INT2LM Climate grid decomposition failed – in #10: INT2LM

in #10: INT2LM

<p> Dear colleagues, I have faced with a problem of “Climate grid decomposition failed” during int2lm run from <span class="caps"> ERA </span> -Interim forcing within the Russian Arctic domain. Since this domain intersects the North Pole, I have switched on the namelist north_add_in=1. However, the error indicates that </p> zmaxlat 89.9835303250248 &gt; 89.4628219604492 zminlon -179.707653101709 &lt; -178.593750000000 zmaxlon 179.784906162209 &gt; 177.890625000000 zmaxlat 89.9835303250248 &gt; 89.4628219604492 zmaxlon 179.784906162209 &gt; 177.890625000000 <p> I have surprised for this, because this script was successfully run earlier without such problems and issues. I have tried to chabge other ‘ <strong> _add_in </strong> namelists without any success. Could you get any hints, where could be the source of error? There are caf* <span class="caps"> ERA </span> -I data, or changes in the int2lm/namelists codes, or any simple errors from my side within script? </p> <p> Thank you for help and excuse me for any inconveniences! </p>

  @vladimirplatonov in #f82c233

<p> Dear colleagues, I have faced with a problem of “Climate grid decomposition failed” during int2lm run from <span class="caps"> ERA </span> -Interim forcing within the Russian Arctic domain. Since this domain intersects the North Pole, I have switched on the namelist north_add_in=1. However, the error indicates that </p> zmaxlat 89.9835303250248 &gt; 89.4628219604492 zminlon -179.707653101709 &lt; -178.593750000000 zmaxlon 179.784906162209 &gt; 177.890625000000 zmaxlat 89.9835303250248 &gt; 89.4628219604492 zmaxlon 179.784906162209 &gt; 177.890625000000 <p> I have surprised for this, because this script was successfully run earlier without such problems and issues. I have tried to chabge other ‘ <strong> _add_in </strong> namelists without any success. Could you get any hints, where could be the source of error? There are caf* <span class="caps"> ERA </span> -I data, or changes in the int2lm/namelists codes, or any simple errors from my side within script? </p> <p> Thank you for help and excuse me for any inconveniences! </p>

INT2LM Climate grid decomposition failed

VP wrote

Dear colleagues, I have faced with a problem of “Climate grid decomposition failed” during int2lm run from ERA -Interim forcing within the Russian Arctic domain. Since this domain intersects the North Pole, I have switched on the namelist north_add_in=1. However, the error indicates that

zmaxlat 89.9835303250248 > 89.4628219604492 zminlon -179.707653101709 < -178.593750000000 zmaxlon 179.784906162209 > 177.890625000000 zmaxlat 89.9835303250248 > 89.4628219604492 zmaxlon 179.784906162209 > 177.890625000000

I have surprised for this, because this script was successfully run earlier without such problems and issues. I have tried to chabge other ‘ _add_in namelists without any success. Could you get any hints, where could be the source of error? There are caf* ERA -I data, or changes in the int2lm/namelists codes, or any simple errors from my side within script?

Thank you for help and excuse me for any inconveniences!

View in channel
<p> Can you try whether the error appears also in the latest version int2lm2.09? https://redc.clm-community.eu/projects/int2lm/wiki/Downloads <br/> The parameters to set are: <br/> <pre> north_add_in=1,east_add_in=1,</pre> </p> <p> <br/> i.e. both have to be set! </p>

  @burkhardtrockel in #ed344b7

<p> Can you try whether the error appears also in the latest version int2lm2.09? https://redc.clm-community.eu/projects/int2lm/wiki/Downloads <br/> The parameters to set are: <br/> <pre> north_add_in=1,east_add_in=1,</pre> </p> <p> <br/> i.e. both have to be set! </p>
BR replied

Can you try whether the error appears also in the latest version int2lm2.09? https://redc.clm-community.eu/projects/int2lm/wiki/Downloads
The parameters to set are: 

  north_add_in=1,east_add_in=1,


i.e. both have to be set!

