mpi overlapping computation and communication added

.gitignore

 # put here the files which are not going to be committed
 *.txt
 *~
+*.bin
\ No newline at end of file

jacobi/mpi/comp_comm/Makefile

+#######################################################################
+# Author: David Goz (david.goz@inaf.it)                               #
+# June  2024                                                          #
+#######################################################################
+#
+# To see all the compilation options
+# $ make info
+#######################################################################
+
+# make.def defines how the application is compiled
+include make.def
+include make_mpi_path
+#######################################################################
+
+.PHONY: info mpi valgrind_memcheck valgrind_callgrind valgrind_cachegrind clean
+
+info:
+	@echo ' '
+	@echo '-----------------------------------------------------------------------------------------'
+	@echo '$$ make                     ---> compile the mpi application                             '
+	@echo '$$ make valgrind_memcheck   ---> run the mpi application using Valgrind under Memcheck   '
+	@echo '$$ make valgrind_callgrind  ---> run the mpi application using Valgrind under Callgrind  '
+	@echo '$$ make valgrind_cachegrind ---> run the mpi application using Valgrind under Cachegrind '
+	@echo '$$ make clean               ---> clean up all                                            '
+	@echo '$$ make info                ---> get make info                                           '
+	@echo '-----------------------------------------------------------------------------------------'
+	@echo ' '
+
+mpi: $(PROG)
+
+valgrind_memcheck: $(PROG_MEMCHECK)
+	@echo 'oooOOO... valgrind_memcheck ...OOOooo'
+	mpirun -n 4 valgrind --tool=memcheck -s --default-suppressions=yes --log-file=valgrind_memcheck_log_%p.txt ./$< 9 2
+	@echo 'oooOOO... valgrind_memcheck ...OOOooo'
+
+valgrind_callgrind: $(PROG_CALLGRIND)
+	@echo 'oooOOO... valgrind_callgrind ...OOOooo'
+	mpirun -n 4 valgrind --tool=callgrind --dump-instr=yes --collect-jumps=yes --log-file=valgrind_callgrind_log_.%p.txt ./$< 9 2
+	@echo ' '
+	@echo 'To generate a function-by-function summary from the profile data file:'
+	@echo '$$ callgrind_annotate --auto=yes callgrind.out.<pid> | less'
+	@echo '(kcachegrind is required in order to visualize the output using the GUI)'
+
+valgrind_cachegrind: $(PROG_CACHEGRIND)
+	@echo 'oooOOO... valgrind_cachegrind ...OOOooo'
+	mpirun -n 4 valgrind --tool=cachegrind --log-file=valgrind_cachegrind_log_.%p.txt ./$< 9 2
+	@echo '$$ cg_annotate --auto=yes cachegrind.out.<pid> | less'
+	@echo '(kcachegrind is required in order to visualize the output using the GUI)'
+	@echo 'oooOOO... valgrind_cachegrind ...OOOooo'
+
+clean:
+	rm -f *~ .*~ ./src/*~ ./src/*# ./include/*~ ./include/*# *~
+	rm -f $(PROG) $(PROG_DEBUG) $(PROG_MEMCHECK) $(PROG_CALLGRIND) $(PROG_CACHEGRIND)
+	rm -f valgrind_*.txt
+	rm -f cachegrind.out.*
+	rm -f callgrind.*
+	rm -f *bin

jacobi/mpi/comp_comm/include/allvars.h

+#pragma once
+
+#define NGHOST   1
+#define NDIM     2 /* 2D */
+#define X        0
+#define Y        1
+#define TOL      1.e-5
+
+#include <mpi.h>
+
+#define MASTER 0
+
+#if defined(SINGLE_PRECISION)
+typedef float MyData;
+#define MPI_MyDatatype MPI_FLOAT_PRECISION
+#else
+/* double precision is assumed by default */
+typedef double MyData;
+#define MPI_MyDatatype MPI_DOUBLE_PRECISION
+#endif /* SINGLE_PRECISION */

jacobi/mpi/comp_comm/include/tools.h

+#pragma once
+
+#include "allvars.h"
+#include <assert.h>
+#include <sys/time.h>
+#include <time.h>
+#include <sys/stat.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+/* function prototypes */
+MyData **Allocate_2DdblArray(const int nx, const int ny);
+void Show_2DdblArray(const MyData **const A,
+                     const int            nx,
+                     const int            ny,
+                     const char *const    string);
+
+double seconds(void);

