Metafor

ULiege - Aerospace & Mechanical Engineering

User Tools

Site Tools


Action disabled: media
devel:slurm

Mémo SLURM

sinfo

similaire à qstat -f de SGE

  • format orienté “queues”
rboman@hmem00 ~]$ sinfo
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
High         up 21-00:00:0      2    mix hmem[01-02]
Middle       up 5-00:00:00      7    mix hmem[03-09]
Low*         up 5-00:00:00     15    mix hmem[03-17]
Fast         up 1-00:00:00      3   idle hmem[18-20]

L'asterisque montre la queue par défaut

  • format orienté “noeuds”
[rboman@hmem00 ~]$ sinfo -Nl
Thu Jan 16 16:46:16 2014
NODELIST        NODES PARTITION       STATE CPUS    S:C:T MEMORY TMP_DISK WEIGHT FEATURES REASON
hmem[01-02]         2      High       mixed   48   4:12:1 512000        0      3   (null) none
hmem[03-09]         7    Middle       mixed   48   4:12:1 256000        0      2   (null) none
hmem[03-15,17]     14      Low*       mixed   48   4:12:1 128000        0     1+   (null) none
hmem16              1      Low*  completing   48   4:12:1 128000        0      1   (null) none
hmem[18-20]         3      Fast        idle    8    4:2:1 128000        0      1   (null) none

squeue

similaire à qstat -f -u '*' de SGE

[rboman@hmem00 ~]$ squeue |more
           JOBID PARTITION     NAME     USER  ST       TIME  NODES NODELIST(REASON)
          236664      High      bse waroquie  PD       0:00      1 (Resources)
          236587      High Blapidar  tlecocq  PD       0:00      1 (Priority)
          236588      High Blapidar  tlecocq  PD       0:00      1 (Priority)
          236589      High Blapidar  tlecocq  PD       0:00      1 (Priority)
...

colonne ST (state) : R=running, PD=pending

squeue --user tlecocq          # montre les jobs de "tlecocq"

(voir aussi –partition, –state, etc.)

scancel

similaire à qdel de SGE

sprio

info sur la priorité des jobs

sbatch: (qsub)

[rboman@hmem00 job1]$ sbatch submit.sh
Submitted batch job 236986
[rboman@hmem00 job1]$ squeue --user rboman
             JOBID PARTITION     NAME     USER  ST       TIME  NODES NODELIST(REASON)
            236986       Low robotest   rboman   R       0:04      1 hmem14
[rboman@hmem00 job1]$ sstat 236986
       JobID  MaxVMSize  MaxVMSizeNode  MaxVMSizeTask  AveVMSize     MaxRSS MaxRSSNode MaxRSSTask     AveRSS   MaxPages MaxPagesNode   MaxPagesTask   AvePages     MinCPU MinCPUNode MinCPUTask     AveCPU   NTasks AveCPUFreq   ReqCPUFreq ConsumedEnergy  MaxDiskRead MaxDiskReadNode MaxDiskReadTask  AveDiskRead MaxDiskWrite   MaxDiskWriteNode MaxDiskWriteTask AveDiskWrite
------------ ---------- -------------- -------------- ---------- ---------- ---------- ---------- ----------   -------- ------------ -------------- ---------- ---------- ---------- ---------- ---------- -------- ----------   ---------- -------------- ------------ --------------- --------------- ------------ ------------ --------------  -- ---------------- ------------
236986.1        202348K         hmem14              0    105032K      1276K     hmem14          0       608K          0       hmem14              0          0  00:00.000     hmem14          0  00:00.000        1      2.20G                           0        0.05M          hmem14               0        0.05M            0           hmem14                0              0
devel/slurm.txt · Last modified: 2016/03/30 15:23 by 127.0.0.1

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki