I'm writing a subroutine that transform a regular vector into the one with only non-zero elements. Say, vector a=(0,0,1,2,3)' (n by 1). Then the non-zero vector is c=(1,2,3), and the row index is recorded as ic=(0,0,0,1,2,3) where ic(1)=0, ic(i+1)-ic(i) is the number of non-zero elements in i-th row. The vector index jc=(1,1,1) with size 3 as there are 3 non-zero entries. See the sparse matrix wiki for FYI: https://en.wikipedia.org/wiki/Sparse_matrix.
Despite its simplicity, I'm having troubles in running the following code named sparsem.f90
!This subroutine coverts a regular sparse matrix a into a CSR form
MODULE SPARSEM
CONTAINS
SUBROUTINE vsparse(a,c,jc,ic,counta,ierr,myid)
IMPLICIT NONE
REAL(8), INTENT(IN):: a(:)
INTEGER, INTENT(IN):: counta,myid
REAL(8), INTENT(OUT):: c(counta)
INTEGER, INTENT(OUT):: jc(counta),ic(size(a)+1)
INTEGER:: ierr,countaa,i
character(len=90):: filename
ierr=0
jc=0
c=0.0d0
ic=0
PRINT *, 'SIZE OF A IN VSPARSE', size(a),count(a>0.0d0),counta
IF (COUNT(a>0.0d0) /= counta) THEN
ierr=1
PRINT *, 'ERROR: number count of non-zero a(i,j) is not', counta
ELSE
countaa=0
ic(1)=0
DO i=1,size(a)
IF (a(i) > 0.0d0 ) THEN
countaa=countaa+1
c(countaa)=a(i)
ic(i+1)=ic(i)+1
jc(countaa)=1
IF (countaa<100) PRINT *,'checkcheckcheck', a(i), &
countaa,jc(countaa),c(countaa)，jc(1:5)
ELSE
ic(i+1)=ic(i)
END IF
END DO
PRINT *, 'JCJCJCJC',jc(1:5)
END IF
IF (myid==7) THEN
WRITE(filename,'("sparsedens_dcheck",I1,".txt")') myid+1
OPEN(UNIT=212101, FILE="/home/wenya/Workspace/Model4/valuef/"//filename,ACTION='write',status='replace')
DO i=1,counta+1
IF (i<=counta) THEN
WRITE(212101,*) c(i),jc(i)
ELSE
WRITE(212101,*) 0.0D0,0
END IF
END DO
CLOSE(212101)
END IF
return
END SUBROUTINE vsparse
END MODULE SPARSEM
So the three print jc codes shall give 1 1 1 1 1.... Yet starting from the second print jc code, the result is 6750960 6750691 6750692 .... The array of jc has size 9,000,000. And I know the first 2250000 element is 0.
To replicate this problem, here is the main program
PROGRAM MAIN
USE SPARSEM
IMPICIT NONE
REAL(8):: dens_last(9000000)
REAL(8), ALLOCATABLE :: dens(:)
INTEGER, ALLOCATABLE :: ic(:),jc(:)
INTEGER:: i
dens_last(1:2250000)=0.0d0
dens_last(2250001:9000000)=1.0d0/6750000.0d0
ncount=count(dens_last>0.0d0)
ALLOCATE(dens(ncount), ic(9000000+1), jc(ncount)_
CALL VSPASEM(dens_last, dens, jc, ic, ncount,ierr)
DEALLOCATE(dens,ic,jc)
END PROGRAM MAIN
I am using gfortran 6.3.0 and openmpi latest version on a UBUNTU 17.04 computer. Although openmpi is not used in this example, it's used in the rest of the program. Any thoughts? Thanks!
Related
I am trying to read a text file using a Fortran code. I have a file with 1999 rows and the number of columns vary with each row. Can someone please tell me how one can code such a problem. This is my code for reading a 4*2 text file but I am using do loops which I can't use in my current case.
