- Last Day of Winter Holiday Shutdown Today; Consulting & Account Support Resume Tomorrow
- Training on Using Perlmutter, January 5-7
- (NEW) BerkeleyGW Events, Week of January 10-14
- Allocation Year 2021 to 2022 Transition on Wednesday, January 19
- New Default Python module at AY transition; All Python modules will support conda activate
- NERSC Federated Identity Pilot Has Begun!
- Please participate in the NERSC Annual User Survey
- Applications for DOE Computational Science Graduate Fellowship Open
- NVIDIA HPC SDK Training, January 12-13
- IDEAS-ECP Webinar on "Lessons Learned & Possibilities for Using the 'Wrong' Programming Approach on Leadership Computing Facility Systems," January 12
January 2022
Su Mo Tu We Th Fr Sa
1
2 *3* 4 *5---6---7* 8 3 Jan Last Day of Winter Holiday [1]
5-7 Jan Using Perlmutter Training [2]
9 *10-*11**12*-13--14* 15 10-14 Jan BerkeleyGW Events [3]
11 Jan Perlmutter Maintenance [4]
12 Jan CSGF Applications Due [5]
12 Jan ECP Monthly Webinar [6]
12-13 Jan NVIDIA HPC SDK Training [7]
16 *17* 18 *19* 20 21 22 17 Jan Martin Luther King Holiday [8]
19 Jan First Day of AY22 [9]
23 24 *25* 26 27 28 29 25 Jan Perlmutter Maintenance [4]
30 31
February 2022
Su Mo Tu We Th Fr Sa
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 *21* 22 23 24 25 26 21 Feb Presidents Day Holiday [10]
27 28
March 2022
Su Mo Tu We Th Fr Sa
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31
- December 24, 2021 - January 3, 2022: Winter Shutdown (Limited Consulting and Account Support)
- January 5-7, 2022: Using Perlmutter Training
- January 10-12 & 13-14, 2022: BerkeleyGW Tutorial Workshop & BESC2022
- January 11 & 25, 2022: Perlmutter Maintenance
- January 12, 2022: Computational Science Graduate Fellowship Applications Due
- January 12, 2022: IDEAS-ECP Monthly Webinar
- January 12-13, 2022: NVIDIA HPC SDK Training
- January 17, 2022: Martin Luther King Jr. Holiday (No Consulting or Account Support)
- January 19, 2022: First day of Allocation Year 2022
- February 21, 2022: Presidents Day Holiday (No Consulting or Account Support)
- All times are Pacific Time zone
-
Upcoming Planned Outage Dates (see Outages section for more details)
- Wednesday, January 5: HPSS Archive (User)
-
Other Significant Dates
- May 30, 2022: Memorial Day Holiday (No Consulting or Account Support)
- June 20, 2022: Juneteenth Holiday (No Consulting or Account Support)
- July 4, 2022: Independence Day Holiday (No Consulting or Account Support)
- September 5, 2022: Labor Day Holiday (No Consulting or Account Support)
- November 24-25, 2022: Thanksgiving Holiday (No Consulting or Account Support)
- December 23, 2022-January 2, 2023: Winter Shutdown (Limited Consulting and Account Support)
NERSC wishes you a Happy New Year! We are looking forward to working with you again this year and enabling the many scientific breakthroughs that arise from the work of our users!
Berkeley Lab, where NERSC is located, is operating under public health restrictions. NERSC continues to remain open while following site-specific protection plans. We remain in operation, with the majority of NERSC staff working remotely, and staff essential to operations onsite. We do not expect significant changes to our operations in the next few months.
You can continue to expect regular online consulting and account support as well as schedulable online appointments. Trainings continue to be held online. Regular maintenances on the systems continue to be performed while minimizing onsite staff presence, which could result in longer downtimes than would occur under normal circumstances.
Because onsite staffing is so minimal, we request that you continue to refrain from calling NERSC Operations except to report urgent system issues.
For current NERSC systems status, please see the online MOTD and current known issues webpages.
Today (Monday, January 3) is the last day of the Berkeley Lab Winter Holiday shutdown period. No consulting or account support is available today.
Regular consulting and account support will resume tomorrow.
NERSC and HPE staff will hold a three-day training event on using Perlmutter this Wednesday through Friday, January 5-7. This event is a continuation and extension of the Introduction to Perlmutter held in June, and is focused on using Perlmutter, including hands-on exercises.
Day 1 will begin with a brief recap and updated overview of the Perlmutter hardware and programming environment, followed by building and running applications on Perlmutter. Day 2 will introduce the HPE profiling and debugging tools, which are primarily CPU focused. The last day will include GPU basics, an overview of GPU programming models, and using Jupyter, Python, and Machine Learning / Deep Learning on Perlmutter. Each day includes hands-on exercises.
For more information and to register, please see https://www.nersc.gov/users/training/events/using-perlmutter-training-jan2022/.
