second version of the code

Makefile

@@ -4,38 +4,32 @@
 
 # Choose pre-processing options
 #   -DSHM	   - enable shared-memory implementation
-#   -DDOUBLE_PREC  - use double-precision (default single)
+#   -DDOUBLE_PREC  - use double-precision
 OPTIONS =
 
 # Choose an FFT engine, available options are:
 #   fftw3      - FFTW version 3.x
 #   generic    - A general FFT algorithm (no 3rd-party library needed)
-FFT=generic
+FFT= generic
 
 # Paths to FFTW 3
-FFTW3_PATH=/home/lining/software/fftw-alpha
+FFTW3_PATH=   # full path of FFTW installation if using fftw3 engine above
 FFTW3_INCLUDE = -I$(FFTW3_PATH)/include
-FFTW3_LIB = -L$(FFTW3_PATH)/lib -lfftw3
+FFTW3_LIB = -L$(FFTW3_PATH)/lib -lfftw3 -lfftw3f
 
-# include PATH 
-ifeq ($(FFT),generic)
-  INC=
-else ifeq ($(FFT),fftw3)
-  INC=$(FFTW3_INCLUDE)
-endif
-
-# library path
-ifeq ($(FFT),generic)
-   LIBFFT=
-else ifeq ($(FFT),fftw3)
-   LIBFFT=$(FFTW3_LIB)
-endif
+# Specify Fortran and C compiler names and flags here
+# Normally, use MPI wrappers rather than compilers themselves 
+# Supply a Fortran pre-processing flag together with optimisation level flags
+# Some examples are given below:
 
-# Compiler 
+#FC =  
+#OPTFC = 
+#CC = 
+#CFLAGS = 
 
 # PGI
 #FC = ftn
-#OPTFC = -O3 -fast -Mpreprocess
+#OPTFC = -fast -O3 -Mpreprocess
 #CC = cc
 #CFLAGS = -O3
 
@@ -46,31 +40,43 @@ endif
 #CFLAGS = -O3
 
 # GNU
-#PREP=/home/lining/build/scalasca/bin/scalasca -inst
-#PREP=
-FC = $(PREP) mpif90
-OPTFC = -g -cpp
+FC = mpif90
+OPTFC = -O3 -funroll-loops -ftree-vectorize -fcray-pointer -cpp
 CC = mpicc
-CFLAGS = -g
+CFLAGS = -O3
 
-#Cray
+# Cray
 #FC = ftn
-#OPTFC = -O3 -F
+#OPTFC = -e Fm
 #CC = cc
-#CFLAGS = -O3
+#CFLAGS = 
+
+#-----------------------------------------------------------------------
+# Normally no need to change anything below
+
+# include PATH 
+ifeq ($(FFT),generic)
+  INC=
+else ifeq ($(FFT),fftw3)
+  INC=$(FFTW3_INCLUDE)
+endif
 
-# NAG
-#FC = f95
-#OPTFC = -O3 -kind=byte -I/usr/local/packages/nag/NAGWare5.1_496/NAGWare_f95-amd64/lib
+# library path
+ifeq ($(FFT),generic)
+   LIBFFT=
+else ifeq ($(FFT),fftw3)
+   LIBFFT=$(FFTW3_LIB)
+endif
 
 # List of source files
 SRC = decomp_2d.f90 glassman.f90 fft_$(FFT).f90 module_param.f90 io.f90 variables.f90 poisson.f90 schemes.f90 convdiff.f90 incompact3d.f90 navier.f90 derive.f90 parameters.f90 tools.f90 visu.f90
 
 #-----------------------------------------------------------------------
 # Normally no need to change anything below
 
-ifneq (,$(findstring DSHM,$(OPTIONS)))   
-OBJ =	$(SRC:.f90=.o) alloc_shm.o
+ifneq (,$(findstring DSHM,$(OPTIONS)))
+SRC := FreeIPC.f90 $(SRC)  
+OBJ =	$(SRC:.f90=.o) alloc_shm.o FreeIPC_c.o
 else
 OBJ =	$(SRC:.f90=.o)
 endif	
@@ -80,6 +86,9 @@ all: incompact3d
 alloc_shm.o: alloc_shm.c
 	$(CC) $(CFLAGS) -c $<
 
