CRAY XC40 Using the Batch System

Introduction

The only way to start a parallel job on the compute nodes of this system is to use the batch system. The installed batch system is based on

• the resource management system torque and
• the scheduler moab

Additional you have to know on CRAY XE6/XC30 the user applications are always launched on the compute nodes using the application launcher, aprun, which submits applications to the Application Level Placement Scheduler (ALPS) for placement and execution.

Detailed information for CRAY XE6 about how to use this system and many examples can be found in Cray Application Developer's Environment User's Guide and Workload Management and Application Placement for the Cray Linux Environment.

Detailed information for CRAY XC30 about how to use this system and many examples can be found in Cray Programming Environment User's Guide and Workload Management and Application Placement for the Cray Linux Environment.

• ALPS is always used for scheduling a job on the compute nodes. It does not care about the programming model you used. So we need a few general definitions :
  • PE : Processing Elements, basically an Unix ‘Process’, can be a MPI Task, CAF image, UPC tread, ...
  • numa_node The cores and memory on a node with ‘flat’ memory access, basically one of the 2 Dies on the Intel and the direct attach memory.
  • Thread A thread is contained inside a process. Multiple threads can exist within the same process and share resources such as memory, while different PEs do not share these resources. Most likely you will use OpenMP threads.

• aprun is the ALPS (Application Level Placement Scheduler) application launcher
  • It must be used to run application on the XE/XC compute nodes interactively and in a batch job
  • If aprun is not used, the application is launched on the MOM node (and will most likely fail)
  • aprun man page contains several useful examples at least 3 important parameter to control:
    • The total number of PEs: -n
    • The number of PEs per node: -N
    • The number of OpenMP threads: -d (the 'stride' between 2 PEs in a node
  • see also understanding aprun
• qsub is the torque submission command for batch job scripts.

Writing a submission script is typically the most convenient way to submit your job to the batch system. You generally interact with the batch system in two ways: through options specified in job submission scripts (these are detailed below in the examples) and by using torque or moab commands on the login nodes. There are three key commands used to interact with torque:

• qsub
• qstat
• qdel

Check the man page of torque for more advanced commands and options

 man pbs

Requesting Resources using batch system TORQUE and ALPS

Batch Mode

Production jobs are typically run in batch mode. Batch scripts are shell scripts containing flags and commands to be interpreted by a shell and are used to run a set of commands in sequence.

• The number of required nodes, cores, wall time and more can be determined by the parameters in the job script header with "#PBS" before any executable commands in the script.

#!/bin/bash
#PBS -N job_name
#PBS -l nodes=2:ppn=32
#PBS -l walltime=00:20:00             
  
# Change to the direcotry that the job was submitted from
cd $PBS_O_WORKDIR

# Launch the parallel job to the allocated compute nodes
aprun -n 64 -N 32 ./my_mpi_executable arg1 arg2 > my_output_file 2>&1

• The job is submitted by the qsub command (all script head parameters #PBS can also be submitted directly by qsub command options).

 qsub my_batchjob_script.pbs

• Overwriting qsub Options:

 qsub -N other_name -l nodes=2:ppn=32,walltime=00:20:00 my_batchjob_script.pbs

• The batch script is not necessarily granted resources immediately, it may sit in the queue of pending jobs for some time before its required resources become available.
• At the end of the execution output and error files are returned to submission directory
• This example will run your executable "my_mpi_executable" in parallel with 64 MPI processes. Torque will allocate 2 nodes to your job for a maximum time of 20 minutes and place 32 processes on each node (one per core). The batch systems allocates nodes exclusively only for one job. After the walltime limit is exceeded, the batch system will terminate your job. The job launcher for the XE6/XC30 parallel jobs (both MPI and OpenMP) is aprun. This needs to be started from a subdirectory of the /mnt/lustre_server (your workspace). The aprun example above will start the parallel executable "my_mpi_executable" with the arguments "arg1" and "arg2". The job will be started using 64 MPI processes with 32 processes placed on each of your allocated nodes (remember that a node consists of 32 cores in the XE6 system and only 16 cores in the XC30 system). You need to have nodes allocated by the batch system before (qsub).

