1 Introduction
The FRB-CESAB server Rossinante is dedicated to medium performance scientific computing (see Table 1.1 for hardware specification). It runs under the Operating System (OS) Ubuntu Server 20.04 LTS (Focal Fossa).
When do you need to use Rossinante?
- You need to analyse large datasets (RAM operations), and/or,
- You need to repeat tasks many times (parallelization on CPU/GPU).
Hardware | Specifications |
---|---|
CPU | 80 threads (2 x Intel Xeon Gold 5218R) |
RAM | 384 GB (12 x Cells 32 Go RDIMM) |
GPU | NVIDIA Quadro RTX 6000 |
Internal storage | 6 TB (8 x 960 GB SSD SATA) |
External storage (Seagate) | 8 TB (1 x 8 To HDD USB 3.0 5900 tr/min) |
External storage (Cervantes) | 8 TB (1 x 8 To HDD USB 3.0 5900 tr/min) |
External storage (Sancho) | 8 TB (1 x 8 To HDD USB 3.0 5900 tr/min) |
You can run programs implemented in R, Python, Julia, Shell script, C, C++, etc. directly from a terminal or using Web clients like RStudio Server or Jupyter Notebook.
What are the available software?
R
(latest version)RStudio Server
Python
(latest version)pip3
andConda
– Python packages managersJupyter Notebook
Julia
LaTeX3
Pandoc
Quarto
git
Docker
And some useful utilities:
CUDA
– Deep LearningFFmpeg
– Transcoding multimedia filesImageMagick
– Image manipulation programPoppler
– PDF utility libraryGDAL
,GEOS
,PROJ
– Spatial librarieshtop
andbtop
– CPU, RAM, and processes monitoring toolnvtop
– NVIDIA GPU monitoring toolnano
andvim
– CLI text editorsscreen
andtmux
– Terminal multiplexerstree
– Recursive directory listing programcurl
andwget
– Download managerszip
andunzip
– ZIP files managers
Can you do everything you want on Rossinante?
No!
Rossinante has only one administrator. Regular user, like you, has limited permissions.
For instance, you have only access to a personal directory: /home/you/
. If you ask you can also have a personal space in one or more external drives (e.g. /media/cervantes/you
).
If you need to use a non-installed software, please contact the administrator. Note that each user has a personal R library in which he/she can install every R packages he/she wants (independently of other users). For Python, packages are shared among all users (except if you work with virtual environments [recommended]).