View in channel
<p> I have set both parameters <br/> north_add_in=1,east_add_in=1, <br/> but unfortunately without any success. <br/> Moreover, I can’t compile the binary for the int2lm2.09 with the same keys and options, as for the current int2lm version, I’m attaching error log and Fopts. I have changed some include paths, but the error log didn’t change. </p>

  @vladimirplatonov in #304b38a

<p> I have set both parameters <br/> north_add_in=1,east_add_in=1, <br/> but unfortunately without any success. <br/> Moreover, I can’t compile the binary for the int2lm2.09 with the same keys and options, as for the current int2lm version, I’m attaching error log and Fopts. I have changed some include paths, but the error log didn’t change. </p>
VP replied

I have set both parameters
north_add_in=1,east_add_in=1,
but unfortunately without any success.
Moreover, I can’t compile the binary for the int2lm2.09 with the same keys and options, as for the current int2lm version, I’m attaching error log and Fopts. I have changed some include paths, but the error log didn’t change.

View in channel
<p> For me it work after setting <br/> <pre><br/> CPPFLAGS = -<span class="caps">DNETCDF</span> -D__INT2LM__ </pre> </p> <p> </p>

  @burkhardtrockel in #3b14a69

<p> For me it work after setting <br/> <pre><br/> CPPFLAGS = -<span class="caps">DNETCDF</span> -D__INT2LM__ </pre> </p> <p> </p>
BR replied

For me it work after setting 



CPPFLAGS     = -DNETCDF -D__INT2LM__

View in channel
<p> Dear Burkhardt, thanks. Unfortunately, adding this line (as well as <span class="caps"> CPPFLAGS </span> , as <span class="caps"> PRAGMA </span> _LIB) get much more errors including netcdf library. Three options -D__INT2LM__ - <span class="caps"> DNETCDF </span> - <span class="caps"> DGRIBDWD </span> get the same result. The original <span class="caps"> PRAGMA </span> _LIB options were - <span class="caps"> DGRIBDWD </span> - <span class="caps"> DNETCDF </span> compiled without netcdf errors. </p>

  @vladimirplatonov in #ce7370e

<p> Dear Burkhardt, thanks. Unfortunately, adding this line (as well as <span class="caps"> CPPFLAGS </span> , as <span class="caps"> PRAGMA </span> _LIB) get much more errors including netcdf library. Three options -D__INT2LM__ - <span class="caps"> DNETCDF </span> - <span class="caps"> DGRIBDWD </span> get the same result. The original <span class="caps"> PRAGMA </span> _LIB options were - <span class="caps"> DGRIBDWD </span> - <span class="caps"> DNETCDF </span> compiled without netcdf errors. </p>
VP replied

Dear Burkhardt, thanks. Unfortunately, adding this line (as well as CPPFLAGS , as PRAGMA _LIB) get much more errors including netcdf library. Three options -D__INT2LM__ - DNETCDF - DGRIBDWD get the same result. The original PRAGMA _LIB options were - DGRIBDWD - DNETCDF compiled without netcdf errors.

