search

MITgcm-contrib / llc_hires

Code customizations and scripts related to LLC4320 family of configurations.
16 stars 12 forks source link

what is the 'mpi_type'? #5

Open daruzi123 opened 3 years ago

daruzi123 commented 3 years ago

when I run MITgcm llc-4320 followed the ， the error is readtile_mpiio.c(49): error: expression must have integral type switch (mpi_type) { ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ I don't know what is the ‘MPI_type’?Does it have anything to do with supercomputers?

menemenlis commented 3 years ago

what instructions did you follow? what platform are you running on?

On Nov 23, 2020, at 1:08 AM, daruzi123 notifications@github.com wrote:

when I run MITgcm llc-4320 followed the https://github.com/MITgcm-contrib/llc_hires/blob/master/llc_4320/readme.txt， the error is readtile_mpiio.c(49): error: expression must have integral type switch (mpi_type) { ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^

readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ I don't know what is the ‘MPI_type’?Does it have anything to do with supercomputers?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/MITgcm-contrib/llc_hires/issues/5, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABH7CSPELWNBMOOKWY7YLHTSRIQ7XANCNFSM4T7GPGVA.

daruzi123 commented 3 years ago

what instructions did you follow? what platform are you running on? … On Nov 23, 2020, at 1:08 AM, daruzi123 @.***> wrote: when I run MITgcm llc-4320 followed the <[](https://github.com/MITgcm-contrib/llc_hires/blob/master/llc_4320/readme.txt)>， the error is readtile_mpiio.c(49): error: expression must have integral type switch (mpi_type) { ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ I don't know what is the ‘MPI_type’?Does it have anything to do with supercomputers? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#5>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABH7CSPELWNBMOOKWY7YLHTSRIQ7XANCNFSM4T7GPGVA.

The model was run on the supercomputer of IOCAS (Institute of Oceanology of the Chinese Academy of Sciences). when I executed the 'make' commond, these errors occurred.

menemenlis commented 3 years ago

Hi, these instructions are specific to pleiades supercomputer. You would need to modify instructions for IOCAS before you can run there. In particular, what genmake option file did you use to replace linux_amd64_ifort+mpi_ice_nas ?

Given that this is a compilation error, in particular, I suggest that you start with smaller problem, first one of the MPI verification experiments to make sure you are able to run MITgcm using MPI, then maybe try to run https://github.com/MITgcm-contrib/llc_hires/tree/master/llc_90 https://github.com/MITgcm-contrib/llc_hires/tree/master/llc_90 before you move to the 1/48-deg simulation.

Regards, Dimitris

On Nov 23, 2020, at 5:48 PM, daruzi123 notifications@github.com wrote:

what instructions did you follow? what platform are you running on? … On Nov 23, 2020, at 1:08 AM, daruzi123 @.***> wrote: when I run MITgcm llc-4320 followed the <>， the error is readtile_mpiio.c(49): error: expression must have integral type switch (mpi_type) { ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ I don't know what is the ‘MPI_type’?Does it have anything to do with supercomputers? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#5>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABH7CSPELWNBMOOKWY7YLHTSRIQ7XANCNFSM4T7GPGVA.

The model was run on the supercomputer of IOCAS (Institute of Oceanology of the Chinese Academy of Sciences). when I executed the 'make' commond, these errors occurred.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

daruzi123 commented 3 years ago

Hi, these instructions are specific to pleiades supercomputer. You would need to modify instructions for IOCAS before you can run there. In particular, what genmake option file did you use to replace linux_amd64_ifort+mpi_ice_nas ? Given that this is a compilation error, in particular, I suggest that you start with smaller problem, first one of the MPI verification experiments to make sure you are able to run MITgcm using MPI, then maybe try to run https://github.com/MITgcm-contrib/llc_hires/tree/master/llc_90 https://github.com/MITgcm-contrib/llc_hires/tree/master/llc_90 before you move to the 1/48-deg simulation. Regards, Dimitris … On Nov 23, 2020, at 5:48 PM, daruzi123 @.> wrote: what instructions did you follow? what platform are you running on? … On Nov 23, 2020, at 1:08 AM, daruzi123 @.> wrote: when I run MITgcm llc-4320 followed the <>， the error is readtile_mpiio.c(49): error: expression must have integral type switch (mpi_type) { ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(51): error: expression must be an integral constant expression case MPI_INT: case MPI_FLOAT: case MPI_REAL4: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ readtile_mpiio.c(55): error: expression must be an integral constant expression case MPI_LONG_INT: case MPI_DOUBLE: case MPI_REAL8: ^ I don't know what is the ‘MPI_type’?Does it have anything to do with supercomputers? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#5>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABH7CSPELWNBMOOKWY7YLHTSRIQ7XANCNFSM4T7GPGVA. The model was run on the supercomputer of IOCAS (Institute of Oceanology of the Chinese Academy of Sciences). when I executed the 'make' commond, these errors occurred. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Thanks. I used the 'linux_amd64_ifort' to instead the 'linux_amd64_ifort+mpi_ice_nas'. And I can use linux_amd64_ifort to build the MITgcm with MPI.