Estás aquí: start » en » centro » servizos » hpc

High Performance Computing (HPC) cluster ctcomp3

High Performance Computing (HPC) cluster ctcomp3

Description

The computing part of the cluster is made up of:

  • 9 servers for general computing.
  • 1 “fat node” for memory-intensive jobs.
  • 4 servers for GPU computing.

Users only have direct access to the login node, which has more limited features and should not be used for computing.
All nodes are interconnected by a 10Gb network.
There is distributed storage accessible from all nodes with 220 TB of capacity connected by a dual 25Gb fibre network.


Name Model Processor Memory GPU
hpc-login2 Dell R440 1 x Intel Xeon Silver 4208 CPU @ 2.10GHz (8c) 16 GB -
hpc-node[1-2] Dell R740 2 x Intel Xeon Gold 5220 @2,2 GHz (18c) 192 GB -
hpc-node[3-9] Dell R740 2 x Intel Xeon Gold 5220R @2,2 GHz (24c) 192 GB -
hpc-fat1 Dell R840 4 x Xeon Gold 6248 @ 2.50GHz (20c) 1 TB -
hpc-gpu[1-2] Dell R740 2 x Intel Xeon Gold 5220 CPU @ 2.20GHz (18c) 192 GB 2x Nvidia Tesla V100S
hpc-gpu3 Dell R7525 2 x AMD EPYC 7543 @2,80 GHz (32c) 256 GB 2x Nvidia Ampere A100 40GB
hpc-gpu4 Dell R7525 2 x AMD EPYC 7543 @2,80 GHz (32c) 256 GB 1x Nvidia Ampere A100 80GB

Accessing the cluster

To access the cluster, access must be requested in advance via incident form. Users who do not have access permission will receive an “incorrect password” message.

The access is done through an SSH connection to the login node:

ssh <nombre_de_usuario>@hpc-login2.inv.usc.es

Storage, directories and filesystems

None of the file systems in the cluster are backed up!!!

The HOME of the users in the cluster is on the file share system, so it is accessible from all nodes in the cluster. Path defined in the environment variable $HOME.
Each node has a local 1TB scratch partition, which is deleted at the end of each job. It can be accessed through the $LOCAL_SCRATCH environment variable in the scripts.
For data to be shared by groups of users, you must request the creation of a folder in the shared storage that will only be accessible by members of the group.

Directory Variable Mount point Capacity
Home $HOME /mnt/beegfs/home/<username> 220 TB*
local Scratch $LOCAL_SCRATCH varía 1 TB
Group folder $GRUPOS/<nombre> /mnt/beegfs/groups/<nombre> 220 TB*

* storage is shared

WARNING

The file share system performs poorly when working with many small files. To improve performance in such scenarios, create a file system in an image file and mount it to work directly on it. The procedure is as follows:

  • Create the image file at your home folder:
## truncate image.name -s SIZE_IN_BYTES
truncate example.ext4 -s 20G
  • Create a filesystem in the image file:
## mkfs.ext4 -T small -m 0 image.name
## -T small optimized options for small files
## -m 0 Do not reserve capacity for root user 
mkfs.ext4 -T small -m 0 example.ext4
  • Mount the image (using SUDO) with the script mount_image.py :
## By default it is mounted at /mnt/imagenes/<username>/ in read-only mode.
sudo mount_image.py example.ext4
  • To unmount the image use the script umount_image.py (using SUDO)

The mount script has this options:

--mount-point path   <-- (optional) This option creates subdirectories under /mnt/imagenes/<username>/<path> 
--rw                  <-- (optional) By default it is mounted readonly, with this option it is mounted readwrite.
Do not mount the image file readwrite from more than one node!!!

The unmounting script has this options:

only supports as an optional parameter the same path you have used when mounting with the option 
--mount-point  <-- (optional)

Transference of files and data

SCP

From your local machine to the cluster:

scp filename <username>@hpc-login2:/<path>

From the cluster to your local machine:

scp filename <username>@<hostname>:/<path>

SCP man page

SFTP

To transfer several files or to navigate through the filesystem.

<hostname>:~$ sftp <user_name>@hpc-login2
sftp>
sftp> ls
sftp> cd <path>
sftp> put <file>
sftp> get <file>
sftp> quit

SFTP man page

RSYNC

SSHFS

Requires local installation of the sshfs package.
Allows for example to mount the user's local home in hpc-login2:

## Mount
sshfs  <username>@ctdeskxxx.inv.usc.es:/home/<username> <mount_point>
## Unmount
fusermount -u <mount_point>

SSHFS man page

Available Software

All nodes have the basic software that is installed by default in AlmaLinux 8.4, in particular:

  • GCC 8.5.0
  • Python 3.6.8
  • Perl 5.26.3

GPU nodes, in addition:

  • nVidia Driver 510.47.03
  • CUDA 11.6
  • libcudnn 8.7

To use any other software not installed on the system or another version of the system, there are three options:

  1. Use Modules with the modules that are already installed (or request the installation of a new module if it is not available).
  2. Use a container (uDocker or Apptainer/Singularity)
  3. Use Conda

A module is the simplest solution for using software without modifications or difficult to satisfy dependencies.
A container is ideal when dependencies are complicated and/or the software is highly customised. It is also the best solution if you are looking for reproducibility, ease of distribution and teamwork.
Conda is the best solution if you need the latest version of a library or program or packages not otherwise available.

Modules/Lmod use

Lmod documentation

# See available modules:
module avail
# Module load:
module <module_name>
# Unload a module:
module unload <module_name>
# List modules loaded in your environment:
module list
# ml can be used as a shorthand of the module command:
ml avail
# To get info of a module:
ml spider <module_name>

Software containers execution

uDocker

uDocker manual
udocker is installed as a module, so it needs to be loaded into the environment:

ml uDocker

Apptainer/Singularity

Apptainer/Singularity documentation
Apptainer/Singularity is installed on each node's system, so you don't need to do anything to use it.

CONDA

Conda Documentation
Miniconda is the minimal version of Anaconda and only includes the conda environment manager, Python and a few necessary packages. From there on, each user only downloads and installs the packages they need.

# Getting miniconda
wget https://repo.anaconda.com/miniconda/Miniconda3-py39_4.11.0-Linux-x86_64.sh
# Install 
sh Miniconda3-py39_4.11.0-Linux-x86_64.sh
#  Initialize for bash shell
~/miniconda3/bin/conda init bash

Using SLURM

The cluster queue manager is SLURM .

The term CPU identifies a physical core in a socket. Hyperthreading is disabled, so each node has as many CPUs available as (number of sockets) * (number of physical cores per socket) it has.
Available resources
hpc-login2 ~]# ver_estado.sh
=============================================================================================================
  NODO     ESTADO                        CORES EN USO                           USO MEM     GPUS(Uso/Total)
=============================================================================================================
 hpc-fat1    up   0%[--------------------------------------------------]( 0/80) RAM:  0%     ---
 hpc-gpu1    up   2%[||------------------------------------------------]( 1/36) RAM: 47%   V100S (1/2)
 hpc-gpu2    up   2%[||------------------------------------------------]( 1/36) RAM: 47%   V100S (1/2)
 hpc-gpu3    up   0%[--------------------------------------------------]( 0/64) RAM:  0%   A100_40 (0/2)
 hpc-gpu4    up   1%[|-------------------------------------------------]( 1/64) RAM: 35%   A100_80 (1/1)
 hpc-node1   up   0%[--------------------------------------------------]( 0/36) RAM:  0%     ---
 hpc-node2   up   0%[--------------------------------------------------]( 0/36) RAM:  0%     ---
 hpc-node3   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node4   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node5   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node6   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node7   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node8   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
 hpc-node9   up   0%[--------------------------------------------------]( 0/48) RAM:  0%     ---
=============================================================================================================
TOTALES: [Cores : 3/688] [Mem(MB): 270000/3598464] [GPU: 3/ 7]
hpc-login2 ~]$ sinfo -e -o "%30N  %20c  %20m  %20f  %30G " --sort=N
# There is an alias for that command:
hpc-login2 ~]$ ver_recursos
NODELIST                        CPUS                  MEMORY                AVAIL_FEATURES        GRES                           
hpc-fat1                        80                    1027273               cpu_intel             (null)                         
hpc-gpu[1-2]                    36                    187911                cpu_intel             gpu:V100S:2                    
hpc-gpu3                        64                    253282                cpu_amd               gpu:A100_40:2                  
hpc-gpu4                        64                    253282                cpu_amd               gpu:A100_80:1(S:0)             
hpc-node[1-2]                   36                    187645                cpu_intel             (null)                         
hpc-node[3-9]                   48                    187645                cpu_intel             (null)
 
