Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

./configure error #35

Open
blazejkow opened this issue Sep 26, 2022 · 6 comments
Open

./configure error #35

blazejkow opened this issue Sep 26, 2022 · 6 comments

Comments

@blazejkow
Copy link

I am new to the software and I am trying to get acquainted with it. When I ran the ./configure script to check if everything checks out before installation I got the following error. I tried downgrading ubuntu and fortran versions but the same error keeps popping up. Do you have any idea how to get around this?
error
fortran_wersja

@mdelorme
Copy link
Contributor

Hi,

I have the same version of gfortran and no problem configuring the code.
Could you check inside the config.log file and check for a line that looks like :

configure:4434: checking whether C++ compiler accepts -fPIC

There should be more information about what is actually failing in the configuration stage here.

@blazejkow
Copy link
Author

Hi,

actually, it looks like the checks do not reach this point. they fail earlier and the script stops at that point (see attachment - configure:3332) so it never reaches configure:4434 which relates to c++, not Fortran btw. I wonder if I can manually add this option to gfortran compiler.
configlog

@mdelorme
Copy link
Contributor

actually, it looks like the checks do not reach this point. they fail earlier and the script stops at that point (see attachment - configure:3332) so it never reaches configure:4434 which relates to c++, not Fortran btw

You're right, my bad.

Looking at the log, it might be because MPI compilation is activated by default, and the system does not find a version of mpif77. Two options : Either you're not planning on running on multiple nodes, in which case MPI is utterly useless and you can try to configure adding the option --disable-mpi or, you want to have mpi activated in which case I would recommend installing the relevant OpenMPI packages in Ubuntu.

@blazejkow
Copy link
Author

It worked:) Thanks. It turned out I simply had to specify some arguments to the ./configure script. I was misled by the manual where it says that running it without any arguments will give you the answer about all the required setup but maybe I just misunderstood it:) anyway it worked after I fired:
./configure --disable-mpi --enable-simd=sse --prefix=/home/blazej/phd/nbody --enable-tools
success

@blazejkow
Copy link
Author

While we still have this thread open maybe you have an idea of how to proceed further. Upon executing 'make' I got the following error. Looks like two variables having different data types in chaos.f file. I assume I am lacking some packages in my environment otherwise this would pop up for everyone?
make

@mdelorme
Copy link
Contributor

I never had this with nbody6 but had this with another code. One way to avoid this is to relax the constraints on the compiler. Try this :

export FCFLAGS='-fallow-argument-mismatch'
./configure [--youroptionshere]

And rebuilding after.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants