Daily Bulletin Archive

Jan. 29, 2018

Users’ Cheyenne jobs sometimes fail because not enough usable memory is available to run them. Determining how much memory a program needs to run is a key step in avoiding this problem, so CISL provides the peak_memusage tool to help. 

See Checking memory use for how to use peak_memusage. Also see this page for how to specify your memory requirements in a job script.

Jan. 29, 2018

CISL staff will perform maintenance on the Geyser and Caldera clusters today, January 29, from approximately 10 a.m. until 2 p.m. MST to update the systems’ directory information services configurations.

No interruptions to Geyser, Caldera, Cheyenne, or the GLADE file system are expected throughout the maintenance period.  Please contact cislhelp@ucar.edu if you experience any disruptions.

Jan. 25, 2018

Registration is now open for an NCAR/CISL series of four one-day workshops on Modern Fortran beginning Tuesday, February 6.

Dan Nagle, CISL Consulting Services Group software engineer and a member of the U.S. Fortran Standards Technical Committee, will provide the training at the NCAR Mesa Lab’s Fleischmann Building in Boulder.

Participants are encouraged to bring their own laptop computers with recent releases of gfortran, mpich, and opencoarrays. Each workshop will begin at 9 a.m. and end at 4 p.m. with an hour break at noon.

  • Scalar Fortran - February 6: Scope, definition, scalar declarations and usage, and interacting with the processor.

  • Vector Fortran - February 20: Arrays, storage order, elemental operations, and array intrinsics.

  • Object-Oriented Fortran - February 27: Derived types, defined operations, defined assignment, and inheritance.

  • Parallel Fortran - March 6: Coarray concepts, declarations, and usage; synchronization and treating failed images.

Use this form to register to attend one or more workshops. The workshops will not be webcast or recorded.

Jan. 22, 2018

The OU Supercomputing Center for Education & Research is offering a free “Supercomputing in Plain English (SiPE)” workshop in weekly sessions beginning January 23 and running through May 1. Participants can attend live in person on the University of Oklahoma campus in Norman or via videoconference. Each hour-long session begins at 11:30 a.m. Mountain time.

Workshop sessions focus on fundamental High Performance Computing (HPC) issues and topics, including: overview of HPC; the storage hierarchy; instruction-level parallelism; high performance compilers; shared memory parallelism (e.g., OpenMP); distributed parallelism (e.g., MPI); HPC application types and parallel paradigms; multicore optimization; high-throughput computing; accelerator computing (e.g., GPUs); scientific and I/O libraries; and scientific visualization. Slides from previous workshops are available here.

Use this form to register for the semester.

SiPE is targeted at an audience of computer scientists and other scientists and engineers, including undergraduates, graduate students, postdocs, faculty, staff, and professionals. Participants should have some recent programming experience or one semester of coursework in Fortan, C, C++, or Java.

Jan. 22, 2018

The Geyser and Caldera clusters will be unavailable Monday, January 22, from approximately 10 a.m. until 2 p.m. to allow CISL staff to perform updates on the systems’ networking configurations.

A system reservation will be in place to prevent batch jobs from executing after 10 a.m. Jobs that are running when the outage begins will be killed. No impact to other NCAR HPC systems or the GLADE file system is expected during the outage.

CISL will inform users through the Notifier service when the system is restored. We apologize for any inconvenience this will cause and thank you for your patience.

Jan. 22, 2018

The Cheyenne, Wyoming, area experienced widespread power outages on Sunday from heavy snow and wind. Following several hours of issues related to power stability, the Cheyenne HPC system was powered down Sunday night to protect the hardware.

As of 9 a.m. MST today, power is being restored and CISL staff expect to have the Cheyenne system back up by 2 p.m. CISL will inform users through the Notifier service when the system is restored.

Jan. 19, 2018

The CISL Help Desk and Consulting Office will be closed Friday from 11:30 am - 2:00 pm for a USS function. For assistance with any urgent issues, call 303-497-2400 or email cislhelp@ucar.edu both of which are always monitored.

Jan. 19, 2018

CISL has added a new script for Cheyenne users that simplifies the launching of resource-intensive compilation jobs on the system’s batch nodes. Running qcmd as shown in this new documentation starts a non-interactive job on a single batch node in the Cheyenne "regular" queue with a default wall-clock time of 1 hour. It is recommended for running resource-intensive tasks such as CESM and WRF builds or any compiles with three or more threads.

Reminder: Use of the Cheyenne login nodes, as noted here, is restricted to running processes that do not consume excessive resources. This is to ensure an appropriate balance between user convenience and login node performance. Users are encouraged to compile on the Cheyenne batch nodes or on the Geyser or Caldera clusters, depending on where they want to run their programs.

Jan. 16, 2018

The Geyser and Caldera clusters will be unavailable from 8 a.m. to 8 p.m. MST on Tuesday, January 16, to allow CISL staff to update hardware and software components.

Users will be unable to log in to Geyser and Caldera during the outage. Reservations will be put in place on batch jobs before the outage begins. After the reservations are activated, jobs that are submitted with specified job times that overlap the reservation period will remain on hold until the systems are restored to service. All processes and running jobs that have not finished when the outage begins will be killed and will need to be resubmitted after the systems are returned to service.

Users will be informed via the CISL Notifier service when the systems are returned to production.

Jan. 16, 2018

Geyser_Caldera downtime: Tuesday, January 16 from 8:00 am - 8:00 pm MST

No downtime: Cheyenne, GLADE and HPSS