# To see current resource use: (CPUS (Allocated/Idle/Other/Total))
hpc-login2 ~]$ sinfo -N -r -O NodeList,CPUsState,Memory,FreeMem,Gres,GresUsed
# There is an alias for that command:
hpc-login2 ~]$ ver_uso
NODELIST            CPUS(A/I/O/T)       MEMORY              FREE_MEM            GRES                GRES_USED
hpc-fat1            80/0/0/80           1027273             900850              (null)              gpu:0,mps:0
hpc-gpu3            2/62/0/64           253282              226026              gpu:A100_40:2       gpu:A100_40:2(IDX:0-
hpc-gpu4            1/63/0/64           253282              244994              gpu:A100_80:1(S:0)  gpu:A100_80:1(IDX:0)
hpc-node1           36/0/0/36           187645              121401              (null)              gpu:0,mps:0
hpc-node2           36/0/0/36           187645              130012              (null)              gpu:0,mps:0
hpc-node3           36/12/0/48          187645              126739              (null)              gpu:0,mps:0
hpc-node4           36/12/0/48          187645              126959              (null)              gpu:0,mps:0
hpc-node5           36/12/0/48          187645              128572              (null)              gpu:0,mps:0
hpc-node6           36/12/0/48          187645              127699              (null)              gpu:0,mps:0
hpc-node7           36/12/0/48          187645              127002              (null)              gpu:0,mps:0
hpc-node8           36/12/0/48          187645              128182              (null)              gpu:0,mps:0
hpc-node9           36/12/0/48          187645              127312              (null)              gpu:0,mps:0

Nodes

A node is SLURM's computation unit and corresponds to a physical server.

# Show node info:
hpc-login2 ~]$ scontrol show node hpc-node1
NodeName=hpc-node1 Arch=x86_64 CoresPerSocket=18 
   CPUAlloc=0 CPUTot=36 CPULoad=0.00
   AvailableFeatures=cpu_intel
   ActiveFeatures=cpu_intel
   Gres=(null)
   NodeAddr=hpc-node1 NodeHostName=hpc-node1 Version=21.08.6
   OS=Linux 4.18.0-305.el8.x86_64 #1 SMP Wed May 19 18:55:28 EDT 2021 
   RealMemory=187645 AllocMem=0 FreeMem=166801 Sockets=2 Boards=1
   State=IDLE ThreadsPerCore=1 TmpDisk=0 Weight=1 Owner=N/A MCS_label=N/A
   Partitions=defaultPartition 
   BootTime=2022-03-01T13:13:56 SlurmdStartTime=2022-03-01T15:36:48
   LastBusyTime=2022-03-07T14:34:12
   CfgTRES=cpu=36,mem=187645M,billing=36
   AllocTRES=
   CapWatts=n/a
   CurrentWatts=0 AveWatts=0
   ExtSensorsJoules=n/s ExtSensorsWatts=0 ExtSensorsTemp=n/s

Partitions

Partitions in SLURM are logical groups of nodes. In the cluster there is a single partition to which all nodes belong, so it is not necessary to specify it when submitting jobs.

# Show partition info:
hpc-login2 ~]$ sinfo
defaultPartition*    up   infinite     11   idle hpc-fat1,hpc-gpu[1-4],hpc-node[1-9]

Jobs

Jobs in SLURM are resource allocations to a user for a given time. Jobs are identified by a sequential number or JOBID.
A JOB consists of one or more STEPS, each consisting of one or more TASKS that use one or more CPUs. There is one STEP for each program that executes sequentially in a JOB and there is one TASK for each program that executes in parallel. Therefore in the simplest case such as launching a job consisting of executing the hostname command the JOB has a single STEP and a single TASK.

Queue system (QOS)

The queue to which each job is submitted defines the priority, the limits and also the relative “cost” to the user.

