Triton cluster

Triton is the Aalto high-performance computing cluster. It serves all researchers of Aalto, but is currently coordinated from within the School of Science. Access is free for researchers (students not doing research should check out our intro for students). It is similar to the CSC clusters, though CSC clusters are larger and Triton is easier to use because it is more integrated into the Aalto environment.

Quick contents and links

Triton contents About Triton Cluster overview Usage policies and legal Acknowledging Triton Getting Help/Contact Triton issue tracker (most requests here, login with HAKA) Suggestions for good support requests Quick Reference Tutorials (start here) Triton accounts About Science-IT and Triton Connecting to Triton Applications Software modules Data storage Interactive jobs Serial Jobs Array jobs Job dependencies Monitoring job progress and job efficiency GPU computing Parallel computing Cluster usage details Parallel jobs (coming, for now see Running programs on Triton) GPU Computing Applications For full contents, see below.

News Shortcuts Issue tracker Quick Reference Triton Cheatsheet Triton FAQ Scicomp Garage Scientific computing resources SCIP – Scientific Computing in Practice courses: organized by SciComp. Including Triton kickstarts and many others Parallel computing Aalto IT Services for Research Hands-on Scientific Computing: map of important computing skills Software Carpentry (scientific computation basics) and Code Refinery (more focused on programming techniques) General links CSC - Finland’s academic computing center. FGCI user’s guide at CSC: That is a general Guide to FGCI resources. Triton is one of them. CSC HPC guides at CSC: a Triton like cluster at CSC. Similar setup, thus examples and instructions can be useful. Aalto research data management information

Overview

• Cluster overview
• Getting Triton help
• Triton accounts
• Usage policies and legal
• Acknowledging Triton

Tutorials

These are designed to be read (or watched) in-order by every Triton user when they get their accounts (except maybe the last ones). In order to use Triton well, in the Hands-on SciComp roadmap you should also know the Basics (A) and Linux (C) levels as a prerequisite.

• About these tutorials
• About Science-IT and Triton
• Connecting to Triton
• Applications
• Software modules
• Data storage
• Interactive jobs: running your first command
• Serial jobs: running in the queue
• Monitoring job progress and job efficiency
• Array jobs
• GPU computing
• Parallel computing
• Job dependencies

Detailed instructions

• Available compilers
• Compiling CUDA code while using conda environments
• Debugging
• Nvidia DGX machines
• Frequently asked questions
• Running programs on Triton
• GPU Computing
• Grid computing
• Grid computing 2
• Monitoring jobs
• Monitoring jobs’ usage of the filesytem
• Libraries
• Storage: local drives
• Storage: Lustre (scratch)
• MPI on Triton
• Profiling
• Quotas
• Singularity Containers
• Small files
• Triton ssh key fingerprints
• Storage
• Compilers and toolchains
• Remote workflows at Aalto

Applications

See our general information and the full list below:

• Applications: General info
• FHI-aims
• Armadillo
• Boost
• COMSOL Multiphysics
• Deep learning software
• Detectron
• Fenics
• FMRIprep
• Freesurfer
• FSL
• GPAW
• Gurobi Optimizer
• Julia
• Jupyter
• Keras
• Lammps
• Using Mathematica on Triton
• Matlab
• MLPack
• MNE
• NVIDIA’s singularity containers
• Octave
• OpenFOAM (with ParaView)
• OpenPose
• Paraview
• Python
• PyTorch
• R
• RStan
• RStudio
• Siesta & Transiesta
• Your own notebooks via sjupyter
• Spyder
• Tensorflow
• Theano
• VASP
• VisIT

Reference and Examples

• Triton quick reference
• Examples