Summary Outline of Topics in this message:
System notices:
- Maintenance outage: May 15th-19th, Monday 8:00am - Friday 5:00pm - Changes in /fastscratch system and policies
- New cluster in the works
Welcome new team members:
- Ayat Mohammad, Computational Scientist
Upcoming conferences and VT workshops
- Super-Computing
- PEARC
How to get help
----------------------------------
All ARC systems will offline for maintenance the week following commencement. In particular, the Tinkercliffs and Infer clusters, the OnDemand website, and ARC's underlying storage systems (/projects, /home, /fastscratch, /vtarchive) will all be inaccessible for the duration of the outage. Jobs which cannot finish before the start of the maintenance will be held in queue until the clusters are released. Please terminate any persistent connections or processes (SSH, screen, tmux) you may have on the login nodes prior to the start of maintenance.
Changes in /fastscratch system and policies
The /fastscratch file system serves as a convenient place for staging jobs and for temporary storage of data. However, it has been persistently full for the last several months which has severely affected its usability. During this maintenance outage we will implement policies to limit each person to 5TB of storage on that system and will also begin to enforce regular cleanup/deletion of data from that file system.
ARC has been able to acquire a new computational resource which should be arriving soon to begin the commissioning process. This will be new cluster mainly comprised of 84 CPU nodes. These nodes are each equipped with:
- 2 AMD
EPYC 9454 "Genoa" series processors for a total of 96 cores
- 768GB
DDR5-4800MHz memory via 24 channels (4 cores per channel, 8GB per core),
- HDR Infiniband connected at 100GbE
The system will be ARC's first water-cooled system which should enable running at boost speed (3.8GHz) "indefinitely".
This acquisition also includes several large memory nodes to provide capabilities for workloads which demand up to 4TB of system memory.
Upcoming conferences and VT workshops
Software and Data Carpentries
Start your summer research right with data and programming workshops!
Students, faculty, staff and community members are invited to join us for free workshops on R, Python, Git, Excel, and more hosted by the University Libraries. All workshops are introductory-level, cover a range of fundamental skills so you can start working with tools right away, and open to anyone! To learn more about the curriculum or sign up, visit the
PDN (faculty/staff/grads) or
Library Calendar (undergrads/community) or contact Data Education Coordinator
Nathaniel Porter.
In-Person (University Mall 2104):
5/15-16 9:00-5:00 Data Carpentry Social Science: R, Excel, and OpenRefine
5/18-19 9:00-5:00 Software Carpentry: Unix, Git, and Python
5/22 9:00-12:00 Software Carpentry: The Unix Shell
5/23 9:00-12:00 Software Carpentry: Version Control with Git
5/24 9:00-4:00 Software Carpentry: Plotting and Programming in Python
5/25 9:00-4:00 Software Carpentry: R for Reproducible Research
PEARC23 (Practice and Experience in Advanced Research Computing)
July 23-27, 2023
November 12-17, 2023
How to get help with using ARC systems:
Best regards and happy computing!
Advanced Research Computing
Virginia Tech