Recent content by cva

Done that. Createa d a back-up arrray an tried to a line like this tmp=ckpsi, but as I said several times the contents of ckpsi are not accessible. In this code with one call, I can access the array ckpsi: do j=1,ly do i=1,lx...

My original code has the arrray global, however if I have a succession of calls (Fourier transforms) I can not access the values of this array . The only explanation is that the cpu does not synchronize. Could gice a small eample of what do you mean by the second suggestion?

Well I need to access this "stack based local array". declaring that array as global does not work. For examplem writing this: do j=1,ly do i=1,lx psi(i,j)=-0.55*(cos(sqrt(3.)*i*dx/2)*cos(j*dy/2)-0.5*cos(j*dy))-0.25 enddo enddo call dfftw_execute_dft_r2c(planpsi,psi,ckpsi) ...

The modified code wold be something like this: open(unit=1,file='hex.in',status='old') do j=1,ly do i=1,lx read(1,*) psi(i,j) enddo enddo close(1)call dfftw_execute(planpsi) ! call to make FT of array psi ! Only once do n=0,nend call update() enddocontains subroutine update() nt=psi**3...

The line psi**3 is ok. The fortran 90 vompilers reconize psi**3 as a nested loop and the result a matrix with each element with power 3. The A*B statement makes A(i,j)*B(i,j), while the psi**3 will result will do psi(i,j)**3 for each element. Arrays psi and ckpsi are golbal. During the...

Hello, Thank you for your reply. I was trying to write in a pseudocode. The ** takes power of each element of a matrix. I uploaded the whole code, but here is the relevant part of the code. open(unit=1,file='hex.in',status='old') do j=1,ly do i=1,lx read(1,*) psi(i,j) enddo enddo...

Hello, I am am confused about how fortran code is being executed and how are the variables in the memory updated during the execution. My code in simple terms looks like this: give some values to a matrix psi(lx,ly) (initialization) main loop do n=1,nend nt=psi^3 make Fourier...

The fftw library seems to install corectley Then as indicated in the manual the flas -lfftw3 and lfftw_3mpi are added to the compiling .

Good point. The MPI version on the server is HP-MPI v2.2.5 implementing MPI 1.2 and full MPI 2 Standard functionality. I compile the my codes with: bash-3.00$ mpif90 -show pgf90 -L/opt/hpmpi/lib/linux_amd64 -I/opt/hpmpi/include/64 -lhpmpio -lhpmpi -ldlWell I was just wondering if someone...

Hello, I've been trying for 2 weeks now to write in fortran a MPI program that computes the Fourier transform off a real matrix and then in vers transform using fftw-3.3alpha1. I installed the libray on evry computer availbale at my school and an I was able to compile yet my code. Even if...

So how do you inverse the function t?

Hello, Meanwhile I was able to solve the problem. The ouptput of fm:=minimize(..., location=true) is an array which contains the value of minima and a set like in the eq down. fm[1] := 0.2075577438e-1, {[{A2 = 0.3304115674e-2, A3 = -.5155123546, A1 = 0.9409892271e-1}, 0.2075577438e-1]}...

Thanks for the answers. Cristian

Thanks for your reply. My problem is that all these variables are function of v. and I need to have the dependence A1(v),A2(v) and A3(v). fm:=minimize(..., location=true) gives an array , but unfortunatley ,aple does not keep the order of the variables, every time is different. For example...

Thanks for your reply. My problem is that all these variables are function of v. and I need to have the dependence A1(v),A2(v) and A3(v). fm:=minimize(..., location=true) gives an array , but unfortunatley ,aple does not keep the order of the variables, every time is different. For example...