View in channel
<p> Dear Burkhardt and colleagues, I have tested different versions of int2lm for this case. I have established that versions int2lm_2.05 and below (int2lm_2.04) compiled successfully, however it crashed during run with the abovementioned error (Climate grid decomposition failed). Higher int2lm versions (2.07, 2.08 and 2.09) have not been compiled because of similar errors provided in the make_err file attached in my previous post: </p> <p> ——————^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(252): error #7002: Error in opening the compiled module file. Check <span class="caps"> INCLUDE </span> paths. [DATA_CONSTANTS] <br/> USE data_constants , <span class="caps"> ONLY </span> : &amp; <br/> ——^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(253): error #6580: Name in only-list does not exist. [REPSILON] repsilon <br/> ——^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(3629): error #6404: This name does not have a type, and must have an explicit type. [REPSILON] IF( <acronym title="Z_VEC(i"> <span class="caps"> ABS </span> </acronym> ) &lt; repsilon) THEN <br/> ——————————————————-^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(3655): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- <span class="caps"> ALPHA </span> _VEC(i))/ <acronym title="DEL*ZT2_VEC(i"> <span class="caps"> MAX </span> </acronym> ,repsilon) <br/> ————————————————————————————————————————^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(4242): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- <span class="caps"> ALPHA </span> _VEC(i,j))/ <acronym title="DEL*ZT2_VEC(i,j"> <span class="caps"> MAX </span> </acronym> ,repsilon) </p> <p> I have tried different Fortran (intel/15.0.3, intel/2019.5) and OpenMPI (openmpi/1.8.4-icc, openmpi/4.1.0-icc) compiler versions without any changes for these last int2lm versions. Therefore, I have two points of possible errors elimination, the 1st one is any adoptations of run script for earlier versions associated with north_add_in=1,east_add_in=1, namelist; the 2nd one is to adopt compiling or to edit somehow the utilities.f90 file in the int2lm code for the latest versions. <br/> I would be grateful for any hints in this case! </p>

  @vladimirplatonov in #87f2674

<p> Dear Burkhardt and colleagues, I have tested different versions of int2lm for this case. I have established that versions int2lm_2.05 and below (int2lm_2.04) compiled successfully, however it crashed during run with the abovementioned error (Climate grid decomposition failed). Higher int2lm versions (2.07, 2.08 and 2.09) have not been compiled because of similar errors provided in the make_err file attached in my previous post: </p> <p> ——————^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(252): error #7002: Error in opening the compiled module file. Check <span class="caps"> INCLUDE </span> paths. [DATA_CONSTANTS] <br/> USE data_constants , <span class="caps"> ONLY </span> : &amp; <br/> ——^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(253): error #6580: Name in only-list does not exist. [REPSILON] repsilon <br/> ——^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(3629): error #6404: This name does not have a type, and must have an explicit type. [REPSILON] IF( <acronym title="Z_VEC(i"> <span class="caps"> ABS </span> </acronym> ) &lt; repsilon) THEN <br/> ——————————————————-^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(3655): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- <span class="caps"> ALPHA </span> _VEC(i))/ <acronym title="DEL*ZT2_VEC(i"> <span class="caps"> MAX </span> </acronym> ,repsilon) <br/> ————————————————————————————————————————^ <br/> /home/vplatonov/COSMO- <span class="caps"> CLM </span> /int2lm-int2lm-2.09/src/utilities.f90(4242): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- <span class="caps"> ALPHA </span> _VEC(i,j))/ <acronym title="DEL*ZT2_VEC(i,j"> <span class="caps"> MAX </span> </acronym> ,repsilon) </p> <p> I have tried different Fortran (intel/15.0.3, intel/2019.5) and OpenMPI (openmpi/1.8.4-icc, openmpi/4.1.0-icc) compiler versions without any changes for these last int2lm versions. Therefore, I have two points of possible errors elimination, the 1st one is any adoptations of run script for earlier versions associated with north_add_in=1,east_add_in=1, namelist; the 2nd one is to adopt compiling or to edit somehow the utilities.f90 file in the int2lm code for the latest versions. <br/> I would be grateful for any hints in this case! </p>
VP replied

Dear Burkhardt and colleagues, I have tested different versions of int2lm for this case. I have established that versions int2lm_2.05 and below (int2lm_2.04) compiled successfully, however it crashed during run with the abovementioned error (Climate grid decomposition failed). Higher int2lm versions (2.07, 2.08 and 2.09) have not been compiled because of similar errors provided in the make_err file attached in my previous post:

