|
Next revision
|
Previous revision
Next revision
Both sides next revision
|
cluster:218 [2022/06/30 13:23]
hmeij07 created |
cluster:218 [2022/09/08 08:51]
hmeij07 [Resources] |
| * August 2022 is designated **migration** period | * August 2022 is designated **migration** period |
| * Queues ''hp12'' and ''mwgpu'' (centos6) will be serviced by Openlava, not Slurm | * Queues ''hp12'' and ''mwgpu'' (centos6) will be serviced by Openlava, not Slurm |
| | |
| | |
| | ==== Quick Start Slurm Guide ==== |
| | |
| | Jump to the **Rocky8/CentOs7 script templates** listed in the menu of this page, top right. |
| | |
| | There is also detailed information on Amber20/Amber22 on this page with script examples. |
| | |
| | * [[cluster:214|Tada]] new head node |
| |
| ==== Basic Commands ==== | ==== Basic Commands ==== |
| |
| # sorta like bqueues | # sorta like bqueues |
| $ sinfo -l | sinfo -l |
| |
| # more node info | # more node info |
| # sorta like bhosts -l | # sorta like bhosts -l |
| scontrol show node n78 | scontrol show node n78 |
| | |
| | # sorta like bstop/bresume |
| | scontrol suspend job 1000001 |
| | scontrol resume job 1000001 |
| |
| # sorta like bhist -l | # sorta like bhist -l |
| |
| * manual pages for conf files or commands, for example | * manual pages for conf files or commands, for example |
| * ''man lsf.conf'' | * ''man slurm.conf'' |
| * ''man sbatch'' | * ''man sbatch'' |
| * etc ...(see above commands) | * etc ...(see above commands) |
| |
| You must request **resources**, that is for example number of cpu cores or which gpu model to use. ** If you do not request resources, Slurm will assume you need all the node's resources** and thus prevent other jobs from running on that node. | You must request **resources**, that is for example number of cpu cores or which gpu model to use. ** If you do not request resources, Slurm will assume you need all the node's resources** and thus prevent other jobs from running on that node. |
| | |
| | Details |
| | |
| | * https://slurm.schedmd.com/cons_res_share.html |
| |
| Some common examples are: | Some common examples are: |
| #SBATCH -n 8 # tasks=S*C*T | #SBATCH -n 8 # tasks=S*C*T |
| #SBATCH -B 2:4:1 # S:C:T=sockets/node:cores/socket:threads/core | #SBATCH -B 2:4:1 # S:C:T=sockets/node:cores/socket:threads/core |
| | #SBATCH --mem=250 # needed to override oversubscribe |
| | #SBATCH --ntasks-per-node=1 # perhaps needed to override oversubscribe |
| | #SBATCH --cpus-per-task=1 # needed to override oversubscribe |
| | |
| |
| GPU control | GPU control |
| #SBATCH --cpus-per-gpu=1 | #SBATCH --cpus-per-gpu=1 # needed to override oversubscribe |
| #SBATCH --mem-per-gpu=7168 | #SBATCH --mem-per-gpu=7168 # needed to override oversubscribe |
| #SBATCH --gres=gpu:geforce_gtx_1080_ti:1 # n[78], amber128 | #SBATCH --gres=gpu:geforce_gtx_1080_ti:1 # n[78], amber128 |
| #SBATCH --gres=gpu:geforce_rtx_2080_s:1 # n[79-90], exx96 | #SBATCH --gres=gpu:geforce_rtx_2080_s:1 # n[79-90], exx96 |
| Slurm has a builtin MPI flavor. I suggest you do not rely on it. The documentation states that on major release upgrades the ''libslurm.so'' library is not backwards compatible. All software using this library would need to be recompiled. | Slurm has a builtin MPI flavor. I suggest you do not rely on it. The documentation states that on major release upgrades the ''libslurm.so'' library is not backwards compatible. All software using this library would need to be recompiled. |
| |
