The Ohio Supercomputer Center (OSC) is an incredible resource for academic researchers across the state … so long as they know how to use it.
Most of my work as a post-doctoral researcher within Ohio State’s chemistry department was computational. And I was lucky when I arrived in the department because my group there had been a significant user of OSC services, meaning I was able to learn from my colleagues. They introduced me to their workflow, along with the scripts and steps they used, before I got started.
Institutional knowledge is the most common way people learn to use supercomputers. If you have a group that has that acquired knowledge, it’s easy to get started. And when you make a mistake, somebody tells you “Yeah, I did that before. Here’s where you went wrong.”
But what if you don’t have that institutional knowledge? For those who fall into that category, it can take time to tap fully into OSC’s deep resources.
That’s a rising challenge now because new groups, new departments, and new disciplines – such as linguistics, computational biology, or economics – are beginning to use high performance computing (HPC) more every day. In those areas, users don’t have many seasoned colleagues to turn to within their own department, or their own college, so it takes them longer to get started and become productive.
That’s a learning gap we’re trying to address through our workshops, which take us to a different section of Ohio each month. OSC has held these workshops regularly throughout the years, but recently we’ve been working to coordinate more stops throughout the state to help introduce our academic clients to HPC and to address whatever roadblocks they may have.
Our goal is to get to each of the different regions of Ohio at least once per year, if not once per semester.
The main workshop we present is specifically for new users and includes:
- An introduction to OSC
- The basics of what HPC is
- Why a client would use HPC
- Examples of projects that have been done on our systems and through HPC in general
- What a supercomputer is
- The parts that make up a supercomputer
- Specifics about hardware
- What types of hardware we have and how a client can use them
- How to get an account
- How to log in
- How to create and submit a job to the batch system
- Specifics on our file systems
- And our OnDemand web portal, so clients can access those systems.
We also have more specialized workshops for people leveraging HPC for Big Data analysis, which incorporates Hadoop and Spark.
For future workshops, we’re creating classes on profiling and debugging, optimizing code, an R workshop for HPC to go along with OnDemand, and potentially, a course for using our new NVIDIA Tesla P100 GPUs.
Academic clients able to deftly use OSC resources have been able to quickly gain incredible insight into their projects. Our goal with these workshops is to help more researchers more quickly discover those amazing breakthroughs that benefit so many of us in our daily lives.
One more note: We’ve been doing these workshops at an institutional level. But if there are groups that would like an introduction to this at a department or a research group level, that’s certainly something we can do, too. If we do it at a department or within a discipline, we can be more specific to meet those needs.