Fat nodes to Bianca

The SNIC board has approved about 10 MSEK in funding to procure more hardware for Bianca, including more GPU nodes, more storage, and more semi-fat compute nodes. With this increase in capacity and lifetime, we will be able to serve our growing community for several more years.

Additionally, an investment by the Kamerlin group has resulted in 144 nodes from the old NGI cluster, Irma, now added to Rackham. These resources, with 16 cores and 256 GB of RAM per node, amounting to about 1.5 million core-hours per month, are already available in the "core" partition (though the Kamerlin group naturally has priority over others). Look for nodes named "r1000" and higher!

During the coming year, Bianca will get four really high-mem nodes, 48 cores each with 768GB RAM. We will also augment the number of 256GB nodes considerably, at least 50 more of those, for a total of more than 260 nodes compared to 204 today. This will also mean that we can give more memory (and possibly cores) to each login node.

UPPMAX has laid an order for 3.75 PB of effective storage that will be added to the Crex storage system in the Spring of next year. This will bring the total volume of storage available on Rackham and Snowy to about 9.75 TB. Part of this space is reserved for system software and central databases, and available quota for projects is expected to increase by over 50%. This should be sufficient until the planned end-of-life of Rackham and Crex at the end of 2024. 

Until the extension is in place, space on Crex is very tight. Please continue to delete unnecessary files, be patient with performance degradations, and please wait until Spring before generating significant amounts of new data.

The SNIC Board has approved funding that will let UPPMAX not only continue to run Rackham and Bianca through the end of 2024, but also make some necessary investments to improve the user experience. Details are still being hammered out, but these investments will take place in early 2022, and will ensure that the user community can continue to grow and evolve. 

UPPMAX is happy to announce that Bianca now offers GPUs! The GPUs that were provided by SNIC are available for all users. GPUs are a type of accelerator that may significantly increase performance for certain workloads, for instance, workloads that contain large amounts of identical operations on independent pieces of data.

Today, the 20 millionth job on Rackham was submitted. 

(This article is about a news article in Swedish)

Ny Teknik har intervjuat en av våra anställda i en artikel om effekterna av metoo-uppropet. För drygt tre år sedan synliggjorde metoo-uppropet #teknisktfel sexuella trakasserier inom teknikbranschen. Sedan dess har framsteg gjorts menar initiativtagarna. Men mycket återstår att förbättra på arbetsplatserna.

Läs artikeln här: https://www.nyteknik.se/ingenjorskarriar/tre-ar-efter-uppropet-teknisktfel-kvinnor-som-ar-utsatta-blir-trodda-nu-7011093

The Swedish Research Council is conducting a survey regarding current practises and available support functions for research data management. This work is part of a commissioned project to coordinate activities about Open Access data.

Answer the survey here: https://survey.vr.se/Survey/601

An estimated 40% of all publications reported to SciLifeLab under 2020 involved work done on UPPMAX resources. 

UPPMAX is gathering interest in user-funded GPU resources on Bianca. Users are invited to contact us.

The EuroCC National Competence Center Sweden (ENCCS) is an initiative to establish a support organisation for HPC competence in general and especially in large-scale applications running on EuroHPC systems (such as LUMI).

A new resource for Artificial Intelligence and Machine Learning AI/ML research has recently opened at C3SE, the national centre for scientific and technical computing at Chalmers.

Several large European HPC-sites are reporting issues with cybersecurity and several SNIC centres have recently restricted access to some of its services.

While the main function of UPPMAX is to provide a world-class digital research environment, as a tax-funded resource we can and should keep public welfare in mind. In the current crisis, HPC resources can play a crucial role in contributing to the understanding of the virus itself as well as mitigation strategies and their impact. 

The UPPMAX board has decided to offer users an opportunity to
co-finance a pool of Nvidia Tesla T4 cards with 16 GB RAM each, at a price of
roughly 20 000 SEK/card. Read on if your research group may be interested in buying access to GPU capability at UPPMAX.

The 6PB large storage system, called Crex, is nearly fully allocated. This is the storage system attached to Rackham and Snowy. When we have allocated all the space available on it, we cannot both grant new projects that need disk space and prolong old ones with large allocations. Please read on to learn more about the situation.