To query further options of aprun, please use

 man aprun 
 aprun -h

Warning: You have to change into a subdirectory of /mnt/lustre_server (your workspace), before calling apron.

Interactive batch Mode

Interactive mode is typically used for debugging or optimizing code but not for running production code. To begin an interactive session, use the "qsub -I" command:

 qsub -I -l nodes=2:ppn=32,walltime=00:30:00

If the requested resources are available and free (in the example above: 2 nodes/32 cores, 30 minutes), then you will get a new session on the mom node for your requested resources. Now you have to use the aprun command to launch your application to the allocated compute nodes. When you are finished, enter logout to exit the batch system and return to the normal command line.

Notes

• Remember, you use aprun within the context of a batch session and the maximum size of the job is determined by the resources you requested when you launched the batch session. You cannot use the aprun command to use more resources than you reserved using the qsub command. Once a batch session begins, you can only use fewer resources than initially requested.
• While your job is running (in Batch Mode), STDOUT and STDERR are written to a file or files in a system directory and the output is copied to your submission directory only after the job completes. Specifying the "qsub -j oe" option here and redirecting the output to a file (see examples above) makes it possible for you to view STDOUT and STDERR while the job is running.

Run job on other Account ID

There are Unix groups associated to the project account ID (ACID). To run a job on a non-default project budget, the groupname of this project has to be passed in the group_list:

qsub -W group_list=<groupname> ...

To get your available groups:

 id 

Warning: It is not possible anymore to use your primary group as groupname!

Usage of a Reservation

For nodes which are reserved for special groups or users, you need to specify an additional option for this reservation:

E.g. a reservation named john.1 will be used with following command:

qsub -W x=FLAGS:ADVRES:john.1 ...

Deleting a Batch Job

 qdel <jobID>
 canceljob <jobID>

This commands enables you to remove jobs from the job queue. If the job is running, qdel will abort it. You can obtain the Job ID from the output of command "qstat" or you remember the output of your qsub command of your job.

Status Information

* Status of jobs:
 qstat
 qstat -a
 showq

• Status of Qeues:

 qstat -q
 qstat -Q
 

• Status of job scheduling

 checkjob <jobID>
 showstart <jobID>

• Status of backfill. This can help you to build small jobs that can be backfilled immediately while you are waiting for the resources to become available for your larger jobs

 showbf

• Status of Nodes/System (see also Gathering Application Status and Information on the Cray System)

 xtnodestat
 apstat

Note: for further details type on the login node:

 man qstat
 man apstat
 man xtnodestat
 showbf -h
 showq -h
 checkjob -h
 showstart -h

Limitations

• see the Batch System Layout and Limits for CRAY XE6
• see the Batch System Layout and Limits for CRAY XC30

Understanding aprun

Running Applications

Core specialization

System 'noise' on compute nodes may significantly degrade scalability for some applications. The Core Specialization can mitigate this problem.

• 1 core per node will be dedicated for system work (service core)
• As many system interrupts as possible will be forced to execute on the service core
• The application will not run on the service core

To get core specialization use aprun -r

 aprun -r1 -n 100 a.out

highest numbered cores will be used, starting with 31 on current nodes. (independent on aprun -j setting)

apcount provided to compute total number of cores required

 man apcount

Hyperthreading only for XC30 system !

Cray XC30 compute nodes are always booted with hyperthreading on ON. User can choose to run with one or two PEs or threads per core. The default is to run with 1. You can make your choice at runtime :

aprun –n### -j1 … -> Single Stream mode, one rank per core

aprun –n### -j2 … -> Dual Stream mode, two ranks per core

The numbering of the cores in single stream mode is 0-7 for die 0 and 8-15 for die 1. If using dual stream mode the numbering of the first 15 cores stays the same and cores 16-23 are on die 0 and 24-31 on die 1. Note that this make the numbering of the cores in hypterthread mode is not contigues :

Mode	cores on die 0	cores on die 1
Single Stream Dual Stream	0-7 8-15	0-7,16-23 8-15,24-31

aprun CPU Affinity control

CLE can dynamically distribute work by allowing PEs and threads to migrate from one CPU to another within a node. In some cases, moving PEs or threads from CPU to CPU increases cache and translation lookaside buffer (TLB) misses and therefore reduces performance. The CPU affinity options enable to bind a PE or thread to a particular CPU or a subset of CPUs on a node.

