MPP-templates.tar

MPP-templates/0000755000175000017500000000000013657173315011637 5ustar  dshdshMPP-templates/Makefile_cc0000755000175000017500000000100213657173241013736 0ustar  dshdshMF=	Makefile_cc

# For NEXGenIO
CC=	mpicxx
#CFLAGS=	-cxx=icpc

# For Cirrus
#CC=	mpicxx
#CFLAGS=	-cxx=icpc

# For ARCHER
#CC=	CC
#CFLAGS=	-g

LFLAGS=	-lm

EXE=	hello

INC= \
#	No user-defined include files at present - list here if required.

SRC= \
	hello.cc

#
# No need to edit below this line
#

.SUFFIXES:
.SUFFIXES: .cc .o

OBJ=	$(SRC:.cc=.o)

.cc.o:
	$(CC) $(CFLAGS) -c $<

all:	$(EXE)

$(OBJ):	$(INC)

$(EXE):	$(OBJ)
	$(CC) $(CFLAGS) -o $@ $(OBJ) $(LFLAGS)

$(OBJ):	$(MF)

clean:
	rm -f $(OBJ) $(EXE) core
MPP-templates/hello.cc0000644000175000017500000000023113620741025013232 0ustar  dshdsh#include "mpi.h"
#include <iostream>

using namespace std;

int main()
{
  MPI_Init(NULL, NULL);

  cout << "Hello World!" << endl;

  MPI_Finalize();
}
MPP-templates/Makefile_f900000644000175000017500000000066013657173214013755 0ustar  dshdshMF=	Makefile_f90

# For NEXTGenIO
FC=	mpif90
#FFLAGS=

# For Cirrus
#FC=	mpif90
#FFLAGS=	-fc=ifort

# For ARCHER
#FC=	ftn

LFLAGS=

EXE=	hello

SRC= \
	hello.f90

#
# No need to edit below this line
#

.SUFFIXES:
.SUFFIXES: .f90 .o

OBJ=	$(SRC:.f90=.o)
MOD=	$(SRC:.f90=.mod)

.f90.o:
	$(FC) $(FFLAGS) -c $<

all:	$(EXE)

$(EXE):	$(OBJ)
	$(FC) $(FFLAGS) -o $@ $(OBJ) $(LFLAGS)

$(OBJ):	$(MF)

clean:
	rm -f $(OBJ) $(EXE) $(MOD) core
MPP-templates/Makefile_c0000755000175000017500000000075713657173171013615 0ustar  dshdshMF=	Makefile_c

# For NEXTGenIO
CC=	mpicc
CFLAGS=

# For Cirrus
#CC=	mpicc
#CFLAGS=	-cc=icc

# For ARCHER
#CC=	cc
#CFLAGS=	-g

LFLAGS=	-lm

EXE=	hello

INC= \
#	No user-defined include files at present - list here if required.

SRC= \
	hello.c

#
# No need to edit below this line
#

.SUFFIXES:
.SUFFIXES: .c .o

OBJ=	$(SRC:.c=.o)

.c.o:
	$(CC) $(CFLAGS) -c $<

all:	$(EXE)

$(OBJ):	$(INC)

$(EXE):	$(OBJ)
	$(CC) $(CFLAGS) -o $@ $(OBJ) $(LFLAGS)

$(OBJ):	$(MF)

clean:
	rm -f $(OBJ) $(EXE) core
MPP-templates/archermpi.pbs0000644000175000017500000000370613620741025014312 0ustar  dshdsh# This is a simple template script to run batch jobs on ARCHER at EPCC
#
# You only have to change two things to run an MPI program with it:
#
# 1) Make a copy of the script with the same name as your MPI executable,
#    eg if the executable is 'myjob' then type: cp archermpi.pbs myjob.pbs
#
# 2) Set the variable "NPROC" appropriately (and "select" if NPROC > 24),
#    and your job will run using NPROC MPI processes.
#
# To run: qsub myjob.pbs (you could try "qsub -q short" for test jobs)
# For supervised practical sessions there may be a queue reserved for you.
#
# All screen output (stdout and stderr) will appear in a file called
# myjob.pbs.oXXXXX, where XXXXX is the job number assigned at submit time.
#
# David Henty, EPCC, 18/09/2016
#

