News

Hoffman2 Cluster Winter Maintenance December 20, 2024 🚧

Important

The Hoffman2 Cluster will be unavailable from 12:00 p.m. on Friday, December 20, 2024, through 1:00 a.m. on Saturday, December 21, 2024 for scheduled maintenance.

WHAT WE ARE DOING…

  • Updating the core research switch infrastructure

  • Migrating project directories to new storage

WHAT TO EXPECT…

Warning

  • Your job will not start if its requested runtime exceeds the number of hours remaining between now and 5:00 a.m. on 12/06/2024!!

  • Your files in $SCRATCH will be DELETED. Please back up any important files you have stored in $SCRATCH!

JOB SCHEDULER INFORMATION…

  • Starting Friday, December 6, we will hourly reduce the maximum job run time in order to completely drain the cluster’s running jobs by December 20. If your job requests more than the remaining available runtime, it will not start.

  • All remaining running/pending jobs will be terminated when maintenance starts. Affected users will have to resubmit their jobs after the outage. Please plan your runs accordingly.

WHERE TO GET HELP…

If you have any questions regarding this maintenance, please submit a ticket to our support site at:

https://support.idre.ucla.edu/helpdesk

Regular email notifications will be sent out prior to the downtime.


OARC/UCLA Workshop: GPU computing in AI and HPC with VAST Data & NVIDIA – REGISTRATION ✅

Registration for the workshop on GPU computing in Artificial Intelligence and High Performance Computing presented by OARC, in conjunction with VAST Data and NVIDIA, on October 15, 2024, is now open.

We invite you to register for the event and to consult the full agenda here:

https://events.vastdata.com/uclaaiworkshopvastnvidia/OARC

Please feel free to distribute this news to anyone in the UCLA community who may have interest in this event.

For further feedback or questions about this email, please open a support ticket at https://support.idre.ucla.edu.


OARC/UCLA Workshop: GPU computing in AI and HPC with VAST Data & NVIDIA, Pre-event Survey – PLEASE RESPOND ☑️

OARC’s Hoffman2 Cluster Team, in conjunction with VAST Data and NVIDIA, is organizing an in-person workshop on GPU computing in Artificial Intelligence and High Performance Computing on October 15, 2024.

We invite you to sign up for more information and to give feedback on this event by completing the following form:

https://forms.office.com/r/1ZMTSkAPK2

Please feel free to distribute this email to anyone in the UCLA community who may have interest in this event.

For further feedback or questions about this email, please open a support ticket at https://support.idre.ucla.edu.


Hoffman2 Cluster Summer 2024 maintenance outage

* WHAT’S NEW * During the outage, we upgraded storage server firmware and migrated numerous filesystems from an old storage server to our new all-flash VAST system. You should experience significantly improved performance on your home directory during periods of high cluster utilization.

We increased the user quota on $HOME from 40GB to 60GB (a 50% increase) and the file count from 500,000 to 1,000,000. We hope this benefits users of Anaconda, virtual environments, and other applications that utilize a significant amount of space in your home directory.

* NOTE * The contents of $SCRATCH have been erased as announced. We have reinstated 2TB quotas on $SCRATCH directories. If you need temporary additional space, please submit a formal request through the ticketing system.


Hoffman2 Cluster Spring 2024 maintenance outage

Attention

The Hoffman2 Cluster will be unavailable from 5:00 a.m. on Sunday, March 24, 2024 through 6:00 p.m. on Tuesday, March 26, 2024 for scheduled maintenance.

Warning

** IMPORTANT JOB SCHEDULER INFORMATION: Your job will not start if its requested runtime exceeds the number of hours remaining between now and 5:00 a.m. on 3/24/2024!!

WHAT WE ARE DOING…

  • Installing additional high-speed flash-based storage storage capacity to expand our scratch filesystem and provide higher performance for certain high-throughput computational workloads.

