'Unknown option: -localonly' when running Coarray program

[FruitBread]

Hello everyone.

I am recently converting an old Fortran 90 serial program to parallel. I intended to use Coarray feature to do this. As I tried running my example code with IFX Intel Fortran Compiler 2024.2.1 (integrated with Microsoft Visual Studio 2022 Community, v17.11.5), the compiling and linking procedure finished without any problem, but the generated exe starts with an error. Code example and exe result are as below:

PROGRAM A
    print*,this_image()	
END

I have already enabled Coarrays for shared memory and set number of images. It seems there is an argument  '-localonly'  which cannot be identified. In Visual Studio project configuration pages I did not find such command option:

I tried to figure out how to disable this argument but failed (did not find where to use commands rather than GUI)

P.S. the oneAPI version (base and HPC toolkits) is 2024.2.1. I tried different configurations (Debug/Release) and platforms (x86/x64), and also tried Intel Fortran Compiler Classic, but none worked.

Can anyone help with this problem? Thanks

Best regards,

FruitBread

[Steve_Lionel]

-localonly is not an Intel Fortran option. I have tried reproducing this problem but was not successful. Would you please ZIP the project folder (and any subfolders) and attach the ZIP here?

I do see that -localonly is an Intel Hydra option, but am unsure if that is relevant here.

[FruitBread]

Hi Lionel!

I have attached the zipped Visual Studio project file in this reply.

Yes, -localonly is an Intel MPI Hydra Option rather than IFX/IVF command. The coarray program relies on MPI library to run, and this might be the problem. Perhaps I should transfer this post to Intel MPI Library forum for help?

Thank you!

[Steve_Lionel · Accepted solution]

Thanks. I ran the exe in the project and did not see the error you showed. There may be something odd in your system's install of Intel MPI.

[FruitBread]

Got it. I will look for the problems in my MPI configurations. Thanks a lot!