#PBS -A y14
#PBS -j oe
#PBS -l walltime=00:01:00
#PBS -l select=1
  
#----------------------------------------------------------------------#
# You should only have to change the following parameter NPROC, unless #
# running on more than 24 processes when "select=1" must be increased. #
# The variable "select" should be set to the number of nodes required. #
# Each node has 24 physical cores so if NPROC=144 then use select=6    #
#----------------------------------------------------------------------#

NPROC=4

#------------------------------------------------------------------------#
# You should not have to edit below here for simple MPI jobs             #
#------------------------------------------------------------------------#

cd $PBS_O_WORKDIR

MPIPROG=`basename $PBS_JOBNAME .pbs`
MPISIZE=$NPROC

echo '--------------------------------------------------------------------------------'

echo 'Running MPI program' $MPIPROG 'on' $MPISIZE 'processes'

echo 'Started at' `date`
echo '--------------------------------------------------------------------------------'

(time aprun -n $MPISIZE ./$MPIPROG) 2>&1

echo '--------------------------------------------------------------------------------'
echo 'Finished at' `date`
MPP-templates/hello.c0000644000175000017500000000022513620741025013072 0ustar  dshdsh#include <stdio.h>
#include <stdlib.h>

#include <mpi.h>

int main(void)
{
  MPI_Init(NULL, NULL);

  printf("Hello World!\n");

  MPI_Finalize();
}
MPP-templates/cirrusmpi.pbs0000644000175000017500000000313313656272500014355 0ustar  dshdsh# This is a simple template script to run batch jobs on Cirrus at EPCC
#
# To submit: qsub cirrusmpinew.pbs
#
# For supervised practical sessions there may be a queue reserved for you,
# so use: qsub -q RXXXXXX cirrusmpi.pbs
#
# All screen output (stdout and stderr) will appear in a file called
# hello.pbs.oXXXXX, where XXXXX is the job number assigned at submit time.
#
# David Henty, EPCC, 11/05/2020
#

# "-N hello" sets the name of this PBS job
# "-j oe" sends error messages to the same log file as program output
# "-l walltime=00:01:00" sets the maximum run time to 1 minute
# "-l place=excl" ensures that you get exclusive access to all nodes

#PBS -A tc010
#PBS -N hello
#PBS -j oe
#PBS -l walltime=00:01:00
#PBS -l place=excl

# You will need to increase the value of "select" below if you run on
# more than 36 processes; however, the value of "ncpus" *is always 36*

#PBS -l select=1:ncpus=36

module load mpt
module load intel-compilers-18

# Change directory back to where the job was submitted from

cd $PBS_O_WORKDIR

echo 'Started at' `date`
echo '--------------------------------------------------------------------------------'

# Run on 4 processes with a maximum of 36 processes on each node.
#
# If you need more than 36 processes you will need to increase the
# "select" option above to ask for more nodes.  For example, for
# between 37 and 72 processes you would use: #PBS -l select=2:ncpus=36
# and run using, e.g.: mpiexec_mpt -ppn 36 -n 72 ./hello

mpiexec_mpt -ppn 36 -n 4 ./hello

echo '--------------------------------------------------------------------------------'
echo 'Finished at' `date`
MPP-templates/README0000644000175000017500000000125413656273010012511 0ustar  dshdshA few useful files for the MPI course.

The file "cirrusmpi.pbs" should be used for submitting MPI jobs to the
backend of Cirrus using the Portable Batch System PBS. Instructions
are included as comments in the file itself.

For very small test jobs on Cirrus, you can run interactively on the
login node using: mpirun -n 4 ./hello

The file "archermpi.pbs" should be used for submitting MPI jobs to the
backend of ARCHER using the Portable Batch System PBS. Instructions
are included as comments in the file itself.

The Makefiles should be easier to use than compiling by hand. They are
set up to deal with multiple source files, although they currently
only contain a single file.
MPP-templates/hello.f900000644000175000017500000000023713620741025013251 0ustar  dshdshprogram hello

  use mpi

  implicit none

  integer :: ierr

  call MPI_Init(ierr)

  write(*,*) 'Hello World!'

  call MPI_Finalize(ierr)

end program hello