Daily Bulletin Archive

Sep. 27, 2017

Acknowledging the support of NCAR and CISL computing when you publish research results helps ensure continued support from the National Science Foundation and other sources of funding for future high-performance computing (HPC) systems. It is also a requirement of receiving an allocation, as noted in your award letter.

The reporting requirements, how to cite your use of various systems, and recommended wording of acknowledgments can be found on this CISL web page. The content of citations and acknowledgments varies depending on the type of allocation that was awarded.

Sep. 27, 2017

CISL staff are recommending that users avoid the tar command’s -S and --sparse options for handling sparse files when creating an archive. A recently discovered data corruption issue resulting from use of those options can affect tarballs that users create and it also has corrupted some automated GLADE user home directory backups over the past several months. Snapshots have not been affected, so users who suspect some file corruption may be able to use those to recover files.

Users are encouraged to avoid using the -S or --sparse options until a fix is in place and to watch for a Notifier message announcing that the work is complete.

Sep. 18, 2017

CISL is now accepting requests from university-based researchers for large-scale allocations for the new 5.34-petaflops Cheyenne cluster; submissions are due September 18.

Large allocations on Cheyenne are those of more than 400,000 core-hours. CISL accepts requests from university researchers for these large-scale allocations every six months. For submission instructions and information regarding available resources, see the CISL HPC Allocations Panel (CHAP) page.

Please contact cislhelp@ucar.edu if you have any questions about this opportunity.

Sep. 15, 2017

NCAR researchers and computational scientists are encouraged to submit requests for NCAR Strategic Capability (NSC) projects to be run on the Cheyenne system.

The deadline for NSC submissions is September 15. NSC allocations target large-scale projects lasting one year to a few years that align with NCAR’s scientific priorities and strategic plans.

For information please see the NSC Allocation homepage
Sep. 15, 2017

Course material for a self-paced Fortran workshop is being made available on GLADE for easy access by Cheyenne users. Files for the Scalar Fortran portion of the workshop are ready now, to be followed by Array Fortran, Object-Oriented Fortran, and Parallel Fortran material.

Users are encouraged to copy the files from this directory to their own GLADE space, where they can read, compile, and execute the examples:


Taken in order, the examples show how definition and scope work in Fortran. Declarations, control flow, bit models, interacting with the operating system, input/output transfers, and formats are covered.

Those interested are encouraged to subscribe to the Fortran Interest Group mailing list for updates and to contact Dan Nagle in the CISL Consulting Services Group with questions or comments.

Sep. 12, 2017

The maximum usable memory on Cheyenne’s batch nodes has been reduced by 1 GB, to 45 GB and 109 GB for the small and large memory nodes respectively. Also, the maximum PBS job script size has been set to 10 MB. Previously no limit was in effect.

The purpose of these changes is to improve overall system stability and the reliability of jobs running on shared nodes. Most users will be unaffected, but jobs that previously were close to using all available job memory may now fail. Please contact cislhelp@ucar.edu if you need assistance modifying your job scripts to meet these new limits.

Sep. 12, 2017

The Cheyenne cluster will be unavailable Tuesday, September 12, from 8 a.m. until noon MDT  to allow CISL staff to perform important system updates.

A system reservation will be put in place on batch jobs Monday afternoon. Jobs submitted with specified job times that overlap the reservation period will remain on hold until the system is restored to service. Running jobs that have not finished when the outage begins may be killed and need to be resubmitted after the system is restored to service. Users are advised to take this into account before submitting long jobs after 8 p.m. on Monday.

We apologize for any inconvenience this might cause. Users will be informed via the CISL Notifier service when the system is returned to production.

Sep. 11, 2017

Users can now get a quick, timely look at Cheyenne system activity with the expanded CISL Resource Status web page. A new table on the page shows the level of activity in each Cheyenne queue, including the number of jobs running, how many nodes are in use, and how many jobs are waiting to start in the queue. It also displays the number of jobs that are being held, which can happen for a variety of reasons.

The information is intended to help Cheyenne users decide when to submit jobs and which queue to specify. Users can check on the status of their own individual jobs by running qstat -u $USER as detailed in Managing and monitoring jobs. The expanded section shows an example of qstat output and addresses what to do if a job is held.

Sep. 8, 2017

CISL’s documentation for Cheyenne supercomputer users now includes details regarding how to load Python modules and custom-built Python packages as well as how to set up and maintain a Python virtual environment.

Setting up a virtual environment is particularly important for those who want to avoid version conflicts when using CISL-built Python packages along with packages that they install or develop themselves. The step-by-step process is described here.

Sep. 7, 2017

Due to unexpected issues encountered in the early stages of the planned outage, the Laramie test cluster will be continue to be unavailable until Friday, September 15, to allow CISL staff to complete important hardware and system software upgrades. This extension is to ensure minimal downtime for Cheyenne in October when similar upgrades will be performed.

Users will not be able to log in to the system throughout the outage. Any jobs that were running on Laramie when the outage began were killed and will need to be resubmitted after the upgrades have been completed.

Users will be informed via the CISL Notifier service when the system is returned to service.