-
Notifications
You must be signed in to change notification settings - Fork 0
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
Compiling the GACODE in the JET Heimdall computer #358
Comments
@franal12 Is there a version of mpif90 built with gfortran? |
Thanks for your response! Yes, a version of openmpi can be loaded along with gfortran. |
Generally for commodity CPU systems I recommend:
For openblas use these settings
Then, the |
@franal12 Would it be possible to add some minimal credentials to you github profile? |
@franal12 has there been progress? |
Hi @jcandy! I compiled the openblas by using the settings mentioned by you and I am getting the following error
I was asking for help from the local IT support so I couldn't get back to you quickly. I updated my profile now on Github. |
I assume |
yes it is generated so I will try to build GACODE and update the thread. |
The make file contains the following lines based on the MINT makefile
and I am getting the following error
Is there a document or a guide to understand how makefiles are created? I found some guides on the internet but they explain basic makefile syntaxes. |
Is it possible for me to get an account on Heimdall? |
I don't think that is possible but we can have a ZOOM meeting and try compiling the Gacode together. |
@jcandy
Yes, there is a cgyro_globals.mod file in gacode/module. There are many instances of “Rank mismatch” errors. Compiling gyro gives this error at the end |
Please generate a make log:
and attach the file here |
Here is the log file after cgyro compilation. |
Are you using an old version of GACODE (like from 2022?) |
Apparently, it is an older version of code from October 2023. I downloaded the latest GACODE and recompiled the cgyro again. I attached the log file for your reference |
The error log shows that you don't have FFTW3 installed. |
The necessary file fftw.f03 is in /usr/include/fftw3.f03. How can I change the makefile to include this file? |
Here are the typical settings:
So, where are the |
Thanks! I found libfftt3.a and libfftw3_omp.a libraries in usr/lib64 directory and I made the corresponding changes in the make file. After compiling the cgyro code, I am getting undefined reference to `cgyro_nl_fftw_comm2_test_' . I have attached the log file for your reference. |
I don't think the iso_fortran_env is the root cause. It is just a warning. @franal12 To be pedantic, Note: cgyro_rhs has no warning, but still fails. |
@sfiligoi I checked the gacode directory and there are no extra files present. |
To recap, please do the usual steps:
|
@franal12 I just tried with conda installed gfortran and mpi, and it compiled without a problem using the MINT platform:
Just had to tweak the paths in make.inc.MINT:
Something seems to be wrong with your setup. |
I also run the CGYRO validation tests, and they all pass.
|
@jcandy and @sfiligoi I checked modules and cgyro/src directories for stale mod files and they are not present. System math librariesLMATH = /home/wh4980/OpenBLAS-develop/libopenblas.a /usr/lib64/libfftw3.a /usr/lib64/libfftw3_omp.a # netCDF (not optional any more)# NETCDF_HOME=/usr/local/depot/netcdf-4.2.1.1/# NETCDF_DIR is set by the moduleNETCDF=-L${NETCDF_DIR}/lib -lnetcdff ArchiveARCH = ar cr |
A make clean was not done here. Can you start with a make clean in the gacode top directory? |
@franal12 What version of fortran and MPI are you using? Here is mine:
|
@sfiligoi Here are the details:
|
That's a very old version of gfortran! |
Sure! I will ask the IT support for a newer version and update the thread. I will confirm if I can use COONDA at JET computers. |
@sfiligoi I installed the necessary modules via conda and compiled the cgyro again after doing the necessary steps.
I am still getting the compilation error and I have attached the log file for your reference |
@franal12 Can we see the new software versions? |
Sure! Here is the version information. |
Further information if needed. |
@jcandy, I reinstalled the conda environment and still got the same error. I don't know what is causing the same error. Do you have any pointers to resolve this problem? |
@franal12 If you are using the latest gacode from git, I do not know what to suggest. Maybe one more thought: |
sorry for my delay in response, I was busy with my EPS conferene preparations in the last couple of weeks. I can compile the gacode on my personal linux machine with the same procedure and the makefile so I beleive the error could be due to the environement settings of JET Heimdall system. |
Hi GACODE team!
I am planning to compile the gacode on the JET Heimdall computer using the existing makefile JAC-64 in the build directory. I can compile TGLF but am unable to compile any other codes due to an error in the compilation of the math. The compilation error is shown below
Any tips on how to resolve this error?
Thanks,
Francis
The text was updated successfully, but these errors were encountered: