CISL Daily Bulletin

The Daily Bulletin delivers news about CISL’s high-performance computing environment, training opportunities, scheduled maintenance, and other events of interest to the user community.

HPC systems returned to service after scheduled maintenance

October 21, 2022

NCAR’s supercomputing and storage systems were all returned to service on Thursday following completion of scheduled preventive maintenance and security updates. As announced, any impact on users should be minimal. Some may notice minor changes in their GLADE file space quota reports as a result of block size adjustments in /glade/u/home directories. The block size for the home file space, as n...

Correction: Globus unavailable November 7-8 for Version 5 update

October 13, 2022 Updated: October 17, 2022

CISL’s Globus file transfer services will be unavailable from 8 a.m. MST on Monday, November 7, through 5 p.m. MST on Tuesday, November 8, as the new Version 5 is implemented. Transfers in progress during that time frame will be canceled and will need to be restarted after the update. Enhancements coming in the new version will make it easier for researchers to share scientific data with collea...

Using Casper GPUs to prep for Derecho

October 17, 2022

Researchers with Casper allocations are welcome to use the analysis and visualization cluster to begin preparing for the extensive GPU capabilities of the new Derecho system when it becomes available in 2023. Casper already has 19 nodes with either NVIDIA GP100 GPUs or V100 GPUs in addition to three nodes with A100 GPUs. More A100s will become available in the next few months. Derecho will get ...

System downtimes for week of October 17-21

October 17, 2022

Cheyenne, Casper, GLADE, Campaign Storage, and JupyterHub will all be down from 10/17-20 for the previously announced fall maintenance. Quasar and Stratus will also be down temporarily..

Match your memory request to your job’s needs

October 10, 2022

A lot can go wrong if a job you try to run on NCAR’s HPC systems requires more memory than what’s available on the nodes you want. If you request a lot more memory than you need, on the other hand, your job may be slow to start. What do you do? CISL provides several options to help you determine how much memory your program is using so you can request the right amount and avoid problems. See Ch...

No system downtimes for week of October 10-14

October 10, 2022

None planned for Cheyenne, Casper, GLADE, Campaign Storage, Quasar, Stratus, or JupyterHub.