Since this spring, we have been offering the Lutra offload storage system suitable for "cold" data which is not accessed or changed frequently, but which still has to be kept available on our clusters.

Due to the popularity of this service, the UPPMAX board wants to offer current
PIs at UPPMAX the possibility to sign up for additional storage. This time,
it will also be possible to buy storage suitable for sensitive personal data. 

Read the full article for more details

Rackham just passed 10 million completed jobs. See the picture below

Seminar by Philipp Schlatter, the chair of the SNIC Allocations Committee (SNAC), "SNIC and SNAC: Handling of large-scale computer time allocations in Sweden", to be held on in the afternoon on September 25th.

As of now, CONDA should work (module load conda) again, but please report anything weird (For CONDA, being slow is not weird).

Due to some changes and external alterations, the entire CONDA installation is in need of being re-installed. This might cause some lag and even some failures in the near future. Hopefully we will have the systems and repositories up and running as soon as possible.

A EuroHPC Joint Undertaking, involving Finland, Belgium, Czech Republic, Denmark, Norway, Poland, Sweden and Switzerland, will be placing a large HPC system at CSC in Finland. Scientists in member countries will be able to use the machine starting Q4 2020.

The SciLifeLab Compute and Storage platform, a part of NBIS and hosted by UPPMAX, is looking for a person with experience working with cloud and/or container  technology in order to support life science researchers. See below for a link to the job posting. Deadline is June 7th! 

For users who have purchased offload or glacial storage on our new system Lutra you may now start sending in your proposals from https://supr.snic.se.

UPPMAX application experts want to share their skills in a 4-day series of courses. We will help you move from being a Linux novice to an UPPMAX expert. If you already have the fundamentals down, you are still sure to enjoy the tips and tricks in the later parts of the course. 

Dates: August 26-29.

A research group used UPPMAX resources in a recent publication that has caused some popular interest. Around 4500 BCE, megalithic funerary monuments began to be built in northwestern Europe. The group's genomic investigation, published in the Proceedings of the National Academy of Sciences (PNAS), provides insight into the people associated with this long-standing megalith funerary tradition, including their social dynamics.

UPPMAX is looking for a new technical system administrator. If you enjoy working with Linux in a dynamic environment and contribute to a world-class research environment, please follow the url below.

Per Stenberg, a researcher at Umeå University and FOI, has used Rackham in a pilot study that explored the spread of organisms that were incidentally caught in more than 15,000 air filters that were archived for radioactivity measurements.

UPPMAX application experts want to share their skills in a 4-day series of courses. We will help you move from being a Linux novice to an UPPMAX expert. If you already have the fundamentals down, you are still sure to enjoy the tips and tricks in the later parts of the course. 

Dates: January 21-24.

If you are writing code that is used in research, then this course is for you

Hands-on informal and interactive three-day event with type-along type of presentations, live coding and demos. Short tutorials alternate with practical exercises.

December 3-5, 2018, 9:00 am to 5:00 pm

Welcome to a short hands on introduction to SNIC Science Cloud from the “beginner’s user/developer” point of view.

An application oriented introduction to cloud computing aimed mainly at PhD students. Basics of Service Oriented Architectures (SOA). Basic concepts of cloud computing such as virtualization and the service layers IaaS, PaaS and SaaS, dynamic provisioning, elasticity. Practical use of available public and private cloud stacks. Introduction to cloud security. Task based programming in cloud environments. Application of key technologies and design patterns to realize scientific applications in cloud computing environments.

The Fall 2018 call for proposals for large allocations of computing time on the High Performance Computing resources of the Swedish National Infrastructure for Computing (SNIC) is open. Deadline 19 October 2018.

The allocations are for the period 1 January 2019 - 31 December 2019.
 

This year's Scientific Visualisation course will be held from 26th to 30th of November in Uppsala, with specially invited guest-lecturer, Alex Telea, who is the author of the course book "Data Visualization: Principles and Practice" on which the course is based itself.

With the retirement of Pica, a new storage system for home folders is being taken into production this week. This system is named Domus, which means "home" in Latin. In addition to home directories, the software library will also be located on Domus.