+FreeIPC_c.o: FreeIPC_c.c
+	$(CC) $(CFLAGS) -c $<
+
 incompact3d : $(OBJ)
 	$(FC) -O3 -o $@ $(OBJ) $(LIBFFT)
 

alloc.f90

@@ -0,0 +1,277 @@
+!=======================================================================
+! This is part of the 2DECOMP&FFT library
+! 
+! 2DECOMP&FFT is a software framework for general-purpose 2D (pencil) 
+! decomposition. It also implements a highly scalable distributed
+! three-dimensional Fast Fourier Transform (FFT).
+!
+! Copyright (C) 2009-2012 Ning Li, the Numerical Algorithms Group (NAG)
+!
+!=======================================================================
+
+  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
+  ! Utility routine to help allocate 3D arrays
+  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
+
+  ! X-pencil real arrays
+  subroutine alloc_x_real(var, opt_decomp, opt_global)
+
+    implicit none
+
+    real(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%xst(1):decomp%xen(1), &
+            decomp%xst(2):decomp%xen(2), decomp%xst(3):decomp%xen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%xsz(1),decomp%xsz(2),decomp%xsz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_x_real
+
+
+  ! X-pencil complex arrays
+  subroutine alloc_x_complex(var, opt_decomp, opt_global)
+
+    implicit none
+
+    complex(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%xst(1):decomp%xen(1), &
+            decomp%xst(2):decomp%xen(2), decomp%xst(3):decomp%xen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%xsz(1),decomp%xsz(2),decomp%xsz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_x_complex
+
+
+  ! Y-pencil real arrays
+  subroutine alloc_y_real(var, opt_decomp, opt_global)
+
+    implicit none
+
+    real(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%yst(1):decomp%yen(1), &
+            decomp%yst(2):decomp%yen(2), decomp%yst(3):decomp%yen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%ysz(1),decomp%ysz(2),decomp%ysz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_y_real
+
+
+  ! Y-pencil complex arrays
+  subroutine alloc_y_complex(var, opt_decomp, opt_global)
+
+    implicit none
+
+    complex(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%yst(1):decomp%yen(1), &
+            decomp%yst(2):decomp%yen(2), decomp%yst(3):decomp%yen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%ysz(1),decomp%ysz(2),decomp%ysz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_y_complex
+
+
+  ! Z-pencil real arrays
+  subroutine alloc_z_real(var, opt_decomp, opt_global)
+
+    implicit none
+
+    real(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%zst(1):decomp%zen(1), &
+            decomp%zst(2):decomp%zen(2), decomp%zst(3):decomp%zen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%zsz(1),decomp%zsz(2),decomp%zsz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_z_real
+
+
+  ! Z-pencil complex arrays
+  subroutine alloc_z_complex(var, opt_decomp, opt_global)
+
+    implicit none
+
+    complex(mytype), allocatable, dimension(:,:,:) :: var
+    TYPE(DECOMP_INFO), intent(IN), optional :: opt_decomp
+    logical, intent(IN), optional :: opt_global
+
+    TYPE(DECOMP_INFO) :: decomp
+    logical :: global
+    integer :: alloc_stat, errorcode
+
+    if (present(opt_decomp)) then
+       decomp = opt_decomp
+    else
+       decomp = decomp_main
+    end if
+
+    if (present(opt_global)) then
+       global = opt_global
+    else
+       global = .false.
+    end if
+
+    if (global) then
+       allocate(var(decomp%zst(1):decomp%zen(1), &
+            decomp%zst(2):decomp%zen(2), decomp%zst(3):decomp%zen(3)), &
+            stat=alloc_stat)
+    else
+       allocate(var(decomp%zsz(1),decomp%zsz(2),decomp%zsz(3)), &
+            stat=alloc_stat)
+    end if
+
+    if (alloc_stat /= 0) then
+       errorcode = 8
+       call decomp_2d_abort(errorcode, &
+            'Memory allocation failed when creating new arrays')
+    end if
+
+    return
+  end subroutine alloc_z_complex