PROGRAM myread2
IMPLICIT NONE
INTEGER, DIMENSION(100) :: a, b
INTEGER :: row,col,max_rows,max_cols
OPEN(UNIT=11, file='text.txt')
DO row = 1,4
READ(11,*) a(row), b(row)
END DO
PRINT *, a(1)
PRINT *, a(4)
PRINT*, b(4)
END PROGRAM myread2
The best way of reading a file like this depends on how you want to store the data. I'm going to use a ragged array as it's probably simplest, although other container types may be better suited depending on your requirements.
Fortran doesn't have ragged arrays natively, so first you need to define a type to hold each row. This can be done as
type :: RowData
integer, allocatable :: cols(:)
end type
type(RowData), allocatable :: rows(:)
When this container is filled out, the value in the i'th column of the j'th row will be accessed as
value = rows(j)%cols(i)
We can then write a program to read the file, e.g.
type :: RowData
integer, allocatable :: cols(:)
end type
type(RowData), allocatable :: rows(:)
integer :: no_rows
integer :: i
open(unit=11, file='text.txt')
no_rows = count_lines(11)
allocate(rows(no_rows))
do i=1,no_rows
rows(i)%cols = read_row(11)
enddo
Now we just need to write the functions count_lines, which counts the number of lines in the file, and read_row, which reads a line from the file and returns the contents of that line as an array of integers.
Following this question, count_lines can be written as
! Takes a file unit, and returns the number of lines in the file.
! N.B. the file must be at the start of the file.
function count_lines(file_unit) result(output)
integer, intent(in) :: file_unit
integer :: output
integer :: iostat
output = 0
iostat = 0
do while (iostat==0)
read(file_unit, *, iostat=iostat)
if (iostat==0) then
output = output+1
endif
enddo
rewind(file_unit)
end function
Writing read_row, to parse a line of unknown length from a file, can be done by following this question.
I need to call some fortran code that dynamically allocates arrays similar to this basic example. (unlike this example in the code I'm working with the size of the array is not known at the beginning of the function)
subroutine getIdentity(n,I)
integer, intent(in) :: n
integer, allocatable,dimension(:,:), intent(out) :: I
integer :: j,k
write(*,*) "creating " ,n, "x", n, "identity matrix"
allocate(I(n,n))
do j=1,n
do k=1,n
if(k==j) then
I(j,k) = 1
else
I(j,k) = 0
end if
end do
end do
end subroutine getIdentity
When I call this with this julia code:
I = zeros(Int32,1,1)
n = Ref{Int32}(3)
ccall((:__myModule_MOD_getidentity,"./test.so"), Cvoid ,
(Ref{Int32},Ref{Int32}), n,I)
println(I)
When I look at I in julia it is just garbage, not the Identity matrix I would expect. What would be the correct way to do this?
Hermite Interpolation woes
I am trying to find the Newton Dividing Differences for the function and derivative values of a given set of x's. I'm running into serious problems with my code working for tiny examples, but failing on bigger one's. As is clearly visible, my answers are very much larger than they original function values.
Does anybody have any idea what I'm doing wrong?