  • Updating network infrastructure

  • Deploying a new software image

WHAT TO EXPECT…

Attention

** IMPORTANT ** - You will not be able to log in, run jobs, check out licenses, or transfer files during the maintenance window. - Your files on $SCRATCH will be DELETED. Please back up any important files on $SCRATCH!

JOB SCHEDULER INFORMATION… - Starting Sunday, March 10, we will hourly reduce the maximum job run time in order to completely drain the cluster’s running jobs by March 24. If your job requests more than the remaining available runtime, it will not start. See How to modify any resource requested in a pending job to learn how to modify the time requested in already submitted jobs.

  • All remaining running/pending jobs will be killed when maintenance starts. Affected users will have to resubmit their jobs after the outage. Please plan your runs accordingly.

WHERE TO GET HELP…

If you have any questions regarding this maintenance, please submit a ticket to our support site at:

https://support.idre.ucla.edu/helpdesk

Regular email notifications will be sent out prior to the downtime.


Visual Studio Code Version Supported by the Hoffman2 Cluster

Attention

As of Janauary 2024, if you are using Visual Studio Code to connect to the cluster via the SSH Remote extension, the highest version of Code compatible with the Hoffman2 Cluster is: VSCode v. 1.85, available at:

https://code.visualstudio.com/updates/v1_85

If you have any questions or problems, please submit a ticket at our support site https://support.idre.ucla.edu/


Hoffman2 Cluster: Winter Campus Closure Information 🏂

Please be aware that the OARC Research Technology Group will have limited staffing during the Winter Campus Closure from 12/23/2023 through 1/2/2024.

We will make every effort to keep the Hoffman2 Cluster operational during the campus closure, within the constraints imposed by limited service, delivery, and staffing schedules. Non-urgent support tickets will be answered after the campus reopens on January 3, 2024.

If you have any questions or problems, please submit a ticket at our support site https://support.idre.ucla.edu/

Please enjoy your break and stay safe and healthy!


Gaussian version 16 revision C.02 and GaussView version 6 now available

Gaussian 16 Rev. C.02 and GuassView version 6 are now available on the Hoffman2 Cluster.

Gaussian provides state-of-the-art capabilities for electronic structure modeling

https://gaussian.com

To learn how to run Gaussian on the Hoffman2 Cluster please visit:

https://www.hoffman2.idre.ucla.edu/Using-H2/Software/Software.html#gaussian

To learn how to run GaussView on the Hoffman2 Cluster please visit:

https://www.hoffman2.idre.ucla.edu/Using-H2/Software/Software.html#gaussview

Documentation is available at:

https://gaussian.com/man/

If you have any questions or problems, please submit a ticket at our support site https://support.idre.ucla.edu/


MATLAB version R2023b now available

MATLAB version R2023b is now available on the Hoffman2 Cluster.

MATLAB is a programming and numeric computing platform, for more details see:

https://www.mathworks.com/products/matlab.html

To learn how to run MATLAB on the Hoffman2 Cluster please visit:

https://www.hoffman2.idre.ucla.edu/Using-H2/Software/Software.html#matlab

Documentation is available at: https://www.mathworks.com/help/?s_tid=srchtitle_site_search_1_documentation

If you have any questions or problems, please submit a ticket at our support site https://support.idre.ucla.edu/


Hoffman2 Summer Maintenance Announcement

Warning

The Hoffman2 Cluster will be unavailable from 5:00 a.m. on Monday, September 18, 2023 through 6:00 p.m. on Wednesday, September 20, 2023 for scheduled maintenance.

WHAT WE ARE DOING:

  • We will be relocating our scratch filesystem cluster and increasing its maximum throughput capabilities, migrating our core virtual machine infrastructure to faster hardware, and implementing various hardware and software updates.

WHAT TO EXPECT:

  • You will not be able to log in, run jobs, check out licenses, or transfer files during the maintenance window.

JOB SCHEDULER INFORMATION:

  • Starting Monday, September 4 2023, we will hourly reduce the maximum job run time in order to completely drain the cluster’s running jobs by September 18. If your job requests more than this maximum time, it will not start.