BerkeleyGW users and other interested parties are invited to attend the 8th BerkeleyGW Tutorial Workshop and the 3rd Berkeley Excited States Conference, beginning January 10.
The 8th Annual BerkeleyGW Tutorial workshop will be held online January 10-12. The workshop will feature hands-on user sessions on the GW and GW-BSE approaches using the BerkeleyGW package. The target participants for the workshop include graduate students, postdocs, and researchers interested in learning about or sharpening their skills on ab initio calculations of many-electron effects in excited-state properties of condensed matter.
On January 13-14, the two-day annual Berkeley Excited States Conference 2022 will also be held online. BESC2022 will be a general topical conference featuring invited talks by experts on recent progress in the theory and applications of ab initio study of quantum many-body effects as well as experiments in excited-state phenomena in materials, showcasing forefront research involving advanced many-body approaches, novel experiments, and new science in the excited state.
For more information and to register please see https://workshop.berkeleygw.org. There are no fees to register but space is limited for the hands-on tutorial sessions.
The initial phase of the Perlmutter supercomputer is in the NERSC machine room, booting and running successfully.
We have added many early users onto the machine. We hope to add even more users soon. Anyone interested in using Perlmutter may apply using the Perlmutter Access Request Form.
The second phase of the machine, consisting of CPU-only nodes, will arrive early this year. After all the new nodes arrive, all of Perlmutter will be taken out of service and integrated over a period that we anticipate could take up to 8 weeks.
This newsletter item will be updated each week with the latest Perlmutter status.
To help ready your account for Perlmutter, please review your dotfiles. The same
home file system is mounted across all NERSC systems, so your
.bashrc/.cshrc/etc. files (dotfiles) need to work on all systems. The
NERSC_HOST variable can help you distinguish between systems and to set
customizations for each system. The NERSC_HOST variable is set automatically to
"perlmutter" on Perlmutter and to "cori" on Cori.
Some users may have older dot files that are setting the NERSC_HOST variable without first checking to see whether it already has a value, which will cause problems on Perlmutter. Please ensure that this is not the case in your dotfiles. Feel free to reach out to NERSC consulting with any questions or issues.
The current Allocation Year (AY21) ends at 7 am Pacific Time on Wednesday, January 19, 2022.
PIs for new and continuing projects will receive notification of the status of their project request later this week.
There are a few changes for AY22. Of note:
- Charge factors have been recalibrated to the new Perlmutter machine. The new currency, "CPU Node Hours", are 1/400th the size of the former "NERSC Hours" currency. Using one hour of time on a Cori Haswell node will cost 0.34 CPU Node Hours, and one hour of time on a Cori KNL will cost 0.2 CPU Node Hours. When we begin charging for Perlmutter (which will likely happen for the second half of the year), the charge for a single Perlmutter CPU node for one hour will be 1.0.
- Two separate allocations for CPU and GPU architectures. Projects will have allocations on CPU and GPU architectures, which are separate and cannot be traded or exchanged.
- New default Python Module at AY Transition. Please see the entry on this topic below.
- Cori defaults will change in March, 2022. Cori defaults will remain the same at the AY transition, but the machine will undergo what we hope will be its final major system upgrade during the March maintenance, at which time the default user environment will be updated.
For more information, please see the Allocation Year Transition page.
Python users take note: On Jan 19, 2022 at the Allocation Year rollover, NERSC will change our default Python and Python3 modules on Cori to python/3.9-anaconda-2021.11. Please note that older Python modules will remain available, but users must specify the full module name to continue to use them.
Updates in this module include:
- Python 3.9
- Support for conda activate
- Mamba 0.7.6 (a faster alternative to conda)
- netcdf4 1.5.8
- mpi4py 3.1.3
- authlib 0.15.4 (support for NERSC's Superfacility API)
This module is available via module load python/3.9-anaconda-2021.11, and
we encourage you to test it now.
At the AY transition, we will retroactively change the behavior of all
Cori Python modules to support conda activate. Please see
these pending updates to our documentation
for more details. As always, if you have a question, please contact us via
our helpdesk.
Note for pip users: pip packages installed via --user will be
installed at $HOME/.local/cori/3.9-anaconda-2021.11 (defined by
$PYTHONUSERBASE).
Note that the python/3.9-anaconda-2021.11 module is already the default on
Perlmutter and conda activate functionality is already supported there. There
are no scheduled Python module changes on Perlmutter.
Berkeley Lab staff can now follow a one-time process to link their Lab identity to their NERSC identity, then subsequently use their Lab credentials to log into resources such as Iris, ServiceNow, and the NERSC web site.
We anticipate that soon, more than two-thirds of our users will be able to use their institutional login credentials to log into these NERSC services.
The appearance of the NERSC login page for these services has now changed: instead of the form requesting your login name and password you will see a menu where you can choose the institution to use for login.
