ATUQ Supercomputer
The ATUQ supercomputer currently has 1116 cores distributed across 39 computing nodes in 8 chassis and 3 rack-mountable servers connected through InfiniBand networks with speeds of 52 Gbps and 100 Gbps.
- 988 cores en slurm actualmente+atuq-7-1+atuq-4-10+atuq-2-8+atuq+nfs
- Out of production: 104 cores
- atuq-7-1: 48 cores
- atuq-7-2: 48 cores
- atuq-7-3: 64 cores
- atuq-4-10: 24 cores
- atuq-2-8: 32 cores
- Storage distribution:
- 7.7 TB users
- 3 TB scratch
- 10 TB users nfs
- 3 TB scratch nfs
- 11 TB apps nfs
- 10 TB backups nfs
- 7.4 TB share nfs
- Node atuq-6-8 increases its number of cores from 24 to 48 cores
Currently, the supercomputer has dozens of active users from universities, public institutions, and private organizations. They utilize the queuing system to access processing resources, including processor cores and memory, for the execution of their scheduled jobs.
Institutions using the HPC-Modemat service:
- Researchers from Escuela Politécnica Nacional,
- Escuela Superior Politécnica del Litoral ESPOL,
- Instituto Nacional de Meteorología e Hidrología INAMHI,
- Laboratorio Mixto Internacional GREATICE,
- Pontificia Universidad Católica del Ecuador,
- Universidad San Francisco de Quito,
- Universidad Técnica de Ambato,
- Facultad Latinoamericana de Ciencias Sociales FLACSO,
- Charles Darwin Foundation,
- Universidad Técnica de Manabí,
- National Institute of Statistics and Census INEC
The Modemat Data Center
The Modemat data center is a physical facility that houses the ATUQ Supercomputer, many computing equipment and servers, as well as storage systems and network equipment, all used for data processing in research projects.
The specialized data center was designed to ensure high performance, availability, and data security. It features:
- Precision air conditioning and humidity control system.
- Raised floor with ventilation.
- Humidity sensors, fire suppression system.
- UPS 40KVA (up to 100 KVA).
- Security cameras.
- Biometric access control.
Operating System
Linux OS for HPC (master-workers) with a kernel optimized for distributed and/or parallel computing and a pre-arranged execution environment (PXE).
File System
NFS (Network File System)
InfiniBand Network
OpenFabrics
InfiniBand MPI (Message Passing Interface)
Open MPI and MVAPICH libraries
Job Scheduler
SLURM (Used by 60% of the Top500 supercomputers)
Scalable Management Tools
SLURM, LDAP
Monitoring Tools
AKCP, GRAFANA, XDMOD
Scientific Computing Applications
- Matlab (License for up to 200 cores)
Development Tools
- C, C++, and Fortran compilers
- Intel Parallel Studio
- Python 2, 3 (interpreted language) with modules for graphics processing and machine learning (TensorFlow)
- Julia (general numerical computing)
- CUDA (GPU computing libraries)
- R (statistical software and data processing)
- FENICS (variational modeling solver software, e.g., fluid dynamics)
- Singularity
Meteorology
- WRF/WRFDA (Weather Research and Forecasting Model)
- COSMO (Consortium for Small-Scale Modeling)
Biomathematics
- RAxML (phylogenetic tree reconstruction software)
Containerization
- Docker