  • All remaining running/pending jobs will be killed when maintenance starts. Affected users will have to resubmit their jobs after the outage. Please plan your runs accordingly.

Regular email notifications will be sent out prior to the downtime.


Questions or comments? Visit our support online help desk at: https://support.idre.ucla.edu.


BRIEF Hoffman2 JOB HOLD on August 7-8, 2023

We will be holding queued jobs between 4:00 AM on Monday, August 7 and 4:00 PM on Tuesday, August 8, during an urgent replacement of the environmental management system in our POD Data Center.

During this maintenance, we need to minimize the heat load in the facility by reducing the number of concurrently running jobs.

You may continue to submit jobs to the queue during this period.

We will make every effort to begin dispatching queued jobs before the officially announced time if weather conditions are appropriate and cooling has stabilized.

We apologize for the brief inconvenience.

Regular email notifications will be sent out prior to the downtime.


Questions or comments? Visit our support online help desk at: https://support.idre.ucla.edu.


SIM, the account management application, back in service 🎉

After a brief outage SIM, the new Hoffman2 Cluster System Identity Manager, is back on line.

If you have not yet accessed it, please follow these instructions:

  1. If you already have an account on the cluster:

  • log into the Hoffman2 Cluster and run the following command at the shell prompt:

$ mylink

and follow the instructions that will be printed on the screen.

  1. If you do not have an account yet (or have forgotten your username and/or password):


Questions or comments? Visit our support online help desk at: https://support.idre.ucla.edu.


SIM, the account management application, currently unavailable ‼️

Warning

The System Identity Manager (SIM) is currently unavailable and is planned to be back on line on Thursday August 3, 2023.

Perspective new users and/or faculty sponsors are currently not be able to create an account.

Current users: password processes are currently unavailable.

Current faculty sponsors: no group/resources modification is available at this time.

We apologize for the inconvenience.


Questions or comments? Visit our support online help desk at: https://support.idre.ucla.edu.


SIM the new account management application is now live 🎉

The Office of Advanced Research Computing (OARC) is pleased to announce the new web-based account management application: System Identity Manager or SIM.

All current and perspective users and faculty sponsors of the Hoffman2 Cluster need to register into the new account management system. To do so:

  1. If you already have an account on the cluster:

  • log into the Hoffman2 Cluster and run the following command at the shell prompt:

$ mylink

and follow the instructions that will be printed on the screen.

  1. If you do not have an account yet (or have forgotten your username and/or password):


How to modify any resource requested in a pending job

In some cases you may wish to modify the resources requested in a currently pending job. For example, you would like to modify the value of the requested memory, the runtime, the number of slots, or the parallel environment.

In particular, if your pending job requests more time than is available before the scheduled maintenance window begins, it will not start. If your job can finish in less time, you could use the qalter command to change its time limit to less than the remaining time.

To learn how to alter a pending job issue:

$ modify_resources_of_pending_job $JOB_ID

where $JOB_ID is the job ID number of one of your pending jobs.

To see which jobs you have pending you can issue:

$ myjob -s p

Hoffman2 Cluster Winter 2022 maintenance outage

Attention

The Hoffman2 Cluster will be unavailable from 5:00AM on Monday, December 12, 2022 through 6:00PM on Tuesday, December 13, 2021 for scheduled maintenance.

Warning

IMPORTANT JOB SCHEDULER INFORMATION: Your job will not start if its requested runtime exceeds the number of hours remaining between now and 5:00 a.m. on 12/12/2022!!

What to expect

  • Starting Monday, November 28, we will hourly reduce the maximum job run time in order to completely drain the cluster’s running jobs by December 12. If your job requests more than this maximum time, it will not start.

  • All jobs (queued or running) at the time of this outage will be killed; affected users will have to resubmit their jobs after the outage. Please plan your runs accordingly.

  • During the maintenance window, you will not be able to log in, run jobs, check out licenses, or transfer files.