program inter
implicit none
integer ::n,m
integer ::i
real(kind=8),allocatable ::xVals(:),fxVals(:),newtonDivDiff(:),dxVals(:),zxVals(:),zdxVals(:),zfxVals(:)
real(kind=8) ::Px
real(kind=8) ::x
Open(Unit=8,File="data/xVals")
Open(Unit=9,File="data/fxVals")
Open(Unit=10,File="data/dxVals")
n = 4 ! literal number of data pts
m = n*2+1
!after we get the data points allocate the space
allocate(xVals(0:n))
allocate(fxVals(0:n))
allocate(dxVals(0:n))
allocate(newtonDivDiff(0:n))
!allocate the zvalue arrays
allocate(zxVals(0:m))
allocate(zdxVals(0:m))
allocate(zfxVals(0:m))
!since the size is the same we can read in one loop
do i=0,n
Read(8,*) xVals(i)
Read(9,*) fxVals(i)
Read(10,*) dxVals(i)
end do
! contstruct the z illusion
do i=0,m,2
zxVals(i) = xVals(i/2)
zxVals(i+1) = xVals(i/2)
zdxVals(i) = dxVals(i/2)
zdxVals(i+1) = dxVals(i/2)
zfxVals(i) = fxVals(i/2)
zfxVals(i+1) = fxVals(i/2)
end do
!slightly modified business as usual
call getNewtonDivDiff(zxVals,zdxVals,zfxVals,newtonDivDiff,m)
do i=0,n
call evaluatePolynomial(m,newtonDivDiff,xVals(i),Px,zxVals)
print*, xVals(i) ,Px
end do
close(8)
close(9)
close(10)
stop
deallocate(xVals,fxVals,dxVals,newtonDivDiff,zxVals,zdxVals,zfxVals)
end program inter
subroutine getNewtonDivDiff(xVals,dxVals,fxVals,newtonDivDiff,n)
implicit none
integer ::i,k
integer, intent(in) ::n
real(kind=8), allocatable,dimension(:,:) ::table
real(kind=8),intent(in) ::xVals(0:n),dxVals(0:n),fxVals(0:n)
real(kind=8), intent(inout) ::newtonDivDiff(0:n)
allocate(table(0:n,0:n))
table = 0.0d0
do i=0,n
table(i,0) = fxVals(i)
end do
do k=1,n
do i = k,n
if( k .eq. 1 .and. mod(i,2) .eq. 1) then
table(i,k) = dxVals(i)
else
table(i,k) = (table(i,k-1) - table(i-1,k-1))/(xVals(i) - xVals(i-k))
end if
end do
end do
do i=0,n
newtonDivDiff(i) = table(i,i)
!print*, newtonDivDiff(i)
end do
deallocate(table)
end subroutine getNewtonDivDiff
subroutine evaluatePolynomial(n,newtonDivDiff,x,Px,xVals)
implicit none
integer,intent(in) ::n
real(kind=8),intent(in) ::newtonDivDiff(0:n),xVals(0:n)
real(kind=8),intent(in) ::x
real(kind=8), intent(out) ::Px
integer ::i
Px = newtonDivDiff(n)
do i=n,1,-1
Px = Px * (x- xVals(i-1)) + newtonDivDiff(i-1)
end do
end subroutine evaluatePolynomial
Values
x f(x) f'(x)
1.16, 1.2337, 2.6643
1.32, 1.6879, 2.9989
1.48, 2.1814, 3.1464
1.64, 2.6832, 3.0862
1.8, 3.1553, 2.7697
Output
1.1599999999999999 62.040113431002474
1.3200000000000001 180.40121445431600
1.4800000000000000 212.36319446149312
1.6399999999999999 228.61845650513027
1.8000000000000000 245.11610836104515
You are accessing array newtonDivDiff out of bounds.
You are first allocating it as 0:n (main program's n) then you are passing to subroutine getNewtonDivDiff as 0:n (the subroutine's n) but you pass m (m=n*2+1) to the argument n. That means you tell the subroutine that the array has bounds 0:m which is 0:9, but it has only bounds 0:4.
It is quite difficult to debug the program as it stands, I had to use valgrind. If you move your subroutines to a module and change the dummy arguments to assumed shape arrays (:,:) then the bound checking in gfortran (-fcheck=all) will catch the error.
Other notes:
kind=8 is ugly, 8 can mean different things for different compilers. If you want 64bit variables, you can use kind=real64 (real64 comes from module iso_fortran_env in Fortran 2008) or use selected_real_kind() (Fortran 90 kind parameter)
You do not have to deallocate your local arrays in the subroutines, they are deallocated automatically.
Your deallocate statement in the main program is after the stop statement, it will never be executed. I would just delete the stop, there is no reason to have it.
I am writing a generic subroutine in fortran90 that will read in a column of data (real values). The subroutine should first check to see that the file exists and can be opened, then it determines the number of elements (Array_Size) in the column by reading the number of lines until end of file. Next the subroutine rewinds the file back to the beginning and reads in the data points and assigns each to an array (Column1(n)) and also determines the largest element in the array (Max_Value). The hope is that this subroutine can be written to be completely generic and not require any prior knowledge of the number of data points in the file, which is why the number of elements is first determined so the array, "Column1", can be dynamically allocated to contain "Array_Size" number of data points. Once the array is passed to the main program, it is transferred to another array and the initial dynamically allocated array is deallocated so that the routine can be repeated for multiple other input files, although this example only reads in one data file.