| There is a handy parallel job launcher which may be of use, it is called ''srun''. ''srun'' commands can be embedded in a job submission script but it can also be used interactively to test commands out. | There is a handy parallel job launcher which may be of use, it is called ''srun''. ''srun'' commands can be embedded in a job submission script but it can also be used interactively to test commands out. The submmited job will have a single JOBPID and launch multiple tasks. |
| |
| <code> | <code> |
| [hmeij@cottontail2 ~]$ module avail | [hmeij@cottontail2 ~]$ module avail |
| |
| ------------------------ /opt/ohpc/pub/moduledeps/gnu9-openmpi4 ------------------------- | ------------------- /opt/ohpc/pub/moduledeps/gnu9-openmpi4 ------------- |
| adios/1.13.1 fftw/3.3.8 netcdf-cxx/4.3.1 petsc/3.16.1 py3-scipy/1.5.1 slepc/3.16.0 | adios/1.13.1 netcdf-cxx/4.3.1 py3-scipy/1.5.1 |
| boost/1.76.0 hypre/2.18.1 netcdf-fortran/4.5.3 phdf5/1.10.8 scalapack/2.1.0 superlu_dist/6.4.0 | boost/1.76.0 netcdf-fortran/4.5.3 scalapack/2.1.0 |
| dimemas/5.4.2 imb/2019.6 netcdf/4.7.4 pnetcdf/1.12.2 scalasca/2.5 tau/2.29 | dimemas/5.4.2 netcdf/4.7.4 scalasca/2.5 |
| example2/1.0 mfem/4.3 omb/5.8 ptscotch/6.0.6 scorep/6.0 trilinos/13.2.0 | example2/1.0 omb/5.8 scorep/6.0 |
| extrae/3.7.0 mumps/5.2.1 opencoarrays/2.9.2 py3-mpi4py/3.0.3 sionlib/1.7.4 | extrae/3.7.0 opencoarrays/2.9.2 sionlib/1.7.4 |
| | fftw/3.3.8 petsc/3.16.1 slepc/3.16.0 |
| | hypre/2.18.1 phdf5/1.10.8 superlu_dist/6.4.0 |
| | imb/2019.6 pnetcdf/1.12.2 tau/2.29 |
| | mfem/4.3 ptscotch/6.0.6 trilinos/13.2.0 |
| | mumps/5.2.1 py3-mpi4py/3.0.3 |
| |
| ------------------------------ /opt/ohpc/pub/moduledeps/gnu9 ------------------------------ | ------------------------- /opt/ohpc/pub/moduledeps/gnu9 ---------------- |
| R/4.1.2 impi/2021.5.1 mpich/3.4.2-ofi openblas/0.3.7 plasma/2.8.0 superlu/5.2.1 | R/4.1.2 mpich/3.4.2-ofi plasma/2.8.0 |
| gsl/2.7 likwid/5.0.1 mpich/3.4.2-ucx (D) openmpi4/4.1.1 (L) py3-numpy/1.19.5 | gsl/2.7 mpich/3.4.2-ucx (D) py3-numpy/1.19.5 |
| hdf5/1.10.8 metis/5.1.0 mvapich2/2.3.6 pdtoolkit/3.25.1 scotch/6.0.6 | hdf5/1.10.8 mvapich2/2.3.6 scotch/6.0.6 |
| | impi/2021.5.1 openblas/0.3.7 superlu/5.2.1 |
| | likwid/5.0.1 openmpi4/4.1.1 (L) |
| | metis/5.1.0 pdtoolkit/3.25.1 |
| |
| -------------------------------- /opt/ohpc/pub/modulefiles ------------------------------- | --------------------------- /opt/ohpc/pub/modulefiles ------------------- |
| EasyBuild/4.5.0 example1/1.0 libfabric/1.13.0 (L) prun/2.2 (L) | EasyBuild/4.5.0 hwloc/2.5.0 (L) prun/2.2 (L) |
| autotools (L) gnu9/9.4.0 (L) ohpc (L) singularity/3.7.1 | autotools (L) intel/2022.0.2 singularity/3.7.1 |
| charliecloud/0.15 hwloc/2.5.0 (L) os ucx/1.11.2 (L) | charliecloud/0.15 libfabric/1.13.0 (L) ucx/1.11.2 (L) |
| cmake/3.21.3 intel/2022.0.2 papi/5.7.0 valgrind/3.18.1 | cmake/3.21.3 ohpc (L) valgrind/3.18.1 |
| | example1/1.0 os |
| | gnu9/9.4.0 (L) papi/5.7.0 |
| |
| --------------------------- /share/apps/CENTOS8/ohpc/modulefiles --------------------------- | ----------------------- /share/apps/CENTOS8/ohpc/modulefiles ------------ |
| amber/20 cuda/11.6 hello-mpi/1.0 hello/1.0 miniconda3/py39 | amber/20 cuda/11.6 hello-mpi/1.0 hello/1.0 miniconda3/py39 |
| |
| |
| |
| * ''/zfshomes/hmeij/slurm/run.rocky'' | * ''/zfshomes/hmeij/slurm/run.rocky'' for tinymem, mw128, amber128, test queues |
| |
| <code> | <code> |
| #SBATCH -B 1:1:1 # S:C:T=sockets/node:cores/socket:threads/core | #SBATCH -B 1:1:1 # S:C:T=sockets/node:cores/socket:threads/core |
| ###SBATCH -B 2:4:1 # S:C:T=sockets/node:cores/socket:threads/core | ###SBATCH -B 2:4:1 # S:C:T=sockets/node:cores/socket:threads/core |
| #SBATCH --cpus-per-gpu=1 | |
| #SBATCH --mem-per-gpu=7168 | |
| # | # |
| # GPU control | # GPU control |
