<div dir="ltr"><div dir="ltr"><div style="font-family:tahoma,sans-serif" class="gmail_default">
<p class="MsoNormal"><span style="font-size:11pt;font-family:"Calibri",sans-serif;color:rgb(31,73,125)">>This line is probably what is limiting you to around 40gb.</span></p>
<p class="MsoNormal"><span style="font-family:"Tahoma",sans-serif">>#SBATCH --mem=38GB</span></p>

</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">Yes. If I change that value, the "ulimit -v" also changes. See below</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">[shams@hpc ~]$ cat slurm_blast.sh | grep mem<br>#SBATCH --mem=50GB<br>[shams@hpc ~]$ cat my_blast.log<br>virtual memory Â  Â  Â  Â  Â (kbytes, -v) 57671680<br>/var/spool/slurmd/job00306/slurm_script: line 13: ulimit: virtual memory: cannot modify limit: Operation not permitted<br>virtual memory Â  Â  Â  Â  Â (kbytes, -v) 57671680<br>Error memory mapping:/home/shams/ncbi-blast-2.9.0+/bin/nr.69.psq openedFilesCount=168 threadID=0<br>Error: NCBI C++ Exception:</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">However, the solution is not to change that parameter. There are two issue with that:</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">1) --mem belongs to the physical memory which is requested by job and is later reserved for the job by slurm.</div><div class="gmail_default" style="font-family:tahoma,sans-serif">So, on a 64GB node, if a user requests --mem=50GB, actually no one else can run a job with 10GB memory need.</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">2) The virtual size of the program (according) to the top is about 140GB. So, if I set --mem=140GB, the job stuck in the queue because requested information is invalid (node has 64GB of memory).</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">I really think there is a problem with slurm but can not find the root of the problem. The slurm config parameters are</div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif">Configuration data as of 2020-01-28T08:04:55<br>AccountingStorageBackupHost = (null)<br>AccountingStorageEnforce = associations,limits,qos,safe,wckeys<br>AccountingStorageHost Â  = hpc<br>AccountingStorageLoc Â  Â = N/A<br>AccountingStoragePort Â  = 6819<br>AccountingStorageTRES Â  = cpu,mem,energy,node,billing,fs/disk,vmem,pages,gres/gpu<br>AccountingStorageType Â  = accounting_storage/slurmdbd<br>AccountingStorageUser Â  = N/A<br>AccountingStoreJobComment = Yes<br>AcctGatherEnergyType Â  Â = acct_gather_energy/none<br>AcctGatherFilesystemType = acct_gather_filesystem/none<br>AcctGatherInterconnectType = acct_gather_interconnect/none<br>AcctGatherNodeFreq Â  Â  Â = 0 sec<br>AcctGatherProfileType Â  = acct_gather_profile/none<br>AllowSpecResourcesUsage = 0<br>AuthAltTypes Â  Â  Â  Â  Â  Â = (null)<br>AuthInfo Â  Â  Â  Â  