• aprun CPU affinity options (see also man aprun)
  • Default settings: -cc cpu (PEs are bound a to specific core, depended on the –d setting)
  • Binding PEs to a specific numa node : -cc numa_node (PEs are not bound to a specific core but cannot ‘leave’ their numa_node)
  • No binding: -cc none
  • Own binding: -cc 0,4,3,2,1,16,18,31,9,....

aprun Memory Affinity control

Cray XC30 systems use dual-socket compute nodes with 2 dies. For 16-CPU Cray XC30 compute node processors, NUMA nodes 0 and 1 have eight CPUs each (logical CPUs 0-7, 8-15 respectively). If your applications use Intel Hyperthreading Technology, it is possible to use up to 32 processing elements (logical CPUs 16-23 are on NUMA node 0 and CPUs 24-31 are on NUMA node 1). Even if you PE and threads are bound to a specific numa_node, the memory used does not have to be ‘local’

• aprun memory affinity options (see also man apron)
  • Suggested setting is –ss (a PE can only allocate the memory local to its assigned NUMA node. If this is not possible, your application will crash.)

Some basic aprun examples

Assuming a XC30 with Sandybridge nodes (32 cores per node with Hyperthreading)

Pure MPI application , using all the available cores in a node

  aprun -n 32 -j2 ./a.out

Pure MPI application, using only 1 core per node

32 MPI tasks, 32 nodes with 32*32 core allocated can be done to increase the available memory for the MPI tasks

 aprun -N 1 -n 32 -d 32 -j2 ./a.out

Hybrid MPI/OpenMP application, 4 MPI ranks per node

32 MPI tasks, 8 OpenMP threads each need to set OMP_NUM_THREADS

 export OMP_NUM_THREADS=8
 aprun -n 32 -N 4 -d $OMP_NUM_THREADS -j2

MPI and OpenMP with Intel PE

Intel RTE creates one extra thread when spawning the worker threads. This makes the pinning for aprun more difficult.

Suggestions:

• Running when “depth” divides evenly into the number of “cpus” on a socket

 export OMP_NUM_THREADS=“<=depth”
 aprun -n npes -d “depth” -cc numa_node a.out

• Running when “depth” does not divide evenly into the number of “cpus” on a socket

 export OMP_NUM_THREADS=“<=depth”
 aprun -n npes -d “depth” -cc none a.out

Multiple Program Multiple Data (MPMD)

aprun supports MPMD – Multiple Program Multiple Data.

• Launching several executables which all are part of the same MPI_COMM_WORLD

 aprun –n 128 exe1 : -n 64 exe2 : -n 64 exe3

• Notice : Each exacutable needs a dedicated node, exe1 and exe2 cannot share a node.

Example : The following commands needs 3 nodes

 aprun –n 1 exe1 : -n 1 exe2 : -n 1 exe3

• Use a script to start several serial jobs on a node :

 aprun –a xt –n 1 –d 32 –cc none script.sh

 >cat script.sh 
 ./exe1& 
 ./exe2& 
 ./exe3&
 wait 
 >

cpu_lists for each PE

CLE was updated to allow threads and processing elements to have more flexibility in placement. This is ideal for processor architectures whose cores share resources with which they may have to wait to utilize. Separating cpu_lists by colons (:) allows the user to specify the cores used by processing elements and their child processes or threads. Essentially, this provides the user more granularity to specify cpu_lists for each processing element.

Here an example with 3 threads :

 aprun -n 4 -N 4 -cc 1,3,5:7,9,11:13,15,17:19,21,23

Job Examples

Warning: Deprecated CRAY qsub syntax using mppwidth, mppnppn, mppdepth, feature, ...

