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
Error: Symbol 'mpi_complex' at (1) has no IMPLICIT type #415
Comments
Not sure why it even uses MPICH when it is disabled for it. Settings can be seen here: https://github.com/macports/macports-ports/blob/master/science/elmerfem/Portfile |
Also, see this forum post with similar error: http://www.elmerfem.org/forum/viewtopic.php?p=28919#p28919 Building Elmer without MPI in Windows 10 using yesterday's devel branch, gives the same error. Rich. |
@richb2k I see, thank you for the info. Looks like the code has actually been messed up (I had no time to go through commit history and Git blame, perhaps something identifiable). I can’t recall why I have set it to off in the port though. Maybe there was some strong reason (a failure with GCC, or with Clang, or dependencies did not support MPICH), maybe just decided to make the port simple initially. I can try using MPICH, normally it works. |
Here is the error message, compiled without MPI, in Windows 10: C:/Elmer/elmerfem-Copy/fem/src/SParIterComm.F90:5019:51: Looks like it may be related to a commit on 3 August 2023. Also, the last successful Windows binary installer build with 'no-mpi' is dated 3 August 2023. |
Looking in the file(s) mpif_stub.h, there is a macro for 'MPI_DOUBLE_COMPLEX', but not for 'MPI_COMPLEX'. Making that change in three places, then allows compilation to finish without an error. Whether that is the proper fix is another question. Rich. |
I think your fix is good, I can do the hounors on monday morning.
-Juha
From: "Rich_B" ***@***.***>
To: "ElmerCSC" ***@***.***>
Cc: "Subscribed" ***@***.***>
Sent: Saturday, 12 August, 2023 18:57:12
Subject: Re: [ElmerCSC/elmerfem] Error: Symbol 'mpi_complex' at (1) has no IMPLICIT type (Issue #415)
Looking in the file(s) mpif_stub.h, there is a macro for 'MPI_DOUBLE_COMPLEX', but not for 'MPI_COMPLEX'. Making that change in three places, then allows compilation to finish without an error. Whether that is the proper fix is another question.
Rich.
—
Reply to this email directly, [ #415 (comment) | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ACTOMSUBOAVKRMI7NLYTOBLXU6RVRANCNFSM6AAAAAA3N4UZLU | unsubscribe ] .
You are receiving this because you are subscribed to this thread. Message ID: ***@***.***>
|
The other way to fix this issue would be to add an entry for MPI_COMPLEX into mpif_stub.h. |
Thanx Rich! I just added the suggested fix. Hope it works. |
I will test on macOS now. |
@raback @richb2k Looks like we got another instance of the same problem:
|
Hi
My take is that the MPI datatypes should read
MPI_CHARACTER ---> MPI_INTEGER1
and
MPI_DOUBLE ---> MPI_DOUBLE_PRECISION
Maybe you could test ?
Br, Juha
From: "Sergey Fedorov" ***@***.***>
To: "ElmerCSC" ***@***.***>
Cc: "juharu" ***@***.***>, "Comment" ***@***.***>
Sent: Monday, 14 August, 2023 06:55:39
Subject: Re: [ElmerCSC/elmerfem] Error: Symbol 'mpi_complex' at (1) has no IMPLICIT type (Issue #415)
[ https://github.com/raback | @raback ] [ https://github.com/richb2k | @richb2k ] Looks like we got another instance of the same problem:
/opt/local/var/macports/build/_opt_PPCSnowLeopardPorts_science_elmerfem/elmerfem/work/elmerfem-79e6d51eb81b6316c87b816190c127fa34439271/fem/src/modules/SaveGridData.F90:1045:53:
1045 | CALL MPI_BCAST(WorkChar, 1, MPI_CHARACTER, 0, ELMER_COMM_WORLD, ierr)
| 1
Error: Symbol 'mpi_character' at (1) has no IMPLICIT type
/opt/local/var/macports/build/_opt_PPCSnowLeopardPorts_science_elmerfem/elmerfem/work/elmerfem-79e6d51eb81b6316c87b816190c127fa34439271/fem/src/modules/SaveGridData.F90:1169:60:
1169 | CALL MPI_REDUCE(Array,PArray,nx*ny*nz,MPI_DOUBLE,MPI_MAX,0,ELMER_COMM_WORLD, ierr)
| 1
Error: Symbol 'mpi_double' at (1) has no IMPLICIT type
—
Reply to this email directly, [ #415 (comment) | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ACTOMSRDROLJKSZSENGPA2LXVGOTXANCNFSM6AAAAAA3N4UZLU | unsubscribe ] .
You are receiving this because you commented. Message ID: ***@***.***>
|
@juharu Where to change those? Just replace names across the source? |
HI
I think these were isolated occurances in fem/modules/SaveGridDate.F90 only ?
Br, Juha
From: "Sergey Fedorov" ***@***.***>
To: "ElmerCSC" ***@***.***>
Cc: "juharu" ***@***.***>, "Mention" ***@***.***>
Sent: Monday, 14 August, 2023 13:29:57
Subject: Re: [ElmerCSC/elmerfem] Error: Symbol 'mpi_complex' at (1) has no IMPLICIT type (Issue #415)
[ https://github.com/juharu | @juharu ] Where to change those? Just replace names across the source?
—
Reply to this email directly, [ #415 (comment) | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ACTOMSUEP5AH7MDNUD5N34DXVH42LANCNFSM6AAAAAA3N4UZLU | unsubscribe ] .
You are receiving this because you were mentioned. Message ID: ***@***.***>
|
This is a new error, apparently. Build from master 2023.07.12 was okay; build from master 2023.08.10 is broken:
The text was updated successfully, but these errors were encountered: