High Performance Computing - Services for Research
High Performance Computing
High performance computing (HPC) means the capacity to apply tens of thousands of processing units (PU) to a single problem. A HPC system can work on a problem in parallel and is more than a large pile of fast laptops. Without parallel processing, such simulations could take hundreds of years to perform. Typical use cases for HPC include simulations and data-analysis.
Simulations use model systems of e.g. proteins, atmosphere or galaxies and based on physics or mathematical approximations study how they behave. Accurate and realistic models can be very large and need long simulations, which require a lot of computing power.
Data-analysis typically applies different algorithms, like pattern matching or deep learning, to find correlations and understand large measurement data sets, like gene sequencing data or texts.
CSC's HPC infrastructure offers significant and versatile computing power, as well as a large selection of preinstalled software and storage. Depending on your use case, choose among our different services what suits you best. Using the HPC environments might require preliminary skills, for which we also offer training and support services.
Puhti caters to a wide range of use cases. It allows you to reserve computing and memory resources flexibly, and you can run anything from interactive single core data processing to medium scale simulations spanning multiple nodes.
Mahti is in particular geared towards medium to large scale simulations requiring Petaflops of compute power and capable interconnect. Mahti was opened into researchers' use in August 2020.
Allas is a general purpose data storage server that provides an environment for storing, sharing and analysing data. Allas can be accessed on the CSC servers as well as from anywhere on the internet.
Coming to use in 2021 is LUMI which will be one of the most competitive supercomputers in the world when it starts operating. LUMI is one of EuroHPC world-class supercomputers which is intended for research purposes.
Quantum compyters differ significantly from their digital counterparts when it comes to basic computational operators and operating principles. Kvasi Quantum Computing Simulator is intended for learning to use and develop new quantum algorithms in order to prepare for the upcoming era of quantum computing.
Grid computing enables applying the resources of many computers or computing clusters in a network to a single problem at the same time. A grid is a virtual working environment that allows deployment of distributed data and computing resources.