The qsub arguments specially available for CRAY XE6 systems (mppwidth, mppnppn, mppdepth, feature) is deprecated in the new batch system version. Most functionalities of those qsub arguments are still available in this new batch system version. Nevertheless, we recommend not to use these qsub arguments anymore. Please use following syntax:

 qsub -l nodes=2:ppn=32:mem32gb <myjobscript>

(replaces: qsub -l mppwidth=64,mppnppn=32,feature=mem32gb)

• nodes: replacement for mppwidth/mppnppn
• ppn: replacement for mppnppn
• mem32gb: replacement for feature=mem32gb

Note: omitting argument ppn=32 will also allocate all 32 cores of a node

This batch script template serves as basis for the aprun expamples given later.

#! /bin/bash

#PBS -N <job_name>
#PBS -l nodes=<number_of_nodes>
#PBS -l walltime=00:01:00

cd $PBS_O_WORKDIR # This is the directory where this script and the executable are located. 
# You can choose any other directory on the lustre file system.

export OMP_NUM_THREADS=<nt>
<aprun_command>

The keywords <job_name>, <number_of_nodes>, <nt>, and <aprun_command> have to be replaced and the walltime adapted accordingly (one minute is given in the template above). The OMP_NUM_THREADS environment variable is only important for applications using OpenMP. Please note that OpenMP directives are recognized by default by the Cray compiler and can be turned off by the -hnoomp option. For the Intel, GNU, and PGI compiler one has to use the corresponding flag to enable OpenMP recognition.

The following parameters for the template above should cover the vast majority of applications and are given for both the XE6 and XC30 platform at HRLS. The <exe> keword should be replaced by your application.

XE6 Platform

The nodes of the XE6 features two Interlagos processors with 16 cores each resulting in a total of 32 cores per node. Each Interlagos processor forms two NUMA domains of size 8 resulting in totally four NUMA domains per node.

• Description: Serial application (no MPI or OpemMP)

 <number_of_nodes>:  1
 <aprun_command>:  aprun -n 1 <exe>

• Description: Pure OpenMP application (no MPI)

 <number_of_nodes>:  1
 <nt>: 32
 <aprun_command>:  aprun -n 1 -d $OMP_NUM_THREADS <exe>

Comment: You can vary the number of threads from 1-32.

• Description: Pure MPI application on two nodes fully packed (no OpenMP)

 <number_of_nodes>:  2
 <aprun_command>:  aprun -n 64 -N 32 <exe>

Comment: The -n specifies the total number of processing elements (PE) and -N the PEs per node. The -n has to be less or equal to 32*<number_of_nodes> and -N less or equal to <number_of_nodes>. Finally, the -n value divided by the -N value has to be less or equal than the <number_of_nodes>. You can increase the number of nodes as needed and vary the remaining parameters accoridingly.

• Description: Pure MPI application on two nodes in wide-AVX mode (no OpenMP)

 <number_of_nodes>:  2
 <aprun_command>:  aprun -n 32 -N 16 -d 2 <exe>

Comment: The -d 2 is used to place the PEs evenly among the cores on the node. This doubles the memory bandwidth and floating point unit per PE.

• Description: Mixed (Hybrid) MPI OpenMP application on two nodes

 <number_of_nodes>:  2
 <nt>: 8
 <aprun_command>:  aprun -n 8 -N 4 -d $OMP_NUM_THREADS <exe>

Comment: In addition to the constraints mentioned above, the -d value times the -N value has to be less or equal to 32. This configuration runs one processing element per NUMA domain and each PE spawns 8 threads.

XC30 Platform

The compute nodes of the XC30 platform Hornet feature two SandyBridge processors with 8 cores and one NUMA domain each resulting in a total of 16 cores and 2 NUMA domains per node. One conceptual difference between the Interlagos nodes on the XE6 and the SandyBridge nodes on the XC30 is the Hyperthreading feature of the SandyBridge processor. Hyperthreading is always booted and whether it is used or not is controlled via the -j option to aprun. Using -j 2 enables Hyperthreding while -j 1 (the default) does not. With Hyperthreading enabled, the compute node on the XC30 disposes 32 cores instead of 16.

• Description: Pure MPI application on two nodes fully packed (no OpenMP) with Hyperthreads

 <number_of_nodes>:  2
 <aprun_command>:  aprun -n 64 -N 32 -j 2 <exe>

• Description: Pure MPI application on two nodes fully packed (no OpenMP) without Hyperthreads

 <number_of_nodes>:  2
 <aprun_command>:  aprun -n 32 -N 16 -j 1 <exe>

Comment: Here you can also omit the -j 1 option as it is the default. This configuration corresponds to the wide-AVX case on the XE6 nodes.