Â  Â  Â  Â = (null)<br>AuthType Â  Â  Â  Â  Â  Â  Â  Â = auth/munge<br>BatchStartTimeout Â  Â  Â  = 10 sec<br>BOOT_TIME Â  Â  Â  Â  Â  Â  Â  = 2020-01-27T09:53:58<br>BurstBufferType Â  Â  Â  Â  = (null)<br>CheckpointType Â  Â  Â  Â  Â = checkpoint/none<br>CliFilterPlugins Â  Â  Â  Â = (null)<br>ClusterName Â  Â  Â  Â  Â  Â  = jupiter<br>CommunicationParameters = (null)<br>CompleteWait Â  Â  Â  Â  Â  Â = 0 sec<br>CoreSpecPlugin Â  Â  Â  Â  Â = core_spec/none<br>CpuFreqDef Â  Â  Â  Â  Â  Â  Â = Unknown<br>CpuFreqGovernors Â  Â  Â  Â = Performance,OnDemand,UserSpace<br>CredType Â  Â  Â  Â  Â  Â  Â  Â = cred/munge<br>DebugFlags Â  Â  Â  Â  Â  Â  Â = Backfill,BackfillMap,NO_CONF_HASH,Priority<br>DefMemPerNode Â  Â  Â  Â  Â  = UNLIMITED<br>DisableRootJobs Â  Â  Â  Â  = No<br>EioTimeout Â  Â  Â  Â  Â  Â  Â = 60<br>EnforcePartLimits Â  Â  Â  = NO<br>Epilog Â  Â  Â  Â  Â  Â  Â  Â  Â = (null)<br>EpilogMsgTime Â  Â  Â  Â  Â  = 2000 usec<br>EpilogSlurmctld Â  Â  Â  Â  = (null)<br>ExtSensorsType Â  Â  Â  Â  Â = ext_sensors/none<br>ExtSensorsFreq Â  Â  Â  Â  Â = 0 sec<br>FairShareDampeningFactor = 5<br>FastSchedule Â  Â  Â  Â  Â  Â = 0<br>FederationParameters Â  Â = (null)<br>FirstJobId Â  Â  Â  Â  Â  Â  Â = 1<br>GetEnvTimeout Â  Â  Â  Â  Â  = 2 sec<br>GresTypes Â  Â  Â  Â  Â  Â  Â  = gpu<br>GpuFreqDef Â  Â  Â  Â  Â  Â  Â = high,memory=high<br>GroupUpdateForce Â  Â  Â  Â = 1<br>GroupUpdateTime Â  Â  Â  Â  = 600 sec<br>HASH_VAL Â  Â  Â  Â  Â  Â  Â  Â = Match<br>HealthCheckInterval Â  Â  = 0 sec<br>HealthCheckNodeState Â  Â = ANY<br>HealthCheckProgram Â  Â  Â = (null)<br>InactiveLimit Â  Â  Â  Â  Â  = 30 sec<br>JobAcctGatherFrequency Â = 30<br>JobAcctGatherType Â  Â  Â  = jobacct_gather/linux<br>JobAcctGatherParams Â  Â  = (null)<br>JobCheckpointDir Â  Â  Â  Â = /var/spool/slurm.checkpoint<br>JobCompHost Â  Â  Â  Â  Â  Â  = hpc<br>JobCompLoc Â  Â  Â  Â  Â  Â  Â = /var/log/slurm_jobcomp.log<br>JobCompPort Â  Â  Â  Â  Â  Â  = 0<br>JobCompType Â  Â  Â  Â  Â  Â  = jobcomp/none<br>JobCompUser Â  Â  Â  Â  Â  Â  = root<br>JobContainerType Â  Â  Â  Â = job_container/none<br>JobCredentialPrivateKey = (null)<br>JobCredentialPublicCertificate = (null)<br>JobDefaults Â  Â  Â  Â  Â  Â  = (null)<br>JobFileAppend Â  Â  Â  Â  Â  = 0<br>JobRequeue Â  Â  Â  Â  Â  Â  Â = 1<br>JobSubmitPlugins Â  Â  Â  Â = (null)<br>KeepAliveTime Â  Â  Â  Â  Â  = SYSTEM_DEFAULT<br>KillOnBadExit Â  Â  Â  Â  Â  = 0<br>KillWait Â  Â  Â  Â  Â  Â  Â  Â = 60 sec<br>LaunchParameters Â  Â  Â  Â = (null)<br>LaunchType Â  Â  Â  Â  Â  Â  Â = launch/slurm<br>Layouts Â  Â  Â  Â  Â  Â  Â  Â  =<br>Licenses Â  Â  Â  Â  Â  Â  Â  Â = (null)<br>LicensesUsed Â  Â  Â  Â  Â  Â = (null)<br>LogTimeFormat Â  Â  Â  Â  Â  = iso8601_ms<br>MailDomain Â  Â  Â  Â  Â  Â  Â = (null)<br>MailProg Â  Â  Â  Â  Â  Â  Â  Â = /bin/mail<br>MaxArraySize Â  Â  Â  