What we are doing

  • We will be upgrading software and firmware on our high performance storage, network fabric, and virtual machine infrastructure.

Where to get help

If you have any questions or problems regarding this maintenance, please submit a ticket at our support site.


Campus VPN required to reach Grid Identity Manager

As of December 2021, in order to access our web-based, account management system, the Grid Identity Manager (GIM), you will need to be connected to the campus VPN. Please note, campus VPN access is restricted to registered students and university employees with an active staff/faculty appointment. Here are instructions on how to install the campus VPN.

If you have any questions or problems, please submit a ticket at our support site.


Hoffman2 Fall 2021 maintenance outage

Attention

The Hoffman2 Cluster will be unavailable from 5:00AM on Monday, September 20, 2021 through 6:00PM on Wednesday, September 22, 2021 for scheduled maintenance.

What is changing

  • The operating system distribution on the cluster will be upgraded from CentOS 6 to CentOS 7.

  • Singularity will be available on the entire cluster.

  • Elliptical curve (ed25519) ssh keys will be available.

  • Home directory quotas will be doubled in size to 40GB and the maximum file count will be increased to 500,000 per user.

What to expect

  • Starting Monday, September 6, we will hourly reduce the maximum job run time in order to completely drain the cluster’s running jobs by September 20. If your job requests more than this maximum time, it will not start. All jobs (queued or running) at the time of this outage will be killed; affected users will have to resubmit their jobs after the outage. Please plan your runs accordingly.

  • During the maintenance window, you will not be able to log in, run jobs, check out licenses, or transfer files.

Warning

During the maintenance the contents of $SCRATCH will be completely erased. If you have files on $SCRATCH from a recently completed job, please copy them before the maintenace outage to persistent storage (your home or project directory).

  • After the maintenance, you may need to recompile your code (including R and Python libraries, etc.) so it can run on CentOS 7. Please refer to the section Code and libraries that might need recompilation.

  • Before the maintenance to reduce work disruption you may consider testing your jobs on the small number of nodes that are currently already running CentOS 7 and can be used to recompile your code and test its compatibility with the new distribution. Details can be found on the Code and libraries that might need recompilation section.

Code and libraries that might need recompilation

In going from CentOS 6 to CentOS 7 the version of the GNU C Library (glibc) and of the default GCC compiler are changing from: 2.12 to 2.17 and 4.4.7 to 4.8.5 respectively. Accordingly you can expect that any component of your code compiled on CentOS 6 may need to be recompiled to take full advantage of the new system libraries. This applies to any C, C++, FORTRAN code that you have compiled on the cluster or it may apply to any python and R libraries you may have installed in your $HOME or project directory.

We encourage our users to test their applications by requesting interactive sessions on the nodes already on the next version of the OS.

Note

When requesting interactive sessions on the nodes already on the next version of the OS any R library will be automatically installed in a new location in your $HOME. This will avoid any conflict between running R on the nodes on the current version of the OS and on nodes on the next version of the OS.

After the transition you might consider archiving the directory: $HOME/R/x86_64-pc-linux-gnu-library/ as all the needed installed R packages will be in your: $HOME/R/x86_64-pc-linux-gnu-library-RH7/.

Note

To prevent possible conflicts with python libraries, we encourage our users to set the $PYTHONUSERBASE environmental variable appropriately when installing libraries on nodes on the next version of the OS to minimize conflict with existing libraries installed in the user $HOME directory. A possible choice would be:

$ export PYTHONUSERBASE=$HOME/.local_rh7

You can set such environmental variable in your initialization files, by appending the following lines in your $HOME/.bashrc or $HOME/.cshrc:

test=$(grep "release 7" /etc/redhat-release)
if [ $? == 0 ]; then
    export PYTHONUSERBASE=$HOME/.local_rh7
fi

After the transition you might consider archiving the directory: $HOME/.local as all the needed installed python libraries will be in your: $HOME/.local_rh7.

