Différences

Ci-dessous, les différences entre deux révisions de la page.

Lien vers cette vue comparative

Les deux révisions précédentes Révision précédente
Prochaine révision
Révision précédente
quick_start [2016/11/30 13:47] – [User workspace] adminquick_start [2018/06/18 14:45] (Version actuelle) – admin
Ligne 2: Ligne 2:
  
 This document is intended for new users of the service. It gives information on such basics as how to log in, how to run application and where to get help. This document is intended for new users of the service. It gives information on such basics as how to log in, how to run application and where to get help.
 +
 +**Computation Workflow on the HPC**
 +
 +A typical Computation workflow will be:
 +  - Connect to the HPC
 +  - Transfer your files to the HPC
 +  - Compile your code and test it
 +  - Create a job script
 +  - Submit your job
 +  - Wait while
 +      * your job gets into the queue
 +      * your job gets executed  
 +      * your job finishes
 +  - Move your results
 +
  
 ====== Connecting to clusters ====== ====== Connecting to clusters ======
Ligne 7: Ligne 22:
  
  
-<note important>**Before starting using the cluster, remember that mesocentre is composed of three clusters:** +<note important>**Before starting using the cluster, remember that mesocentre is composed of one cluster and one shared multi-core machine:**
    
   -<color red> mesologin1.univ-fcomte.fr </color>    dedicated for sequential, array tasks, MPI and openMP programs (only with SGE)   -<color red> mesologin1.univ-fcomte.fr </color>    dedicated for sequential, array tasks, MPI and openMP programs (only with SGE)
Ligne 16: Ligne 30:
 </note> </note>
  
-Example for MPI usage :+Example, connect and change password:
  
 <code bash> <code bash>
Ligne 32: Ligne 46:
 The easiest way to connect to the cluster is by using SSH protocol. Windows users can use  **PuTTY**. The easiest way to connect to the cluster is by using SSH protocol. Windows users can use  **PuTTY**.
  
-However you can use [[nx_client|FreeNX client]], a powerful graphical user interface to run interactive applications like matlab on mesoshared machine. +However you can use [[x2go|X2GO client]], a powerful graphical user interface to run interactive applications like matlab on ''mesoshared.univ-fcomte.fr'' machine. 
  
  
Ligne 41: Ligne 55:
 Once connected each user will have a home directory organized as follow : Once connected each user will have a home directory organized as follow :
  
-  * ''HOME'': this is the home directory, total allowed size: 10G +  * ''HOME'': this is the home directory, total allowed size: **10GB** 
-  * ''WORK'': Computation and software must be **Started** from here, total allowed size: 600G+  * ''WORK'': Computation and software must be **Started** from here, total allowed size: **1TB**
  
 Read more [[environnement_travail|here]]  Read more [[environnement_travail|here]] 
Ligne 49: Ligne 63:
 <note>See the list of installed software and how to use them [[logiciels_installes|]]</note> <note>See the list of installed software and how to use them [[logiciels_installes|]]</note>
  
-Several software are installed on the clusters (mesocluster, mesoseq). You can show them by : +Several software are installed on the clusters. You can show them by : 
-<code> +<code bash> 
-$module avail+$ module avail
 </code> </code>
 This will show the list of software installed and their version. This will show the list of software installed and their version.
  
-the output, on mesocluster, looks like this :+the output looks like this :
  
 <code bash> <code bash>
Ligne 123: Ligne 137:
  
 <code bash> <code bash>
-#!/bin/bash -l+ 
 +#!/bin/bash
  
  
Ligne 152: Ligne 167:
 Where ''jobscript.sh'' is the script that you have prepared.\\ Where ''jobscript.sh'' is the script that you have prepared.\\
  
-These files will be in the directory you submitted the jobs from if you use the -cwd option, and in your home directory otherwise. The names are constructed from the name of the job script, ''.o'' for output or ''.e'' for errors, and the job number. So the command above might create files such as 'jobscript.sh.o12345' and 'jobscript.sh.e12345' where ''12345'' is the number that is assigned to your job when you submit it.+These files will be stored in the directory you submitted the jobs from if you use the -cwd option, and in your home directory otherwise. The names are constructed from the name of the job script, ''.o'' for output or ''.e'' for errors, and the job number. So the command above might create files such as 'jobscript.sh.o12345' and 'jobscript.sh.e12345' where ''12345'' is the number that is assigned to your job when you submit it.
 <note important> <note important>
 JOB must bhe launched from WORK directory.  JOB must bhe launched from WORK directory. 
Ligne 179: Ligne 194:
  
 <code bash> <code bash>
-#!/bin/bash -l+#!/bin/bash
  
  
Ligne 199: Ligne 214:
  
 <code bash> <code bash>
-#!/bin/bash -l+#!/bin/bash 
  
 #$ -N test_sge #$ -N test_sge
Ligne 214: Ligne 229:
  
 <code bash> <code bash>
-#!/bin/bash -l+#!/bin/bash 
  
 #$ -N test_sge #$ -N test_sge
Ligne 231: Ligne 246:
 ==== MPI script ==== ==== MPI script ====
 <note tip> <note tip>
-To use MPI, you need to specify that you wish to use that Parallel Environment (-pe impi_tight), and declare the number of slots</note>+To use MPI, you need to specify that you wish to use that Parallel Environment (-pe mpi), and declare the number of slots</note>
  
  
 <code bash> <code bash>
-#!/bin/bash -l+#!/bin/bash 
  
 #$ -N test_sge #$ -N test_sge
  • quick_start.1480513630.txt.gz
  • Dernière modification : 2016/11/30 13:47
  • de admin