jacobi/mpi/comp_comm/make.def

+CC     ?= gcc
+CFLAGS ?= -Wall -Wextra -march=native
+LIBS   ?= -lm -lmpi
+
+SYSTYPE = $(strip $(shell uname -n))
+
+PROG           ?= jacobi_mpi_comp_comm_$(SYSTYPE)
+PROG_DEBUG      = $(PROG)_DEBUG
+PROG_MEMCHECK   = $(PROG)_MEMCHECK
+PROG_CALLGRIND  = $(PROG)_CALLGRIND
+PROG_CACHEGRIND = $(PROG)_CACHEGRIND
+
+HEADERS       = $(wildcard ./include/*.h)
+SOURCES       = $(wildcard ./src/*.c)
+DEPENDENCIES  = $(SOURCES) $(HEADERS) Makefile
+
+$(PROG): $(DEPENDENCIES)
+	$(CC) $(CFLAGS) -O3 -I./include -I$(MPI_INC) $(SOURCES) -o $@ -L$(MPI_LIB) $(LIBS)
+	@echo ' '
+	@echo 'Program' $(PROG) 'compiled for' $(SYSTYPE) 'machine'
+	@echo ' '
+
+$(PROG_DEBUG): $(DEPENDENCIES)
+	$(CC) $(CFLAGS) -Og -ggdb3 -fno-omit-frame-pointer -I./include I$(MPI_INC) $(SOURCES) -o $@ -L$(MPI_LIB) $(LIBS)
+	@echo ' '
+	@echo 'Program' $(PROG_DEBUG) 'compiled for' $(SYSTYPE) 'machine'
+	@echo ' '
+
+$(PROG_MEMCHECK): $(DEPENDENCIES)
+	$(CC) $(CFLAGS) -Og -I./include -I$(MPI_INC) $(SOURCES) -o $@ -L$(MPI_LIB) $(LIBS)
+	@echo ' '
+	@echo 'Program' $(PROG_MEMCHECK) 'compiled for' $(SYSTYPE) 'machine'
+	@echo ' '
+
+$(PROG_CALLGRIND): $(DEPENDENCIES)
+	$(CC) $(CFLAGS) -g -O3 -I./include -I$(MPI_INC) $(SOURCES) -o $@ -L$(MPI_LIB) $(LIBS)
+	@echo ' '
+	@echo 'Program' $(PROG_CALLGRIND) 'compiled for' $(SYSTYPE) 'machine'
+	@echo ' '
+
+$(PROG_CACHEGRIND): $(DEPENDENCIES)
+	$(CC) $(CFLAGS) -g -O3 -I./include -I$(MPI_INC) $(SOURCES) -o $@ -L$(MPI_LIB) $(LIBS)
+	@echo ' '
+	@echo 'Program' $(PROG_CACHEGRIND) 'compiled for' $(SYSTYPE) 'machine'
+	@echo ' '

jacobi/mpi/comp_comm/make_mpi_path

+# set the MPI install path
+
+MPI_INSTALL_PATH = /home/gozzilla/software/openmpi/openmpi-5.0.3
+
+
+
+MPI        = $(MPI_INSTALL_PATH)
+MPI_INC    = $(MPI)/include
+MPI_LIB    = $(MPI)/lib
\ No newline at end of file

jacobi/mpi/comp_comm/src/tools.c

+/* ///////////////////////////////////////////////////////////////////// */
+/* Authors:  A. Mignone (mignone@to.infn.it)                             */
+/*           V. Cesare  (valentina.cesare@inaf.it)                       */
+/*           D. Goz     (david.goz@inaf.it)                              */
+/*                                                                       */
+/* Date   : June 2024                                                    */
+/*                                                                       */
+/* ///////////////////////////////////////////////////////////////////// */
+
+#include "tools.h"
+
+/* ********************************************************************* */
+MyData **Allocate_2DdblArray(const int nx, const int ny)
+/*
+ * Allocate memory for a double precision array with
+ * nx rows and ny columns
+ *********************************************************************** */
+{
+  MyData **buf = malloc(nx * sizeof(MyData *));
+  assert(buf != NULL);
+
+  buf[0] = (MyData *) malloc(nx * ny * sizeof(MyData));
+  assert(buf[0] != NULL);
+  
+  for (int j=1 ; j<nx ; j++)
+    buf[j] = buf[j-1] + ny;
+    
+  return buf;
+}
+
+/* ********************************************************************* */
+void Show_2DdblArray(const MyData **const A,
+		     const int            nx,
+		     const int            ny,
+		     const char *const    string)
+/* *********************************************************************** */
+{
+  printf ("%s\n",string);
+  printf ("------------------------------\n");
+  for (int i=0 ; i<nx ; i++)
+    {
+      for (int j=0 ; j<ny ; j++)
+	{
+	  printf ("%8.2f  ", A[i][j]);
+	}
+      printf ("\n");
+    }
+  printf ("------------------------------\n");
+}
+/* ********************************************************************* */
+
+double seconds()
+{
+  struct timeval tmp;
+  gettimeofday(&tmp, (struct timezone *)0);
+  double sec = tmp.tv_sec + ((double)tmp.tv_usec)/1000000.0;
+
+  return sec;
+}