Note

Software available in /u/local/apps has been recompiled or re-installed, should a code you might need not be available let us know by submitting a ticket at our support site.

What we are doing

We will be upgrading the cluster operating system from CentOS 6 to Centos 7, performing data center maintenance, upgrading the research network, migrating data to allow for expanded storage capacity, and more.

Where to get help

If you have any questions or problems regarding this maintenance, please submit a ticket at our support site.


Nodes available on the next version of the OS

A number of nodes are now available on the next version of the Operating System. Building and installing applications on these nodes will be generally easier and more expedite as the nodes have newer default system libraries than those present on the nodes running the current default version of the Operating System (e.g., the login nodes and a large part of the compute nodes).

We encourage our users to test this newer version of the Operating Systems as we will be transitioning to it soon. Because the default version of the GCC compiler, the system libraries and the version of the Intel compiler are newer on these nodes, your applications, python or R libraries may need to be recompiled/installed.

To access these nodes interactively please use:

$ qrsh -l rh7

you can customize the qrsh command by requesting resources such as runtime, memory, etc. To learn more please see the sections: Examples of how to request resources and: Examples of how to request multiple cores.

To access these nodes in batch jobs, please add the following line to the preamble of your job submission script:

#$ -l rh7

to learn how to write your batch job submission script please see the section: Use qsub with a submission script.

Note

The -l highp option cannot be used to submit jobs to nodes on the next version of the OS unless your group has transitioned part, or all of its nodes, to this version of the OS.

If you have any questions or problems, please submit a ticket at our support site.


GPU nodes available

A number of nodes with various types of GPU cards are available on the Hoffman2 Cluster. To learn how to access these nodes please see the sections: How to access GPU nodes and: GPU cards available on the Hoffman2 Cluster.

If you have any questions or problems, please submit a ticket at our support site.


Mathematica version 12.1

Mathematica Version 12.1 is now available on the Hoffman2 Cluster.

Mathematica is a modern technical computing system spanning most areas of technical computing.

To learn how to run Mathematica on Hoffman2 please refer to:

https://www.hoffman2.idre.ucla.edu/Using-H2/Software/Software.html#mathematica

If you have any questions or problems, please submit a new ticket at our support site https://support.idre.ucla.edu/


MATLAB version 9.8 (R2020a) Total-head-count-license

MATLAB Version 9.8 (R2020a) is now available on the Hoffman2 Cluster.

This installation of MATLAB under the Total Academic Headcount License (https://www.mathworks.com/academia/tah-portal/ucla-31454052.html) supports access to the full suite of packages and virtually unlimited licenses.

MATLAB is a proprietary multi-paradigm programming language and numerical computing environment developed by MathWorks.

To run MATLAB on Hoffman2 please refer to:

https://www.hoffman2.idre.ucla.edu/Using-H2/Software/Software.html#matlab

If you have any questions or problems, please submit a new ticket at our support site https://support.idre.ucla.edu/


Q-Chem version 5.3.0

Q-Chem Version 5.3.0 is now available on the Hoffman2 Cluster.

Q-Chem is a comprehensive ab initio quantum chemistry software for accurate predictions of molecular structures, reactivities, and vibrational, electronic and NMR spectra. The new release of Q-Chem 5 represents the state-of-the-art of methodology from the highest performance DFT/HF calculations to high level post-HF correlation methods (https://www.q-chem.com/).

To run Q-Chem on Hoffman2:

To run the shared memory version (from a bash script):

. /u/local/Modules/default/init/modules.sh module load qchem/5.3.0_sm qchem -nt $NSLOTS sample.in sample.out_$JOB_ID

To run the shared memory version (from a bash script):

. /u/local/Modules/default/init/modules.sh module load qchem/5.3.0_sm qchem -mpi -nt 1 -np $NSLOTS sample.in sample.out_$JOB_ID

where you will substitute sample.in and sample.out with the actual input and output file names.

If you have any questions or problems, please submit a new ticket at our support site https://support.idre.ucla.edu/