-on learning environment that will teach the basics of using the command-line and the top 10 most essential Unix commands. Please bring your laptop and already have an FASRC account.
FAS: Thursday, February 16th 11:00AM in NWL 426
HSPH: Thursday, February 23rd 12:30PM in Kresge 205
This session will dive further into Unix core-utilities including sed, grep, and awk. We will also cover the basics in bash shell scripting, which should help you think about automating processes and building more complex job scripts.
FAS: Thursday, March 2nd 11:00AM in NWL 426
HSPH: Thursday, March 9th 12:30PM in Kresge 205
Modules and Software
Having issues wading through which of the 3000+ pieces of software installed on Odyssey you should be using? Still getting "command not found" or missing a shared object dependency? This session will illuminate the simplicities and intricacies of using software on Odyssey. We'll address: setting up your environment; the Lmod module system and how it works to your advantage and updating perl / R / python modules & packages in your environment.
FAS: Thursday, March 16th 11:00AM in NWL 426
HSPH: Thursday, March 23rd 12:30AM in Kresge 205
Choosing Resources Wisely
Now that you know how to submit a basic job script to SLURM, knowing the appropriate resource needs (queues, number of cores, memory, storage, time, etc) for your jobs can be puzzling, but it can be a relatively straightforward task. In this session, we will provide you the tools needed to make the most efficient use of RC cluster resources, which will ultimately lead to more research throughput.
FAS: Thursday, March 30th 11:00AM in NWL 426
HSPH: Thursday, April 13th 12:30AM in Kresge 205
Are you not getting the expected output from your job? Is your job running slower than expected? In this seminar, we will focus on increasing your troubleshooting skills and how to translate common errors. There are a few simple things to include in every job script that will help, as well, there are some common Unix tools to learn. In addition, we will cover the types of information you need to provide RC when you do encounter the unexpected. More-or-less, "help us, help you" to accelerate the time to solution. You are encouraged to bring your laptop and follow along with the examples.
FAS: Thursday, April 6th 11:00AM in NWL 426
HSPH: Thursday, April 27th 12:30AM in Kresge 205
Parallel Job Workflows on Odyssey
There are three main types of parallel workflows: pleasantly parallel (job-array), tightly coupled (SMP threads), distributed parallelism (MPI). Each of these workflows has a different requirement when submitting to the SLURM job scheduler. In this session, we will cover the basic concepts around parallel processing and provide examples on how best to submit these types of jobs.
FAS: Thursday, April 20th 11:00AM in NWL 426
HSPH: Thursday, May 11th 12:30AM in Kresge 205
Last updated: February 18, 2017 at 16:56 pm
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Permissions beyond the scope of this license may be available at Attribution.