Â  Â  Â = 1001<br>MaxJobCount Â  Â  Â  Â  Â  Â  = 10000<br>MaxJobId Â  Â  Â  Â  Â  Â  Â  Â = 67043328<br>MaxMemPerNode Â  Â  Â  Â  Â  = UNLIMITED<br>MaxStepCount Â  Â  Â  Â  Â  Â = 40000<br>MaxTasksPerNode Â  Â  Â  Â  = 512<br>MCSPlugin Â  Â  Â  Â  Â  Â  Â  = mcs/none<br>MCSParameters Â  Â  Â  Â  Â  = (null)<br>MessageTimeout Â  Â  Â  Â  Â = 10 sec<br>MinJobAge Â  Â  Â  Â  Â  Â  Â  = 300 sec<br>MpiDefault Â  Â  Â  Â  Â  Â  Â = none<br>MpiParams Â  Â  Â  Â  Â  Â  Â  = (null)<br>MsgAggregationParams Â  Â = (null)<br>NEXT_JOB_ID Â  Â  Â  Â  Â  Â  = 305<br>NodeFeaturesPlugins Â  Â  = (null)<br>OverTimeLimit Â  Â  Â  Â  Â  = 0 min<br>PluginDir Â  Â  Â  Â  Â  Â  Â  = /usr/lib64/slurm<br>PlugStackConfig Â  Â  Â  Â  = /etc/slurm/plugstack.conf<br>PowerParameters Â  Â  Â  Â  = (null)<br>PowerPlugin Â  Â  Â  Â  Â  Â  =<br>PreemptMode Â  Â  Â  Â  Â  Â  = OFF<br>PreemptType Â  Â  Â  Â  Â  Â  = preempt/none<br>PreemptExemptTime Â  Â  Â  = 00:00:00<br>PriorityParameters Â  Â  Â = (null)<br>PrioritySiteFactorParameters = (null)<br>PrioritySiteFactorPlugin = (null)<br>PriorityDecayHalfLife Â  = 14-00:00:00<br>PriorityCalcPeriod Â  Â  Â = 00:05:00<br>PriorityFavorSmall Â  Â  Â = No<br>PriorityFlags Â  Â  Â  Â  Â  =<br>PriorityMaxAge Â  Â  Â  Â  Â = 1-00:00:00<br>PriorityUsageResetPeriod = NONE<br>PriorityType Â  Â  Â  Â  Â  Â = priority/multifactor<br>PriorityWeightAge Â  Â  Â  = 10<br>PriorityWeightAssoc Â  Â  = 0<br>PriorityWeightFairShare = 10000<br>PriorityWeightJobSize Â  = 100<br>PriorityWeightPartition = 10000<br>PriorityWeightQOS Â  Â  Â  = 0<br>PriorityWeightTRES Â  Â  Â = cpu=2000,mem=1,gres/gpu=400<br>PrivateData Â  Â  Â  Â  Â  Â  = none<br>ProctrackType Â  Â  Â  Â  Â  = proctrack/linuxproc<br>Prolog Â  Â  Â  Â  Â  Â  Â  Â  Â = (null)<br>PrologEpilogTimeout Â  Â  = 65534<br>PrologSlurmctld Â  Â  Â  Â  = (null)<br>PrologFlags Â  Â  Â  Â  Â  Â  = (null)<br>PropagatePrioProcess Â  Â = 0<br>PropagateResourceLimits = ALL<br>PropagateResourceLimitsExcept = (null)<br>RebootProgram Â  Â  Â  Â  Â  = (null)<br>ReconfigFlags Â  Â  Â  Â  Â  = (null)<br>RequeueExit Â  Â  Â  Â  Â  Â  = (null)<br>RequeueExitHold Â  Â  Â  Â  = (null)<br>ResumeFailProgram Â  Â  Â  = (null)<br>ResumeProgram Â  Â  Â  Â  Â  = /etc/slurm/resumehost.sh<br>ResumeRate Â  Â  Â  Â  Â  Â  Â = 4 nodes/min<br>ResumeTimeout Â  Â  Â  Â  Â  = 450 sec<br>ResvEpilog Â  Â  Â  Â  Â  Â  Â = (null)<br>ResvOverRun Â  Â  Â  Â  Â  Â  = 0 min<br>ResvProlog Â  Â  Â  Â  Â  Â  Â = (null)<br>ReturnToService Â  Â  Â  Â  = 2<br>RoutePlugin Â  Â  Â  Â  Â  Â  = route/default<br>SallocDefaultCommand Â  Â = (null)<br>SbcastParameters Â  Â  Â  Â = (null)<br>SchedulerParameters Â  Â  = (null)<br>SchedulerTimeSlice Â  Â  Â = 30 sec<br>SchedulerType Â  Â  Â  Â  Â  = sched/backfill<br>SelectType Â  Â  Â  Â  Â  Â  Â = select/cons_res<br>SelectTypeParameters