As written below, the program compiles just fine on the Intel fortran compiler; however, when it runs it gives me a severe (174): SIGSEV fault. I place the write(,) statements before and after the allocate statement in the subroutine and it prints the first statement "Program works here", but not the second, which indicates that the problem is occurring at the ALLOCATE (Column1(Array_Size)) statement, between the two write(,) statements. I re-compiled it with -C flag and ran the executable, which fails again and states severe (408): "Attempt to fetch from allocatable variable MISC_ARRAY when it is not allocated". The variable MISC_ARRAY is the dummy variable in the main program, which seems to indicate that the compiler wants the array allocated in the main program and not in the subprogram. If I statically allocate the array, the program works just fine. In order to make the program generic and not require any knowledge of the size of each file, it needs to be dynamically allocated and this should happen in the subprogram, not the main program. Is there a way to accomplish this that I am not seeing?
PROGRAM MAIN
IMPLICIT NONE
! - variable Definitions for MAIN program
INTEGER :: n
! - Variable Definitions for EXPENSE READER Subprograms
REAL, DIMENSION(:), ALLOCATABLE :: Misc_Array,MISC_DATA
INTEGER :: Size_Misc
REAL :: Peak_Misc_Value
! REAL :: Misc_Array(365)
CHARACTER(LEN=13) :: File_Name
File_Name = "Misc.txt"
CALL One_Column(File_Name,Size_Misc,Peak_Misc_Value,Misc_Array)
ALLOCATE (MISC_DATA(Size_Misc))
DO n = 1,Size_Misc ! Transfers array data
MISC_DATA(n) = Misc_Array(n)
END DO
DEALLOCATE (Misc_Array)
END PROGRAM MAIN
SUBROUTINE One_Column(File_Name,Array_Size,Max_Value,Column1)
IMPLICIT NONE
REAL, DIMENSION(:), ALLOCATABLE,INTENT(OUT) :: Column1
! REAL :: Column1(365)
REAL, INTENT(OUT) :: Max_Value
CHARACTER,INTENT(IN) :: File_Name*13
INTEGER, INTENT(OUT) :: Array_Size
INTEGER :: Open_Status,Input_Status,n
! Open the file and check to ensure it is properly opened
OPEN(UNIT=100,FILE = File_Name,STATUS = 'old',ACTION = 'READ', &
IOSTAT = Open_Status)
IF(Open_Status > 0) THEN
WRITE(*,'(A,A)') "**** Cannot Open ",File_Name
STOP
RETURN
END IF
! Determine the size of the file
Array_Size = 0
DO 300
READ(100,*,IOSTAT = Input_Status)
IF(Input_Status < 0) EXIT
Array_Size = Array_Size + 1
300 CONTINUE
REWIND(100)
WRITE(*,*) "Program works here"
ALLOCATE (Column1(Array_Size))
WRITE(*,*) "Program stops working here"
Max_Value = 0.0
DO n = 1,Array_Size
READ(100,*) Column1(n)
IF(Column1(n) .GT. Max_Value) Max_Value = Column1(n)
END DO
END SUBROUTINE One_Column
This is an educated guess: I think that the subroutine One_Column ought to have an explicit interface. As written the source code has 2 compilation units, a program (called main) and an external subroutine (called One_Column).
At compile-time the compiler can't figure out the correct way to call the subroutine from the program. In good-old (emphasis on old) Fortran style it takes a leap of faith and leaves it to the linker to find a subroutine with the right name and crosses its fingers (as it were) and hopes that the actual arguments match the dummy arguments at run-time. This approach won't work on subroutines returning allocated data structures.
For a simple fix move end program to the end of the source file, in the line vacated enter the keyword contains. The compiler will then take care of creating the necessary interface.