• Description: Mixed (Hybrid) MPI OpenMP application on two nodes with Hyperthreading.

 <number_of_nodes>:  2
 <nt>: 2
 <aprun_command>:  aprun -n 32 -N 16 -d $OMP_NUM_THREADS -j 2 <exe>

---

General remarks for both platforms

The aprun allows to start an application with more OpenMP threads than compute cores available. This oversubscription results in a substantial performance degradation. The same happens if the -d value is smaller than the number of OpenMP threads used by the application. Furthermore, for the Intel programming environment an additional helper thread per processing element is spawned which can lead to an oversubscription. Here, one can use the -cc numa_node or the -cc none option to aprun to avoid this obersubscription of hardware. The default behavrior, i.e. if no -cc is specified, is as if -cc cpu is used which means that each processing element and thread is pinned to a processor. Please consult the aprun man page. Another popular option to aprun is -ss which forces memory allocation to be constrained in the same node as the processing element or thread is constrained. One can use the xthi.c utility to check the affinity of threads and processing elements.

---

Special Jobs / Special Nodes

MPP nodes with different memory (ONLY for CRAY XE6!)

3072 nodes of total 3552 nodes are installed with 32GB memory; 480 nodes are installed with 64GB memory.

32 GB nodes or 64 GB nodes

• If your job has not defined any node feature, then your job gets a default feature "mem32gb" which will allocate nodes with 32GB memory (see job examples above).

• If you want one or more of the 64GB nodes, then you have to specify the node feature "mem64gb":

 qsub -l nodes=1:mem64gb <my_batchjob_script.pbs>

Or inside a simple script my_batchjob_script.pbs

#!/bin/bash
#PBS -N job_name
#PBS -l nodes=1:mem64gb
#PBS -l walltime=00:20:00             
  
# Change to the direcotry that the job was submitted from
cd $PBS_O_WORKDIR

# Launch the parallel job to the allocated compute nodes
aprun -n 64 -N 32 ./my_mpi_executable arg1 arg2 > my_output_file 2>&1

Jobs with mixed node features

If your job needs some of the 64GB nodes and some of the 32GB nodes at the same time, then your job submission options looks totally different. Do not specify additional -l feature=<nodefeature>! You only need to specify resource name nodes=<node count>:ppn=<process count per node>:mem64gb+<node count>:ppn=<proc count per node>:mem32gb:

 qsub -l nodes=1:ppn=32:mem64gb+64:ppn=32:mem32gb,walltime=3600 my_batchjob_script.pbs

The example above will allocate 65 nodes to your job for a maximum time of 3600 seconds and can place 32 processes on one node with 64GB memory and 32 processes on each of the 64 allocated nodes with 32GB memory. Important is option ppn=32 to get all cores of the allocated mpp nodes. Now you need to select your different allocated nodes for your aprun command in your script my_batchjob_script.pbs:

#!/bin/bash
#PBS -N mixed_job
#PBS -l nodes=1:ppn=32:mem64gb+2:ppn=32:mem32gb
#PBS -l walltime=300

### defining the number of PEs (processes per node ( max 32 for hermit | max 16 for hornet) ###
# p32: number of PEs (Processing Elements) on 32GB nodes
# p64: number of PEs (Processing Elements) on 64GB nodes
#-------------------------------------------------------
p32=32
p64=16

# Change to the direcotry that the job was submitted from
cd $PBS_O_WORKDIR


### selecting nodes with different memory ###
#---------------------------------------
# 1. getting all nodes of my job
nids=$(/opt/hlrs/system/tools/getjhostlist)

# 2. getting the nodes with feature mem32gb of my job
nid32=$(/opt/hlrs/system/tools/hostlistf mem32gb "$nids")
# how many nodes do I have with mem32gb:
i32=$(/opt/hlrs/system/tools/cntcommastr "$nid32")