Â  Â = CR_CORE_MEMORY<br>SlurmUser Â  Â  Â  Â  Â  Â  Â  = root(0)<br>SlurmctldAddr Â  Â  Â  Â  Â  = (null)<br>SlurmctldDebug Â  Â  Â  Â  Â = info<br>SlurmctldHost[0] Â  Â  Â  Â = hpc(10.1.1.1)<br>SlurmctldLogFile Â  Â  Â  Â = /var/log/slurm/slurmctld.log<br>SlurmctldPort Â  Â  Â  Â  Â  = 6817<br>SlurmctldSyslogDebug Â  Â = unknown<br>SlurmctldPrimaryOffProg = (null)<br>SlurmctldPrimaryOnProg Â = (null)<br>SlurmctldTimeout Â  Â  Â  Â = 300 sec<br>SlurmctldParameters Â  Â  = (null)<br>SlurmdDebug Â  Â  Â  Â  Â  Â  = info<br>SlurmdLogFile Â  Â  Â  Â  Â  = /var/log/slurm/slurmd.log<br>SlurmdParameters Â  Â  Â  Â = (null)<br>SlurmdPidFile Â  Â  Â  Â  Â  = /var/run/slurmd.pid<br>SlurmdPort Â  Â  Â  Â  Â  Â  Â = 6818<br>SlurmdSpoolDir Â  Â  Â  Â  Â = /var/spool/slurmd<br>SlurmdSyslogDebug Â  Â  Â  = unknown<br>SlurmdTimeout Â  Â  Â  Â  Â  = 300 sec<br>SlurmdUser Â  Â  Â  Â  Â  Â  Â = root(0)<br>SlurmSchedLogFile Â  Â  Â  = (null)<br>SlurmSchedLogLevel Â  Â  Â = 0<br>SlurmctldPidFile Â  Â  Â  Â = /var/run/slurmctld.pid<br>SlurmctldPlugstack Â  Â  Â = (null)<br>SLURM_CONF Â  Â  Â  Â  Â  Â  Â = /etc/slurm/slurm.conf<br>SLURM_VERSION Â  Â  Â  Â  Â  = 19.05.2<br>SrunEpilog Â  Â  Â  Â  Â  Â  Â = (null)<br>SrunPortRange Â  Â  Â  Â  Â  = 0-0<br>SrunProlog Â  Â  Â  Â  Â  Â  Â = (null)<br>StateSaveLocation Â  Â  Â  = /var/spool/slurm.state<br>SuspendExcNodes Â  Â  Â  Â  = (null)<br>SuspendExcParts Â  Â  Â  Â  = (null)<br>SuspendProgram Â  Â  Â  Â  Â = /etc/slurm/suspendhost.sh<br>SuspendRate Â  Â  Â  Â  Â  Â  = 4 nodes/min<br>SuspendTime Â  Â  Â  Â  Â  Â  = NONE<br>SuspendTimeout Â  Â  Â  Â  Â = 45 sec<br>SwitchType Â  Â  Â  Â  Â  Â  Â = switch/none<br>TaskEpilog Â  Â  Â  Â  Â  Â  Â = (null)<br>TaskPlugin Â  Â  Â  Â  Â  Â  Â = task/affinity<br>TaskPluginParam Â  Â  Â  Â  = (null type)<br>TaskProlog Â  Â  Â  Â  Â  Â  Â = (null)<br>TCPTimeout Â  Â  Â  Â  Â  Â  Â = 2 sec<br>TmpFS Â  Â  Â  Â  Â  Â  Â  Â  Â  = /state/partition1<br>TopologyParam Â  Â  Â  Â  Â  = (null)<br>TopologyPlugin Â  Â  Â  Â  Â = topology/none<br>TrackWCKey Â  Â  Â  Â  Â  Â  Â = Yes<br>TreeWidth Â  Â  Â  Â  Â  Â  Â  = 50<br>UsePam Â  Â  Â  Â  Â  Â  Â  Â  Â = 0<br>UnkillableStepProgram Â  = (null)<br>UnkillableStepTimeout Â  = 60 sec<br>VSizeFactor Â  Â  Â  Â  Â  Â  = 110 percent<br>WaitTime Â  Â  Â  Â  Â  Â  Â  Â = 60 sec<br>X11Parameters Â  Â  Â  Â  Â  = (null)<br></div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div class="gmail_default" style="font-family:tahoma,sans-serif"><br></div><div><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><font face="tahoma,sans-serif">Regards,<br>Mahmood</font><br><br><br></div></div></div><br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div lang="EN-US"><div class="gmail-m_4357538815788791602WordSection1"><div>
</div>
</div>
</div>

</blockquote></div></div>