# Show queues
hpc-login2 ~]$ sacctmgr show qos
# There is an alias that shows only the relevant info:
hpc-login2 ~]$ ver_colas
      Name    Priority                                  MaxTRES     MaxWall            MaxTRESPU MaxJobsPU MaxSubmitPU 
----------  ---------- ---------------------------------------- ----------- -------------------- --------- ----------- 
   regular         100                cpu=200,gres/gpu=1,node=4  4-04:00:00       cpu=200,node=4        10          50 
interactive        200                                   node=1    04:00:00               node=1         1           1 
    urgent         300                        gres/gpu=1,node=1    04:00:00               cpu=36         5          15 
      long         100                        gres/gpu=1,node=4  8-04:00:00                              1           5 
     large         100                       cpu=200,gres/gpu=2  4-04:00:00                              2          10 
     admin         500                                                                                                 
     small         150                             cpu=6,node=2    04:00:00              cpu=400        40         100 

# Priority: is the relative priority of each queue.
# DenyonLimit: job will not be executed if it doesn't comply with the queue limits
# UsageFactor: relive cost for the user to execute jobs on that queue
# MaxTRES: limnits applied to each job
# MaxWall: maximum time the job can run
# MaxTRESPU: global limits per user
# MaxJobsPU: Maximum number of jobs a user can have running simultaneously.
# MaxSubmitPU: Maximum number of jobs that a user can have in total both queued and running.

Sending a job to the queue system

Requesting resources

By default, if you submit a job without specifying anything, the system submits it to the default (regular) QOS and assigns it a node, a CPU and 4 GB. The time limit for job execution is that of the queue (4 days and 4 hours). This is very inefficient, the ideal is to specify as much as possible at least three parameters when submitting jobs:

  1. Node number (-N or --nodes), tasks (-n or --ntasks) and/or CPUs per task (-c or --cpus-per-task).
  2. Memory (--mem) per node or memory per cpu (--mem-per-cpu).
  3. Job execution time ( --time )

In addition, it may be interesting to add the following parameters:

-J --job-name Job name. Default: executable name
-q --qos Name of the queue to which the job is sent. Default: regular
-o --output File or file pattern to which all standard and error output is redirected.
--gres Type and/or number of GPUs requested for the job.
-C --constraint Para especificar que se quieren nodos con procesadores Intel o AMD (cpu_intel o cpu_amd)
--exclusive To specify that you want nodes with Intel or AMD processors (cpu_intel or cpu_amd)
-w --nodelist List of nodes to run the job on
How resources are allocated

The default allocation method between nodes is block allocation (all available cores on a node are allocated before using another node). The default allocation method within each node is cyclic allocation (the required cores are distributed equally among the available sockets in the node).

Priority calculation

When a job is submitted to the queuing system, the first thing that happens is that the requested resources are checked to see if they fall within the limits set in the corresponding queue. If it exceeds any of them, the submission is cancelled.
If resources are available, the job is executed directly, but if not, it is queued. Each job is assigned a priority that determines the order in which the jobs in the queue are executed when resources are available. To determine the priority of each job, 3 factors are weighted: the time it has been waiting in the queue (25%), the fixed priority of the queue (25%) and the user's fairshare (50%).
The fairshare is a dynamic calculation made by SLURM for each user and is the difference between the resources allocated and the resources consumed over the last 14 days.

hpc-login2 ~]$ sshare -l 
      User  RawShares  NormShares    RawUsage   NormUsage   FairShare 
---------- ---------- ----------- ----------- -----------  ---------- 
                         1.000000     2872400                0.500000 
                    1    0.500000     2872400    1.000000    0.250000 
user_name         100    0.071429        4833    0.001726    0.246436

# RawShares: Is the amount of resources allocated to the user in absolute terms . It is the same for all users.
# NormShares: This is the above amount normalised to the total allocated resources.
# RawUsage: The number of seconds/cpu consumed by all user jobs.
# NormUsage: RawUsage normalised to total seconds/cpu consumed in the cluster.
# FairShare: The FairShare factor between 0 and 1. The higher the cluster usage, the closer to 0 and the lower the priority.

Job submission
  1. sbatch
  2. salloc
  3. srun

1. SBATCH
Used to send a script to the queuing system. It is batch-processing and non-blocking.

# Crear el script:
hpc-login2 ~]$ vim test_job.sh
    #!/bin/bash
    #SBATCH --job-name=test              # Job name
    #SBATCH --nodes=1                    # -N Run all processes on a single node   
    #SBATCH --ntasks=1                   # -n Run a single task   
    #SBATCH --cpus-per-task=1            # -c Run 1 processor per task       
    #SBATCH --mem=1gb                    # Job memory request
    #SBATCH --time=00:05:00              # Time limit hrs:min:sec
    #SBATCH --qos=urgent                 # Queue
    #SBATCH --output=test%j.log          # Standard output and error log
 
    echo "Hello World!"
 
hpc-login2 ~]$ sbatch test_job.sh 

2. SALLOC
It is used to immediately obtain an allocation of resources (nodes). As soon as it is obtained, the specified command or a shell is executed.