For a more scalable fix, put the subroutine into a module and use-associate it.
I think it is important to show the corrected code so that future users can read the question and also see the solution. I broke the subroutine into a series of smaller functions and one subroutine to keep the data as local as possible and implemented it into a module. The main program and module are attached. The main program includes a call to the functions twice, just to show that it can be used modularly to open multiple files.
PROGRAM MAIN
!
! - Author: Jonathan A. Webb
! - Date: December 11, 2014
! - Purpose: This code calls subprograms in module READ_COLUMNAR_FILE
! to determine the number of elements in an input file, the
! largest element in the input file and reads in the column of
! data as an allocatable array
!***************************************************************************
!***************************************************************************
!********************* **********************
!********************* VARIABLE DEFINITIONS **********************
!********************* **********************
!***************************************************************************
!***************************************************************************
USE READ_COLUMNAR_FILE
IMPLICIT NONE
CHARACTER(LEN=13) :: File_Name
INTEGER :: Size_Misc,Size_Bar,Unit_Number
REAL :: Peak_Misc_Value,Peak_Bar_Value
REAL, DIMENSION(:), ALLOCATABLE :: Misc_Array,Bar_Array
!***************************************************************************
!***************************************************************************
!********************* **********************
!********************* FILE READER BLOCK **********************
!********************* **********************
!***************************************************************************
!***************************************************************************
! - This section reads in data from all of the columnar input decks.
! User defines the input file name and number
File_Name = "Misc.txt"; Unit_Number = 100
! Determines the number of rows in the file
Size_Misc = File_Length(File_Name,Unit_Number)
! Yields the allocatable array and the largest element in the array
CALL Read_File(File_Name,Unit_Number,Misc_Array,Peak_Misc_Value)
File_Name = "Bar.txt"; Unit_Number = 100
Size_Bar = File_Length(File_Name,Unit_Number)
CALL Read_File(File_Name,Unit_Number,Bar_Array,Peak_Bar_Value)
END PROGRAM MAIN
MODULE READ_COLUMNAR_FILE
!***********************************************************************************
!***********************************************************************************
! ***
! Author: Jonathan A. Webb ***
! Purpose: Compilation of subprograms required to read in multi-column ***
! data files ***
! Drafted: December 11, 2014 ***
! ***
!***********************************************************************************
!***********************************************************************************
!
!-----------------------------------
! Public functions and subroutines for this module
!-----------------------------------
PUBLIC :: Read_File
PUBLIC :: File_Length
!-----------------------------------
! Private functions and subroutines for this module
!-----------------------------------
PRIVATE :: Check_File
!===============================================================================
CONTAINS
!===============================================================================
SUBROUTINE Check_File(Unit_Number,Open_Status,File_Name)
INTEGER,INTENT(IN) :: Unit_Number
CHARACTER(LEN=13), INTENT(IN) :: File_Name
INTEGER,INTENT(OUT) :: Open_Status
! Check to see if the file exists
OPEN(UNIT=Unit_Number,FILE = File_Name,STATUS='old',ACTION='read', &
IOSTAT = Open_Status)
IF(Open_Status .GT. 0) THEN
WRITE(*,*) "**** Cannot Open ", File_Name," ****"
STOP
RETURN
END IF
END SUBROUTINE Check_File
!===============================================================================
FUNCTION File_Length(File_Name,Unit_Number)
INTEGER :: File_Length
INTEGER, INTENT(IN) :: Unit_Number
CHARACTER(LEN=13),INTENT(IN) :: File_Name
INTEGER :: Open_Status,Input_Status
! Calls subroutine to check on status of file
CALL Check_File(Unit_Number,Open_Status,File_Name)
IF(Open_Status .GT. 0)THEN
WRITE(*,*) "**** Cannot Read", File_Name," ****"
STOP
RETURN
END IF
! Determine File Size
File_Length = 0
DO 300
READ(Unit_Number,*,IOSTAT = Input_Status)
IF(Input_Status .LT. 0) EXIT
File_Length = File_Length + 1
300 CONTINUE
CLOSE(Unit_Number)
END FUNCTION File_Length
!===============================================================================
SUBROUTINE Read_File(File_Name,Unit_Number,Column1,Max_Value)
INTEGER, INTENT(IN) :: Unit_Number
REAL, DIMENSION(:), ALLOCATABLE,INTENT(OUT) :: Column1
CHARACTER(LEN=13),INTENT(IN) :: File_Name
REAL, INTENT(OUT) :: Max_Value
INTEGER :: Array_Size,n
! Determines the array size and allocates the array
Array_Size = File_Length(File_Name,Unit_Number)
ALLOCATE (Column1(Array_Size))
! - Reads in columnar array and determines the element with
! the largest value
Max_Value = 0.0
OPEN(UNIT= Unit_Number,File = File_Name)
DO n = 1,Array_Size
READ(Unit_Number,*) Column1(n)
IF(Column1(n) .GT. Max_Value) Max_Value = Column1(n)
END DO
CLOSE(Unit_Number)
END SUBROUTINE Read_File
!===============================================================================
END MODULE READ_COLUMNAR_FILE
I am doing something I've done time after time, setting an array to zero in Fortran 90. However, for some reason in this case it is not working, and I have no idea why.
