Documentation

There is a wiki available with technical information, a how-to section, user tips, etc.

• HPC userwiki

User support

Mailing lists

If you request an HPC account you are automatically added to these two mailing lists:

• hpc-announce: low bandwidth channel that is used for official announcements. Any outages or expected downtimes will be communicated through this list.
  
• hpc-users: a forum for the UGent HPC community discussions and communicating with other colleagues who use HPC resources. Users have the option to unsubscribe from the hpc-users list.

Optionally, UGent personnel without an HPC account who wish to be informed can be added to either or both mailing lists (self service - click links above).

Helpdesk

In case of questions or requests, you can contact the HPC team through the UGent ICT Helpdesk. The easiest and best way is to send an email to helpdesk@ugent.be and mention "HPC" in the subject. Other ways to contact the helpdesk:

• HPC@UGent.be
• tel: 09 264 4747

Please mention your name, HPC account id (vsc40xxx), problem description and error message and output associated with the problem.

Training

• May 30, 2012 - June 6, 2012: Getting Started with High-Performance Computing (Doctoral School of Natural Sciences), prof. P. Dawyndt and K. Hoste
• SWPC2012 - Specialist Workshops in Parallel Computing 2012

• April 19-20, 2012: Multithreading and multiprocessing (Georg Hager and Jan Treibig)
• April 27, 2012: GPGPU: Considerations for parallelizing code with CUDA (Carsten Griwods)
• May 4, 2012: Message Passing Interface (MPI) (Jan Fostier)
• May 11, 2012: MapReduce and Hadoop (Robin Aly)

Specialized R&D support

Research groups which have special needs can contact us. Our team is specialized in HPC application, parallel programming, parallel debugging and Linux support.

• Performance improvement of existing applications;
• Algorithm development and innovation;
• Parallel application support - multi-core or multi-node;
• Research-driven procurement;
• Code porting, i.e. transfer of HPC codes to our infrastructure;
• One-to-one guidance (best practices, specialization, ...).