For user news, see the news section of

December 2018, New hardware has arrived

Our first round of new hardware has arrived.  It includes approximately 1000 new CPUs across 48 new CPU nodes, each with 192GB memory.  We also are receiving 40 NVidia Tesla V100 GPU cards for use in artificial intelligence research.  Various other servers allow us to continue to grow and expand our user base.  This is added to, not replaces, our existing cluster.  This hardware is currently being configured and will be available in early 2019.  Should you be interested in beta testing, please see our issue tracker.

Autumn 2018, Academy funding renewed: FGCI2 to begin

Earlier this year, the Academy of Finland has renewed funding of the Finnish Grid and Cloud Infrastructure for another two years.  Aalto’s new compute and GPU servers will be ordered in the spring/summer arrive in the autumn with a continuing emphasis on data science and deep learning.  We would like to remind users that our procurement can be used to directly purchase server-class dedicated hardware, either as part of Triton or standalone usage, and invite Aalto units to take this opportunity to join the Science-IT project.

Summer 2018, CodeRefinery partnership

is a cross-Nordic partnership to provider higher-level training in scientific programming.  After two successful CodeRefinery workshops in December and May, we have begun a partnership to host two CodeRefinery workshops at Aalto every year, coordinated with our HPC Kickstart courses.  For announcements, see the Scientific Computing in Practice announcement pages.

May 2018, Store data securely with Science-IT

As you have undoubtably heard, the GDPR is coming into effect and researchers will need to store data securely.  Science-IT helps to provide storage to the CS, NBE, and PHYS departments which satisfies these needs.  In addition, Triton is suitable for confidential data.  Together, these form a natural hierarchy (department storage for backed up data, Triton for large active data).  Science-IT is working with Aalto ITS to bring similar quality and simplicity of data storage to all Aalto as well.  We also provides extensive information on data storage at Aalto and in the world, and you can refer to the Aalto list for secure research data storage as well.

June 2017, New Nvidia P100 GPU capacity

We have procured five new GPU servers containing 4 Nvidia P100 GPUs each, which will be installed over the summer.  These provide solid base for deep learning and artificial intelligence research until our next round of procurement in 2018.

Summer 2017, New Aalto computing user guide launched,

The site has been launched.  This replaces the previous closed, wiki-based Triton User guide as well as the CS, (and somewhat NBE and PHYS) department IT instructions.  It also adds sections on Data Management and the general Aalto IT environment, so that researchers will have a one-stop site which can support all of their computational needs.  The site is open-source (CC-BY-4.0) and hosted on Github, so that users may contribute and our other partners may benefit from our work.

June 2016, Cloud capacity

As part of the current Finnish Grid and Cloud Infrastructure procurement, we are investing part of our funding into CSC’s cPouta cloud service.  While managed by CSC’s Cloud team, this ensures that Aalto researchers will always have the necessary CSC billing units to do their work, without an additional application process.  The exact procedure to request this capacity is not currently known, please contact us for more information.

Spring 2016, Academy of Finland funding renewed: FGCI

The Academy of Finland has chosen to fund the Finnish Grid and Cloud Infrastrucre (FGCI), the successor to the Finnish Grid Infrastructure (FGI).  The new project has essentially the same focus and same partners, but the new name reflects an increased emphasis in data science and cloud computing type applications.  A new cluster will be created with the new hardware, data will be migrated, and then old hardware will be added back to our new cluster.  We expect at least 100 new compute servers, including GPU machines.

December 2015, New 2PB Lustre filesystem

We have procured a new 2PB Lustre filesystem from DDN.  This filesystem has 10x the IOPS performance and 4x the capacity of our previous system, and will provide a solid base without capacity restrictions for large-scale data science for the next five years.  Integration and migration will take place over the next six months in conjunction with the upcoming procurement.

June 9-11, 2014, Scientific computing in practice: kickstart 2014

Following success of the SCiP kickstart 2013 we go for the three days HPC maraton in June 2014. This time, in addition to Triton user tutorial, Grid computing and CSC resources, you get to participate into Python for scientific computing and GPU computing parts. Details at Scientific computing in practice: kickstart 2014Note: further training will be announced only on the Scientific Computing in Practice page.

Vuori nodes migrated to FGI / Aalto

Former CSC’s cluster Vuori has moved to FGI, from Feb 2014 operated by Science IT support team at Aalto. Extra nodes available to the end users through FGI or locally.

New HP SL230s G8 compute nodes

Triton has got a full rack of 48x HP SL230 G8 compute nodes. Two 10-cores Ivy Bridge Xeon E5 2680 v2 CPU cores each. 24 nodes have 256GB of memory and the rest 64GB. Available to the end users since Jan 2014.

Hardware upgrade

Computing capacity will be extended during the autumn of 2013. This operation will not affect day-to-day usage of the Triton cluster.

June 10-12, 2013, Lecture: Scientific computing in practice

Three days of non-stop computing course oriented on new comers to fields of scientific computing at Physics, Chemistry, Computer Science, Math or any else at Aalto University. Or those who needs an intensive intro into computational resources available to Aalto users. Covered topics: HPC basics, Triton practicalities, Condor, MPI programming, Matlab, Hadoop, GPU programming, CSC resources, Grid computing on FGI. See details at Scientific Computing in Practice: kickstart 2013