During this first phase, if you are not Berkeley Lab staff, simply select "NERSC" as the authentication source, and you will be sent to the familiar NERSC authentication form. If you are Lab staff, we encourage you to select the "Berkeley Lab" option and try it out!
NERSC has engaged a professional survey company, the National Business Research
Institute (NBRI), to conduct our annual user survey. Users can search for
a reminder email from [email protected] in their inboxes
with a personalized link to the user survey.
We value your response to the survey, which helps inform future plans for improvements to benefit our users. Please take the survey to let us know what we've done well and how we can better serve you!
Are you a US citizen or lawful permanent resident planning to embark on your first or second year of PhD study in physical, engineering, computer, mathematical, or life sciences at an accredited US college or university in the fall of 2022? If so, you may be eligible to apply for the Department of Energy's Computational Science Graduate Fellowship. Benefits include a yearly stipend of $38,000 and payment of full tuition and fees during the up to 4 years of total support, a 12-week practicum experience at a DOE national laboratory, access to DOE supercomputers, and more!
For more information, please see the CSGF Webpage. Applications are due next Wednesday, January 12.
NVIDIA will present a two-day training for users about the GPU programming models supported by NVIDIA's HPC SDK compilers, including Standard Language Acceleration and Libraries, OpenACC, OpenMP offload, and CUDA. The NVIDIA compilers are the default, recommended compiling environment for Perlmutter. Basic GPU architecture and profiling tools will also be presented. Attendees will have the opportunity to do hands-on and homework exercises on the GPUs on Perlmutter.
For more information and to register, please see https://www.nersc.gov/users/training/events/nvidia-hpcsdk-training-jan2022/.
IDEAS-ECP Webinar on "Lessons Learned & Possibilities for Using the 'Wrong' Programming Approach on Leadership Computing Facility Systems," January 12
The next webinar in the Best Practices for HPC Software Developers series is entitled "Wrong Way: Lessons Learned and Possibilities for Using the 'Wrong' Programming Approach on Leadership Computing Facility Systems," and will take place Wednesday, January 12, at 10:00 am Pacific time.
In this webinar, Philip C. Roth (Oak Ridge National Laboratory) will discuss the impact of using a "wrong" programming approach for a given system, in the context of the accelerated large-scale computing systems deployed and being deployed at US Department of Energy Computing facilities. In this webinar, he will present a few of these "wrong" programming approaches, and lessons learned in cases where the approach has been attempted.
There is no cost to attend, but registration is required. Please register here.
NERSC currently has several openings for postdocs, system administrators, and more! If you are looking for new opportunities, please consider the following openings:
- Scientific Data Architect: Collaborate with scientists to meet their Data, AI, and Analytics needs on NERSC supercomputers.
- Exascale Computing Postdoctoral Fellow: Collaborate with ECP math library and scientific application teams to enable the solution of deep, meaningful problems targeted by the ECP program and other DOE/Office of Science program areas.
- Data & Analytics Team Group Lead: Provide vision and guidance and lead a team that provides data management, analytics and AI software, support, and expertise to NERSC users.
- Cyber Security Engineer: Help protect NERSC from malicious and unauthorized activity.
- Machine Learning Engineer: Apply machine learning and AI to NERSC systems to improve on their ability to deliver productive science output.
- HPC Performance Engineer: Join a multidisciplinary team of computational and domain scientists to speed up scientific codes on cutting-edge computing architectures.
- Software Integration Engineer: Develop and support software integration with Continuous Integration in collaboration with ECP.
(Note: You can browse all our job openings by first navigating to https://jobs.lbl.gov/jobs/search/. Under "Business," select "View More" and scroll down to find and select the checkbox for "NE-NERSC".)
We know that NERSC users can make great NERSC employees! We look forward to seeing your application.
- Cori
- 01/19/22 07:00-20:00 PST, Scheduled Maintenance
- Systems will be unavailable during the Allocation Year transition
- 01/19/22 07:00-20:00 PST, Scheduled Maintenance
- Perlmutter
- 01/11/22 07:00-20:00 PST, Scheduled Maintenance
- 01/25/22 07:00-20:00 PST, Scheduled Maintenance
- HPSS Archive (User)
- 01/05/22 09:00-13:00 PST, Scheduled Maintenance
- 01/12/22 09:00-13:00 PST, Scheduled Maintenance
- HPSS Regent (Backup)
- 01/12/22 09:00-13:00 PST, Scheduled Maintenance
- Iris
- 01/19/22 07:00-09:30 PST, Scheduled Maintenance
- Systems will be unavailable during the Allocation Year transition
- 01/19/22 07:00-09:30 PST, Scheduled Maintenance
Visit http://my.nersc.gov/ for latest status and outage information.
You are receiving this email because you are the owner of an active account at NERSC. This mailing list is automatically populated with the email addresses associated with active NERSC accounts. In order to remove yourself from this mailing list, you must close your account, which can be done by emailing [email protected] with your request.