| | #SBATCH --cpus-per-gpu=1 |
| | #SBATCH --mem-per-gpu=7168 |
| ###SBATCH --gres=gpu:geforce_gtx_1080_ti:1 # n78 | ###SBATCH --gres=gpu:geforce_gtx_1080_ti:1 # n78 |
| #SBATCH --gres=gpu:quadro_rtx_5000:1 # n[100-101] | #SBATCH --gres=gpu:quadro_rtx_5000:1 # n[100-101] |
| cd $MYLOCALSCRATCH | cd $MYLOCALSCRATCH |
| |
| ### AMBER20 | ### AMBER20 works via slurm's imaged nodes, test and amber128 queues |
| #source /share/apps/CENTOS8/ohpc/software/amber/20/amber.sh | #source /share/apps/CENTOS8/ohpc/software/amber/20/amber.sh |
| # OR # | # OR # |
| ==== CentOS7 Slurm Template ==== | ==== CentOS7 Slurm Template ==== |
| |
| In this job template I have it setup to run ''pmemd.MPI'' but could also invoke ''pmemd.cuda'' with proper parameter settings. I could also toggle between amber16 or amber20 which on queues ''mwgpu'' and ''exx96'' are local disk CentOS7 software installations. Amber16 will not run on Rocky8 (tried it but forgot error message...we can expect problems like this, hence testing!). | In this job template I have it setup to run ''pmemd.MPI'' but could also invoke ''pmemd.cuda'' with proper parameter settings. On queues ''mwgpu'' and ''exx96'' amber[16,20] are local disk CentOS7 software installations. Amber16 will not run on Rocky8 (tried it but forgot error message...we can expect problems like this, hence testing!). |
| |
| Note also that we're running mwgpu's K20 cuda version 9.2 on exx96 queue (default cuda version 10.2). Not proper but it works. Hence this script will run on both queues. Oh, now I remember, it is that amber16 was compiled with cuda 9.2 drivers which are supported in cuda 10+_ but not in cuda 11+. So Amber 16, if needed, would need to be compiled in Rocky8 environment. (that may work like amber20). | Note also that we're running mwgpu's K20 cuda version 9.2 on exx96 queue (default cuda version 10.2). Not proper but it works. Hence this script will run on both queues. Oh, now I remember, it is that amber16 was compiled with cuda 9.2 drivers which are supported in cuda 10.x but not in cuda 11.x. So Amber 16, if needed, would need to be compiled in Rocky8 environment (and may work like amber20 module). |
| |
| * ''/zfshomes/hmeij/slurm/run.centos'' | * ''/zfshomes/hmeij/slurm/run.centos'' for mwgpu, exx96 queues |
| |
| <code> | <code> |
| # | # |
| # GPU control | # GPU control |
| ###SBATCH --gres=gpu:tesla_k20m:1 # n[33-37] | |
| ###SBATCH --gres=gpu:geforce_rtx_2080_s:1 # n[79-90] | |
| ###SBATCH --cpus-per-gpu=1 | ###SBATCH --cpus-per-gpu=1 |
| ###SBATCH --mem-per-gpu=7168 | ###SBATCH --mem-per-gpu=7168 |
| | ###SBATCH --gres=gpu:tesla_k20m:1 # n[33-37] |
| | ###SBATCH --gres=gpu:geforce_rtx_2080_s:1 # n[79-90] |
| # | # |
| # Node control | # Node control |
| #SBATCH --nodelist=n88 | #SBATCH --nodelist=n88 |
| |
| | # may or may not be needed, centos7 login env |
| | source $HOME/.bashrc |
| | which ifort # should be the parallel studio 2016 version |
| |
| # unique job scratch dirs | # unique job scratch dirs |
| |
| |
| ###source /usr/local/amber16/amber.sh | |
| source /usr/local/amber20/amber.sh | ###source /usr/local/amber16/amber.sh # works via slurm's mwgpu |
| | source /usr/local/amber20/amber.sh # works via slurm's exx96 |
| # stage the data | # stage the data |
| cp -r ~/sharptail/* . | cp -r ~/sharptail/* . |
| July 2022 is for **testing...** lots to learn! | July 2022 is for **testing...** lots to learn! |
| |
| Kudos to Abhilash for working our way through all this. | Kudos to Abhilash and Colin for working our way through all this. |
| |
| \\ | \\ |
| **[[cluster:0|Back]]** | **[[cluster:0|Back]]** |
| |