# Get 5 nodes and launch a job.
hpc-login2 ~]$ salloc -N5 myprogram
# Get interactive access to a node (Press Ctrl+D to exit):
hpc-login2 ~]$ salloc -N1 
# Get interactive EXCLUSIVE access to a node
hpc-login2 ~]$ salloc -N1 --exclusive

3. SRUN
It is used to launch a parallel job (preferable to using mpirun). It is interactive and blocking.

# Launch the hostname command on 2 nodes
hpc-login2 ~]$ srun -N2 hostname
hpc-node1
hpc-node2

GPU use

To specifically request a GPU allocation for a job, options must be added to sbatch or srun:

--gres Request gpus per NODE --gres=gpu[[:type]:count],...
--gpus o -G Request gpus per JOB --gpus=[type]:count,...

There are also the options --gpus-per-socket,--gpus-per-node y --gpus-per-task,
Ejemplos:

## See the list of nodes and gpus:
hpc-login2 ~]$ ver_recursos
## Request any 2 GPUs for a JOB, add:
--gpus=2
## Request a 40G A100 at one node and an 80G A100 at another node, add:
--gres=gpu:A100_40:1,gpu:A100_80:1 

Job monitoring

## List all jobs in the queue
hpc-login2 ~]$ squeue
## Listing a user's jobs            
hpc-login2 ~]$ squeue -u <login>
## Cancel a job:
hpc-login2 ~]$ scancel <JOBID>
## List of recent jobs:
hpc-login2 ~]$ sacct -b
## Detailed historical information for a job:
hpc-login2 ~]$ sacct -l -j <JOBID>
## Debug information of a job for troubleshooting:
hpc-login2 ~]$ scontrol show jobid -dd <JOBID>
## View the resource usage of a running job:
hpc-login2 ~]$ sstat <JOBID>

Configure job output

Exit codes

By default these are the output codes of the commands:

SLURM command Exit code
salloc 0 success, 1 if the user's command cannot be executed
srun The highest among all executed tasks or 253 for an out-of-mem error.
sbatch 0 success, if not, the corresponding exit code of the failed process
STDIN, STDOUT y STDERR

SRUN:
By default stdout and stderr are redirected from all TASKS to srun's stdout and stderr, and stdin is redirected from srun's stdin to all TASKS. This can be changed with:

-i, --input=<option>
-o, --output=<option>
-e, --error=<option>

And options are:

  • all: by default.
  • none: Nothing is redirected.
  • taskid: Redirects only to and/or from the specified TASK id.
  • filename: Redirects everything to and/or from the specified file.
  • filename pattern: Same as the filename option but with a file defined by a pattern .

SBATCH:
By default “/dev/null” is open in the script's stdin and stdout and stderror are redirected to a file named “slurm-%j.out”. This can be changed with:

-i, --input=<filename_pattern>
-o, --output=<filename_pattern>
-e, --error=<filename_pattern>

The reference of filename_pattern is here .

Sending mail

JOBS can be configured to send mail in certain circumstances using these two parameters (BOTH ARE REQUIRED):

--mail-type=<type> Options: BEGIN, END, FAIL, REQUEUE, ALL, TIME_LIMIT, TIME_LIMIT_90, TIME_LIMIT_50.
--mail-user=<user> The destination mailing address.

Status of Jobs in the queuing system

hpc-login2 ~]# squeue -l
JOBID PARTITION     NAME     USER      STATE       TIME  NODES NODELIST(REASON)
6547  defaultPa  example <username>  RUNNING   22:54:55      1 hpc-fat1
 
## Check status of queue use:
hpc-login2 ~]$ estado_colas.sh
JOBS PER USER:
--------------
       usuario.uno:  3
       usuario.dos:  1
 
JOBS PER QOS:
--------------
             regular:  3
                long:  1
 
JOBS PER STATE:
--------------
             RUNNING:  3
             PENDING:  1
==========================================
Total JOBS in cluster:  4

Common job states:

  • R RUNNING Job currently has an allocation.
  • CD COMPLETED Job has terminated all processes on all nodes with an exit code of zero.
  • F FAILED Job terminated with non-zero exit code or other failure condition.
  • PD PENDING Job is awaiting resource allocation.

Full list of possible job statuses .

If a job is not running, a reason will be displayed underneath REASON: reason list for which a job may be awaiting execution.