About BC Research Services
Research Services provides consultation, training, and tutorials to support computing and research at Boston College. The Research Services staff has expertise in statistics, statistical software, qualitative research, surveys, GIS, scientific computation, data analysis, computational mathematics, and visualization. We also provide assistance with coding, databases, Web technologies, and administration and consultation for research oriented computing hardware based within the departments and institutes.
Andromeda Support Help Sessions
Research Services is available to assist with Andromeda. We also offer walk-in office hours for support with cluster-related questions.
Time:
Monday (Online): 1:00-2:00 p.m.
Thursday (Hybrid): 1:30-3:00 p.m.
Location:
Online Meeting (Monday and Thursday): https://meet.google.com/qvh-kojm-vuv
In-person Support (Thursday): O’Neill Library 528 (ITS Conference Room)
Announcements
Did you know that all BC Researchers have access to JMP Statistical Software for Free?
JMP is a powerful and intuitive point-and-click program for statistical analysis. Download JMP Pro using your BC login credentials. To learn more, attend our JMP & Jamovi Tutorial or attend an Intro talk at the free online JMP Discovery Summit.
Dedicated Debugging GPU Node
A new interactivegpu single-node partition is now available to help you test and troubleshoot your GPU-accelerated code.
Node Specification
4 GPUs: 192 GB RAM
60 CPU cores: 512 GB RAM
NVIDIA Driver: 575.57.08
CUDA Version: 12.9
Usage Policy
Purpose: GPU task debugging, code development, and interactive testing only. This node should not be used for long-running batch production jobs.
Time Limit: 1 hour per session. Jobs will automatically terminate after 1 hour to keep the resource available for others.
Recent Update (Released on September 18)
Dear Andromeda Users
We have completed the patching to the Slurm software and have removed the reservation from Andromeda. Submitted jobs will start to run.
I know that downtime is inconvenient and we will always attempt to keep downtime limited and combine patching cycles. However, since Spring of last year we have seen a 9-fold increase in critical and overall patches. This is largely due to AI tools similar to Mythos finding vulnerabilities in software. The latest case is a critical vulnerability in Slurm that could lead to an authorized process gaining root access to Andromeda.
As always we will first try to minimize impact to the Research Community.
Thanks
Matt
HPC Services Overview
Andromeda is in a healthy state.
The old cluster (“Andromeda 1”) has been permanently offline since 09/15/2025.
If you have any questions or require assistance, please reach out to us via Research Help.