Table des matières

Below you can see a gallery that illustrate the computer room evolutions.

STAGE 1 - initial situation

This was the situation at my arrival, in December 2010. The computer room contained around 120 computing cores (with different generations of CPU technologies) and no real storage capacity.


The right enclosure was mainly dedicated for old and dysfunctional UPS, network switches and PVM. In addition, every machines with a yellow post-it on were non-functioning computing machines.

STAGE 2 - Situation before the computer room rebuilding

Before rebuilding the computer room, due to some problems of 3rd floor load resistance (more than 2.8T of computing materials in the room), we had slightly less than 1000 computing cores (868 in HPC clusters) and around 154TB raw for storage.

STAGE 3 - After the computer room rebuilding and till 2019

Before Lucifer cluster dismantling: 1828 CPU cores, ~108k GPU cores and around 750TB raw for storage (archive, home and workdir volumes).

In addition, in these new enclosures are hosted 3 virtualization clusters; one dedicated for LBT computation resources managment and the 2 others for Information Systems (one of which is shared with the IBPC).

As you can notice, everything is now black and white, proper and clean… ;-)

STAGE 4 - The current situation

This is the current situation.

After Lucifer and Hades clusters dismantling: 984 CPU cores, 285k CUDA cores, ~3.6TB RAM and around 720TB raw for storage (archive and workdir volumes).