Earlier this month, I sat down with Sylvain Pitre, Manager of Research Computing Services, to talk about what he and his team do for Carleton researchers.
What is it that your team does?
The three of us, Andrew, Blake and I, help faculty and students who are working on projects with advanced research computing needs. These projects generally involve big data, large computational power or modelling that cannot be handled by standard computing infrastructure.
We work with these faculty members to understand their research computing needs and provide the support or computing resources they require.
There are so many new opportunities in big data, particularly for the non-technical departments. At Big Data 3.0 researchers from Public Policy and Administration, History, Journalism and Health Sciences spoke to developments, needs and opportunities for Big Data in their departments.
Our services are available to any researcher, in any department, who is in need of high computing resources.
What kind of support do you provide?
We offer a number of different support services depending on the individuals needs.
We’ve assisted with the porting, optimization and running of code and applications.
For example, a researcher from the Economics department is using big data to conduct time series analysis. We helped him by looking at the code he was running and tweaked it to run faster. This helped him to focus more on his research and less on the necessary, but often time intensive, computing processes behind analyzing the research.
Another researcher in Biology is looking at genetics of adaptations. He and his student were working with some Open Source code and needed help writing code around it in order to reduce run time. What would have taken six months on a laptop now takes three days using a cluster. Because they need to re-run the code on a regular basis, they can now get the results they need to do their analysis faster.
We’ve helped faculty with grant proposals by contributing sections that describe the interplay between their research and advanced research computing resources. Granting agencies, like SSHRC, want to know what resources the university provides and they will then often match it. We help researchers with the details in describing the resources and big data tools that Carleton provides.
In addition, we host and offer a number of workshops on working with big data. We recently hosted Watson Analytics Workshop. This workshop focused on business analytics to discover insights to: recruitment, enrolment, financial aid, institutional research, student experience, student retention, res life, diversity, facilities, campus safety, and more
We’re also providing the DATA-5000 class (in the collaborative Master’s in Data Science program) support with access to Carleton’s Virtual Computing Lab (VCL) and IBM Cognos Business Intelligence software.
This is an interesting program in that it brings together students from science, engineering, arts, social science, public affairs and Sprott. They sign up for a masters degree in one of these departments and then they add a specialization in data science. This year the program had 40 students (double than last year) – 20 from the technical programs, 20 from the non-technical.
What kind of computing resources do you offer to researchers?
Right now we offer access to open and commercial software through the VCL. Software packages available include IBM Cognos Business Intelligence, IBM SPSS Statistics, MathWorks’ Matlab, Stata/SE 13, MPICH and Hadoop.
These services are available at no cost to researchers in all faculties at Carleton.
Additional Computing Resources
There are currently a number of specialized computing resources on campus – Physics Research Computing Cluster, Mechanical and Aerospace Engineering Computing Cluster, School of Computer Science Research Facilities, and VSIM.
We offer computing resources for researchers who do not require or are not eligible to use these specialized services.
For researchers working with big data requiring national level support, Compute Canada is a great resource. To be able to use these resources, researchers much submit a proposal (with which we’re happy to help), but the reality is that researchers are getting on average 64% of their requested resources. This leaves some researchers who are in the middle – needing more than a laptop but not as much as Compute Canada offers – to fend for themselves.
Our goal is not to discourage researchers from using Compute Canada, but to offer something in between. Which is why we’re creating a Research Computing Cluster.
Tell me more about the Research Computing Cluster.
We’re working on opening a new research computing cluster here in ITS in late spring 2016. It’s funded by the Faculty of Science, Faculty of Engineering and Design and ITS and the goal is to support researchers from across campus.
There are so many new opportunities in big data for the non-technical departments, for example, mining social media data in Journalism. This computing cluster will be available to any researcher who is in need of high computing resources.
It will be an OpenStack-based cloud solution with 160 IBM Power8 compute cores, 192 Intel, Xeon compute cores for a total of 1.8TB storage (128GB per node) and 60 TB all-flash storage (SSD). A modern laptop might have 4 cores. There are 352 total in the cluster.
What do you see in the future for research computing?
We plan to continue to grow our client base and offer services to researchers who would are looking for compute and storage resources on campus. We want to be the first point of contact for any researcher on campus looking for these services and find them what they need, on or off-campus.
Working close with researchers and discussing their requirements helps us determine services that are not currently within our scope but that could be opportunities in the future.
For more information about the services that Sylvain and his team provide, visit the Research Computing Services website.