Social Science Computing Cooperative

SSCC News for September 2025

Posted on

Welcome to the SSCC!

If this is your first SSCC News, we welcome you to the Social Science Computing Cooperative! Whether you’re a new faculty member, a new grad student, or an undergrad taking a class that uses SSCC services, we’re here to help you succeed. Visit the SSCC web site to learn about the services we provide.

SSCC News comes out roughly once every two months. If you’d rather not receive it, feel free to reply with ‘unsubscribe’. But if you’re not interested because you no longer need your SSCC account, let us know and we’ll close it for you.

Outage for All SSCC Services Night of 9/16-9/17

On the night of September 16th all power in the Sewell Social Sciences Building will be shut down due to construction work. This will require shutting down the equipment in the SSCC datacenter and all the computers in the building.

Please power off your office computer at the end of the day. We will attempt to remotely shut down all the computers we manage in the Sewell Social Sciences Building starting at 6:00PM on the 16th. We will begin shutting down SSCC services (Winstat, Linstat, etc.) starting at 7:00PM. All services will be available again at 8:00AM on Wednesday, September 17th.

Please send any questions to the SSCC Help Desk. We apologize for any problems this outage causes, but it is out of our control.

SSCC Training

We’ve got several popular training workshops coming up, including Functions and Iteration in R, Stata Macros and Loops, Using the SSCC Linux Servers, and more. Visit the training page for the full schedule.

In response to your requests, we’ve added an Introduction to Stata session for this Friday afternoon (9/12).

Caitlin will be leaving us in December, so if you’re interested in qualitative analysis and software like NVivo or MaxQDA, now is the time to take her workshops or meet with her one-on-one!

Windows 10: The End is Near

Support for Windows 10 ends on October 14th. After that, security problems will no longer be patched, and hackers are probably saving up exploits for use after that date. Computers running Windows 10 will be in extreme danger of being compromised and of compromising other computers they interact with, so the time to upgrade or replace any computers that run Windows 10 is now. The SSCC will begin blocking Windows 10 computers from SSCC resources after October 14th.

There are still some SSCC-managed computers that need to be upgraded. If you have one, please turn it on and connect it to the SSCC network, either by VPN or by plugging it into the wired network in the Sewell Social Sciences Building. Automatic tools will take it from there, given time. Upgrading Windows will take a while, so don’t wait until you need that machine. If you’re not sure about the status of an SSCC-managed machine, reach out to the SSCC Help Desk.

Upgrade-Proof Your Python Code and VSCode Settings with Conda

As part of last month’s updates we upgraded the default Python on all SSCC servers and, as expected, that caused problems for some Python users. In some cases VSCode was unhappy when the Python version it had been told to expect disappeared.

A conda environment contains the Python version of your choice and compatible versions of the packages you need, and you control when and if it is upgraded. We strongly recommend you create conda environments for your Python projects so you can avoid problems caused by Python updates and incompatible packages. Note that you’ll need to use the Linux servers rather than Winstat to create conda environments, but we suggest serious Python users use Linux anyway. (Sign up for a workshop on using the SSCC’s Linux servers and we’ll show you how.)

Slurm Tips

The Slurm cluster is often “full” in the sense that there are jobs waiting in the queue (though having to wait for more than a few hours is very rare) but it could do more work if jobs used the available resources more efficiently. Some tips:

  • Read the efficiency reports included in the email notifying you that your job finished. If your job didn’t use most of the resources you reserved, you can reserve less next time. We see too many jobs reserving several times more memory than they need or reserving dozens of cores and then actually using one.
  • The server reserved for short jobs (<6 hours) is frequently idle–and it’s a high-memory server. Submit jobs to the short partition to use it (--part=short). They’ll run on other servers too if they’re available, so there’s no downside to using the short partition rather than the default sscc partition if your jobs will finish in less than six hours.
  • It’s fairly rare for jobs to be preempted, but if your jobs are very long (>24 hours) consider adding the --nopreempt switch. This will prevent your job from running on servers where someone else has higher priority than you. Your job may take longer to start as a result, but it will never be preempted.
  • When jobs are preempted, it’s usually jobs submitted to the econ partition being preempted by jobs submitted to econ-fac or econ-grad, which have higher priority. Submitting to econ will give your job more server options (assuming you’re in Econ), but if the job is long you’re probably better off using econ-fac or econ-grad and --nopreempt so it will only run on the servers where you have top priority. (econ-fac is for Econ faculty and econ-grad is for Econ grad students.)

For more information, see the Partitions & Priorities section in the Slurm chapter of the Guide to Research Computing at the SSCC.