# 3. getting the nodes with feature mem64gb of my job
nid64=$(/opt/hlrs/system/tools/hostlistf mem64gb "$nids")
# how many nodes do I have with mem64gb:
i64=$(/opt/hlrs/system/tools/cntcommastr "$nid64")


(( P32 = $i32 * $p32 ))
(( P64 = $i64 * $p64 ))
(( D32 = 32 / $p32 ))
(( D64 = 32 / $p64 ))

# Launch the parallel job to the allocated compute nodes using
# Multi Program, Multi Data (MPMD) mode (see "man aprun")
# -------------------------------------------------
# $nid64 : node list with 64GB memory
# $i64     : number of nodes with 64GB memory
# $p64    : number of PEs per node on nodes with 64GB
# $P64    : total number of PEs (processing elements) on nodes with 64GB
# ----
# $nid32 : node list with 32GB memory
# $i32     : number of nodes with 32GB memory
# $p32    : number of PEs per node on nodes with 32GB
# $P32    : total number of PEs on nodes with 32GB
# ----------
# The "env OMP_NUM_THREADS=...." parts of the aprun command below are only useful for OpenMP (hybrid) programs.
#
aprun -L $nid64 -n $P64 -N $p64 -d $D64 env OMP_NUM_THREADS=$D64 ./my_executable1 : -L $nid32 -n $P32 -N $p32 -d $D32 env OMP_NUM_THREADS=$D32 ./my_executable2

By defining p64 and p32 in the example above you can control the number of processes on each node for the different node types (64GB memory and 32GB memory). Important to know is the maximum value is 32, the number of cores of each mpp node.

---

Nodes with Kepler accelerator cards (ONLY for CRAY XE6!)

28 nodes of total 3552 nodes are installed with 32GB memory, 16 cores each and with an additional Kepler accelerator card installed.

There are 2 ways to submit a job allocating the nodes with the Kepler accelerators.

• Using the queue gpgpu:

 qsub -q gpgpu -l mppwidth=16,mppnppn=16 <myjobscript> 

This allocates 1 node (16 cores) with Kepler accelerator, respectively nodes with feature tesla .

• Using the node feature tesla:

 qsub -l mppwidth=16,mppnppn=16,feature=tesla <myjobscript> 

This also allocates 1 node (16 cores) with Kepler accelerator.

For ccm jobs you need to use the second way submitting to the queue ccm :

  qsub -q ccm -l mppwidth=16,mppnppn=16,feature=tesla <myjobscript>

---

Pre- and Postprocessing/Visualization nodes with large memory (ONLY for CRAY XE6!)

A few number of external nodes are standard cluster nodes with 128GB memory and 32 cores of Intel Xeon CPU X7550. These nodes are not connected to the GEMINI interconnect network of the CRAY mpp compute nodes. But these nodes have the same workspace and home filesystem mounted and they have a graphic NVIDIA Quadro 6000 installed for visualization. To get one of the pre-postprocessing node, following qsub options are needed:

 qsub -l nodes=1:mem128gb,walltime=3600 ./my_batchjob_script.pbs

This allocates one of the pre-postprocessing node for 3600 seconds. It's not possible to get more than 1 of those nodes in one job.

Detail information for visualization are available on Graphic Environment. There you will find wrapper scripts and environment settings for your visualization work.

---

Shared node SMP with very large memory (ONLY for CRAY XE6!)

One external node has 1TB memory and 64 cores of Intel Xeon CPU X7550. This node is not connected to the GEMINI interconnect network of the CRAY XE6 mpp compute nodes. But this node has the same workspace and home filesystem mounted. This node is shared by several users and jobs at the same time! Users need to request the number of cores and the maximum of total memory needed by all processes you want to run. Otherwise the requested job gets very small defaults. The requested memory will be enforced, which means the job will be killed in case the job allocates more memory than requested.

To submit a job to this node you need following qsub options:

 qsub -q smp -l nodes=1:smp:ppn=2,vmem=100gb,walltime=3600 ./my_batchjob_script.pbs

This allocates 2 core (ppn=2) of the SMP node for 3600 seconds and limits the total memory used by all processes to 100GByte.

Note: All your processess will be grouped on 2 cores on this node. (see Linux cgroups mechanism)