Page History

Upute koje koriste sudo pri izgradnji nisu moguće na Supeku i Padobranu jer korisnici nemaju takve ovlasti.

Pri izgradnji ili nadogradnji na Supeku i Padobranu koristite opciju fakeroot (npr. za interaktivnu nadogradnju):

[korisnik@x3000c0s27b0n0 ~]$ apptainer shell --writable --fakeroot ...

Zbog bind model spajanja, potrebno je koristiti operativne sustave koji su slični operativnom sustavu Supeka Red Hat Enterprise Linux 8.6.

Neki od mogućih su:

• Rocky Linux 8
• AlmaLinux 8
• CentOS Stream 8

Bootstrap: localimage
From: /apps/utils/apptainer-images/rockylinux-8-mpich-4.1.1.sif

%files

  #kopiranje datoteka
  mpi.latency.c         /testmpi/

%post

  # prevedi aplikaciju u izvršnu datoteku
  mpicc     /testmpi/mpi.latency.c             -o /testmpi/mpi.latency

#PBS -q cpu
#PBS -l select=2:ncpus=1
#PBS -l place=scatter

cd $PBS_O_WORKDIR

module load utils/apptainer-bind/cray-mpich

mpiexec --no-transfer apptainer exec mpich.sif /testmpi/mpi.latency

.mpitutorial.com
// This code is provided freely with the tutorials on mpitutorial.com. Feel
// free to modify it for your own use. Any distribution of the code must
// either provide a link to www.mpitutorial.com or keep this header intact.
//
// An intro MPI hello world program that uses MPI_Init, MPI_Comm_size,
// MPI_Comm_rank, MPI_Finalize, and MPI_Get_processor_name.
//
#include <mpi.h>
#include <stdio.h>

int main(int argc, char** argv) {
  // Initialize the MPI environment. The two arguments to MPI Init are not
  // currently used by MPI implementations, but are there in case future
  // implementations might need the arguments.
  MPI_Init(NULL, NULL);
  
  // Get the number of processes
  int world_size;
  MPI_Comm_size(MPI_COMM_WORLD, &world_size);
  
  // Get the rank of the process
  int world_rank;
  MPI_Comm_rank(MPI_COMM_WORLD, &world_rank);

  // Get the name of the processor
  char processor_name[MPI_MAX_PROCESSOR_NAME];
  int name_len;
  MPI_Get_processor_name(processor_name, &name_len);

  // Print off a hello world message
  printf("Hello world from processor %s, rank %d out of %d processors\n",
         processor_name, world_rank, world_size);

  // Finalize the MPI environment. No more MPI calls can be made after this
  MPI_Finalize();
}

# spajanje na pristupni poslužitelj
[korisnik@kompjuter:~] $ ssh mkvakic@login-gpu.hpc.srce.hr

# izgradnja pytorch kontejnera
[korisnik@x3000c0s27b0n0:~] $ apptainer build pytorch_1.13.1.sif docker://pytorch/pytorch:1.13.1-cuda11.6-cudnn8-runtime

# print raznih informacija korištenjem --nv
[korisnik@x3000c0s27b0n0:~] $ apptainer run --nv ./pytorch_1.13.1.sif python -c "import torch; print('Is cuda available?:', torch.cuda.is_available())"
Is cuda available?: True
[korisnik@x3000c0s27b0n0:~] $ apptainer run --nv ./pytorch_1.13.1.sif python -c "import torch; print('Device count:', torch.cuda.device_count())"
Device count: 1
[korisnik@x3000c0s27b0n0:~] $ apptainer run --nv ./pytorch_1.13.1.sif python -c "import torch; print('Current device:', torch.cuda.current_device())"
Current device: 0
[korisnik@x3000c0s27b0n0:~] $ apptainer run --nv ./pytorch_1.13.1.sif python -c "import torch; print('Device name:', torch.cuda.get_device_name(0))"
Device name: NVIDIA A100-PCIE-40GB

# izgradnja test.sandbox sandbox slike kontejnera

# spajanje na pristupni poslužitelj
[korisnik@kompjuter:~] $ ssh mkvakic@login-gpu.hpc.srce.hr

# izgradnja pytorch kontejnera
[korisnik@x3000c0s27b0n0:~] $ apptainer build pytorch_1.13.1.sif docker://pytorch/pytorch:1.13.1-cuda11.6-cudnn8-runtime

# print raznih informacija korištenjem --nv
[korisnik@x3000c0s27b0n0:~] $ apptainer build run--sandbox --nvfix-perms ./pytorch_1.13.1.sif python -c "import torch; print('Is cuda available?:', torch.cuda.is_available())"
Is cuda available?: Truetest.sandbox test.def

# modifikacija slike kontejnera
[korisnik@x3000c0s27b0n0:~] $ touch test.sanbox/usr/bin/nvidia-smi
[korisnik@x3000c0s27b0n0:~] $ touch test.sanbox/usr/bin/nvidia-debugdump
[korisnik@x3000c0s27b0n0:~] $ apptainer run --nv ./pytorch_1.13.1.sif python -c "import torch; print('Device count:', torch.cuda.device_count())"
Device count: 1touch test.sanbox/usr/bin/nvidia-persistenced
[korisnik@x3000c0s27b0n0:~] $ touch test.sanbox/usr/bin/nvidia-cuda-mps-control
[korisnik@x3000c0s27b0n0:~] $ touch test.sanbox/usr/bin/nvidia-cuda-mps-server
[korisnik@x3000c0s27b0n0:~] $ apptainer runmkdir --nvp test./pytorch_1.13.1.sif python -c "import torch; print('Current device:', torch.cuda.current_device())"
Current device: 0sandbox/var/run/nvidia-persistenced/
[korisnik@x3000c0s27b0n0:~] $ touch test.sanbox/var/run/nvidia-persistenced/socket

# pokretanje "sandbox" kontainera sa --nv opcijom
[korisnik@x3000c0s27b0n0:~] $ apptainer runshell --nv ./pytorch_1.13.1.sif python -c "import torch; print('Device name:', torch.cuda.get_device_name(0))"
Device name: NVIDIA A100-PCIE-40GB--writable --fakeroot test.sandbox