Computing – MpdRoot

Computing TDR

Network and computing infrastructure for the NICA complex at JINR. Technical Design

LHEP cluster

NICA Cluster

How to use NICA cluster of LHEP

To obtain account for the NICA cluster it is necessary to fill the registration form.
Please follow the interactive guide on How To Use Clusters
The rules of the using the NICA cluster:
Now each user have new home-dir: /eos/lhep/username WITH QUOTA up to 50GB!
The old users can copy files from the old-home: /nica/user/*/username.
( Use command eoscp or xrdcp ).
For the data storage the disk space in directory /eos/nica is divided to
3th project:
BMN – /eos/nica/bmn,
MPD – /eos/nica/mpd,
SPD – /eos/nica/spd.
You should send a request to the project admin for receiving the place
into these directories of your project!

Recommendations for using the NICA cluster

Do not use /weekly for long-term storage of files – quickly transfer them to EOS!
Do not copy file from /weekly to /eos, use /tmp or /toEos on ncx104, /weekly->/toEos->/eos more than 10 times faster then /weekly->/eos
Run Your jobs from /tmp, and use /tmp as data output for small file (size of /tmp ~220 GB,) as data output for big files please use /weekly
Don’t forget clean working folders in /tmp and /weekly after using

HybriLIT & Govorun

The HybriLIT heterogeneous cluster is a computation component of the Multifunctional center for data storage processing and analysis of LIT JINR, which contains a multicore component and computation accelerators: NVIDIA graphic processors and Intel Xeon Phi coprocessors. The detailed information on the heterogeneous cluster, rules and registration form can be found on the JINR HybriLIT Web-site.

Please follow the interactive guide on How To Use Clusters
If you have time-consuming tasks, many simple tasks or a lot of files to process, you should use batch system SLURM of the HybriLIT heterogeneous cluster and GOVORUN. For an access to the GOVORUN supercomputer run the following command

module add GVR/v1.0-1

Simple example of user job for SLURM

#!/bin/sh
#SBATCH --ntasks-per-node=1
#SBATCH --mem-per-cpu=8192M  # that much memory for simulation and reconstruction jobs only
#SBATCH --time=3-0:0:0
#SBATCH -p cascade

source /cvmfs/nica.jinr.ru/sw/os/login.sh
module add mpdroot

root -l -q -b runMC.C\(\"Path to Your input file\",\"Path to Your output file\",0,1000,kFALSE,0\)

LIT Clusters

MICC

The Multifunctional Information and Computing Complex (MICC) has the main objective to meet the needs of the JINR scientific community to the maximum extent possible in order to solve urgent tasks, from theoretical research and experimental data processing, storage and analysis to the solution of applied tasks in the field of life sciences. The tasks of the NICA project, the neutrino programme, the tasks of processing data from the experiments at the LHC and other large-scale experiments, as well as support for users of the JINR Laboratories and its Member States will be the priorities.

Monitoring

JINR Central Information and Computer Complex (CICC),

To monitor the CICC infrastructure and its maintenance the system of centralized local monitoring “litmon” is used which allows to provide round the-clock control of all resources (Tier1, Tier2, Hybrilit, Cloud, dCache, UPS, etc.) timely notifies about failures and allows to conduct complex analysis of the complex.

Parallelization

NICA-Scheduler is a module package developed for the ROOT- and FairRoot-based frameworks, such as the BmnRoot and MpdRoot software. It uses an existing batch system (SLURM, SGE and Torque are supported) to distribute user jobs on clusters (as well as supports multithreading on local machines), and simplifies parallel job execution without knowledge of the batch system. NICA-Scheduler can be used as a submodule inside your project or can be compiled as a standalone package.