IDRIS tools and computing experience feedback

IDRIS has developed tools to help users employ the intensive calculation means in a maximally efficient way. Moreover, the user support team provides advanced support responding to specific problems of user calculation codes. In addition, research studies concerning technical issues are also conducted by IDRIS engineers in order to help the entire user community to better use the supercomputers.

Other relevant information may be found in the archive section of our website.

• idrmem : Memory consumption measurement tool for the Ada and Turing machines
• libhpcidris : Library for measuring execution time, the performance of compute codes and memory consumption on the Turing Blue Gene/Q

• MAR advanced support report : Optimisation of memory consumption and sequential performance and MPI of MAR (Modèle Atmosphérique Régional)
• RAMSES advanced support report : Improvement of parallel scalability and reduction of memory imprint of the RAMSES astrophysics compute code

• Introduction aux entrées/sorties parallèles (Introduction to parallel input/output)
• Introduction aux systèmes de fichiers parallèles (Introduction to parallel file systems)
• Expériences d'entrées/sorties sur architectures massivement parallèles (Input/output experiments on massively parallel architectures)

• Introduction to parallel filesystems
• Best practice hints for parallel IO and MPI-IO