I allocate the array and use A = 0.d0 but when I write out one of the components it prints as 0.4xxx
My array is a module-level array if this makes any difference, and I am initializing it within a subroutine.
Does anybody have an idea why this could be happening?
EDIT: Sorry I have been away therefore haven't responded. This is still happening. I am using gfortran 4.3. I have changed a few things to see if they will help but they haven't. Notice below that I set the elements to zero two ways. Within the loop they are definitely being set to zero, but after the loop at least one element is becoming non-zero for no apparent reason. I know that all other elements are non-zero as well. I changed the array concerned to be a local subroutine array but this has no effect. The following is the code that is giving me wrong output:
subroutine coeff_cube(f, Ng,x_max_8,coeffs)
integer, intent(in) :: Ng
real(8), intent(in) :: f(Ng,Ng,Ng)
real(8), intent(in) :: x_max_8
integer :: i,j,k,ii,jj,kk
real(8) :: Ints(Ng,nmax+1)
real(8), intent(out) :: coeffs(nmax+1,nmax+1,nmax+1)
call cube_ints(x_max_8,Ng,Ints)
write(*,*) "NOW NMAX IS: ", nmax !Prints '24'
coeffs = 0.0d0
do i=1,nmax+1
do j=1,nmax+1
do k=1,nmax+1
coeffs(i,j,k) = 0.d0
write(*,*) coeffs(i,j,k) !Prints 0.0000000000000000 for all i,j,k
end do
end do
end do
write(*,*) coeffs(1,3,28) !Prints a non-zero number
coeffs(1,3,28) = 0.0d0
write(*,*) coeffs(1,3,28) !Prints 0.0000000000000000
do k=1,nmax+1
i=1
j=1
if (i+j+k .GT. nmax+1)then
exit
end if
do j=1,nmax+1
i=1
if (i+j+k .GT. nmax+1)then
exit
end if
do i=1,nmax+1
if (i+j+k .GT. nmax+1)then
exit
end if
do kk=1,Ng
do jj = 1,Ng
do ii = 1,Ng
coeffs(i,j,k) = coeffs(i,j,k) + &
& f(ii,jj,kk)*Ints(ii,i)*Ints(jj,j)*Ints(kk,k)
if(i==1.AND.j==3.AND.k==28)then
if (kk==1) then
write(*,*) coeffs(i,j,k)
end if
end if
end do
end do
end do
end do
end do
end do
write(*,*) coeffs(1,3,28) !Prints 0.0000000000000000
end subroutine
Does anyone have any ideas?
Thanks.
Sorry I have figured it out...
It does in fact set the array to zero, however for some reason I am writing an element that is not in the array (out of bounds). I would have expected it to give me an error rather than write out any old thing, but I guess that's just fortran...
Are you aware that you can initialize an entire array with a single statement?
coeffs = 0.0d
will set all elements of the array to 0.