Hp XC System 2.x Software Manuel d'utilisateur Page 76

  • Télécharger
  • Ajouter à mon manuel
  • Imprimer
  • Page
    / 154
  • Table des matières
  • MARQUE LIVRES
  • Noté. / 5. Basé sur avis des utilisateurs
Vue de la page 75
Each partition’s node limits sup ersede those specified by -N. Jobs that request more nodes than
the partition allows n ever leave the PENDING state. To use a specific partition, use the srun
-p option. Com binations of -n and -N control how job processes are distrib uted among nodes
according to the following srun policies:
-n/-N com binations srun infers your intended number of processes per nod e if you
specify both the number of processes and the number of nodes
for your job. Thus -n16 -N8 normally results in running 2
processes/node. But, see t he next policy for exceptions.
Minimum interpretation srun interprets all n ode requests as m inimu m node req uests (-N16
means "at least 16 nodes"). If some nodes lack enough CPUs to
cover the process count specified by -n, srun will automatically
allocate more nodes (than mentioned with -N) to meet the need. For
example, if not all nodes have 2 w orking CPUs, then -n32 -N16
together will allocate more than 16 nodes so that all processes are
supported. The actual number of nodes assigned (not the num ber
requested) is stored in environment variable SLURM_NNODES.
CPU overcommitment By default, srun never allocates more than one process per CPU. If
you intend to assign multiple processes per CPU, you must invoke
the srun -O option along with -n and -N.Thus,-n16 -N4 -O
together allow 2 processes per CPU on the 4 allocated 2-CPU nodes.
Inconsistent allocation
srun rejects as e rrors inconsistent -n/-N combinations. For
example, -n15 -N16 requests the impossible assignment of 15
processes to 16 nodes.
-c cpt (--cpus-per-task=cpt)
The -c cpt option assigns cpt CPUs p er process for this job ( defau lt is one CPU per process).
This o ption supports m ultith readed programs that requ ire more than a single CPU per process
for best performance.
For mu lti thr eaded programs where the density of CPUs is m ore important than a specific node
count, use both -n and -c on the srun execute line rather than -N. The options -n16 and
-c2 result in whatever node allocation is needed to yield the requested 2 CPUs/process. This is
the r everse of CPU overcommitm ent (see -N and -O options).
-p part (--partition=part)
The -p part option requests nodes only from the part partition. The default partition is
assigned by the system administrator.
-t minutes (--time=minutes)
The -t minutes option a llocates a total number o f minutes for this job to run (default is the
current partition’s time limit). If the number of minutes exceeds the partition’s time limit, then
the job never leaves the PENDING state. When the tim e limit has been reached, SLUR M
sends each job process SIGTERM followed (after a pause specified by SLURM’s KillWait
configuration parameter) by SIGKILL.
-T nthreads ( --threads=nthreads)
The -T nthreads option requests that srun allocate nthreads threads to initiate and
control the parallel tasks in this job. The default is the smaller of either 10 or th e number of
nodes actually allocated, SLURM_NNODES.
6-6 Using SLURM
Vue de la page 75
1 2 ... 71 72 73 74 75 76 77 78 79 80 81 ... 153 154

Commentaires sur ces manuels

Pas de commentaire