——————^
/home/vplatonov/COSMO- CLM /int2lm-int2lm-2.09/src/utilities.f90(252): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [DATA_CONSTANTS]
USE data_constants , ONLY : &
——^
/home/vplatonov/COSMO- CLM /int2lm-int2lm-2.09/src/utilities.f90(253): error #6580: Name in only-list does not exist. [REPSILON] repsilon
——^
/home/vplatonov/COSMO- CLM /int2lm-int2lm-2.09/src/utilities.f90(3629): error #6404: This name does not have a type, and must have an explicit type. [REPSILON] IF( ABS ) < repsilon) THEN
——————————————————-^
/home/vplatonov/COSMO- CLM /int2lm-int2lm-2.09/src/utilities.f90(3655): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- ALPHA _VEC(i))/ MAX ,repsilon)
————————————————————————————————————————^
/home/vplatonov/COSMO- CLM /int2lm-int2lm-2.09/src/utilities.f90(4242): error #6363: The intrinsic data types of the arguments must be the same. [MAX] (1.0_wp- ALPHA _VEC(i,j))/ MAX ,repsilon)

I have tried different Fortran (intel/15.0.3, intel/2019.5) and OpenMPI (openmpi/1.8.4-icc, openmpi/4.1.0-icc) compiler versions without any changes for these last int2lm versions. Therefore, I have two points of possible errors elimination, the 1st one is any adoptations of run script for earlier versions associated with north_add_in=1,east_add_in=1, namelist; the 2nd one is to adopt compiling or to edit somehow the utilities.f90 file in the int2lm code for the latest versions.
I would be grateful for any hints in this case!

View in channel
<p> Dear colleagues, I have tested all working int2lm versions (2.04 and 2.05) for other domains without pole transection and, therefore, without ‘north_add_in=1,east_add_in=1,’ namelists. All these runs worked well. Therefore, I can conclude the problem is in the <span class="caps"> GRID </span> _IN namelists, maybe specific for the int2lm 2.04 and 2.05 versions. However, this script worked well a feew months before. The other possible issue could be changes in the <span class="caps"> ERA </span> -I forcing data prepared on the <span class="caps"> DKRZ </span> if any, because I have dowloaded them recently. <br/> So, if anybody could suggest any solutions or other sources of problem, or knows any useful information about the cited issues, I would be very grateful for any hints! Thank you very much! </p>

  @vladimirplatonov in #1dac45b

<p> Dear colleagues, I have tested all working int2lm versions (2.04 and 2.05) for other domains without pole transection and, therefore, without ‘north_add_in=1,east_add_in=1,’ namelists. All these runs worked well. Therefore, I can conclude the problem is in the <span class="caps"> GRID </span> _IN namelists, maybe specific for the int2lm 2.04 and 2.05 versions. However, this script worked well a feew months before. The other possible issue could be changes in the <span class="caps"> ERA </span> -I forcing data prepared on the <span class="caps"> DKRZ </span> if any, because I have dowloaded them recently. <br/> So, if anybody could suggest any solutions or other sources of problem, or knows any useful information about the cited issues, I would be very grateful for any hints! Thank you very much! </p>
VP replied

Dear colleagues, I have tested all working int2lm versions (2.04 and 2.05) for other domains without pole transection and, therefore, without ‘north_add_in=1,east_add_in=1,’ namelists. All these runs worked well. Therefore, I can conclude the problem is in the GRID _IN namelists, maybe specific for the int2lm 2.04 and 2.05 versions. However, this script worked well a feew months before. The other possible issue could be changes in the ERA -I forcing data prepared on the DKRZ if any, because I have dowloaded them recently.
So, if anybody could suggest any solutions or other sources of problem, or knows any useful information about the cited issues, I would be very grateful for any hints! Thank you very much!

View in channel
<p> I tried int2lm2.05_clm2 and int2lm2.09. Both run fine. See the <span class="caps"> INPUT </span> file for the settings. </p>

  @burkhardtrockel in #f62435f

<p> I tried int2lm2.05_clm2 and int2lm2.09. Both run fine. See the <span class="caps"> INPUT </span> file for the settings. </p>
BR replied

I tried int2lm2.05_clm2 and int2lm2.09. Both run fine. See the INPUT file for the settings.

View in channel
<p> Dear Burkhardt, thanks a lot for providing <span class="caps"> INPUT </span> file example. I have tried and adapted most of namelists provided in your <span class="caps"> INPUT </span> file. Unfortunately, there are no changes in int2lm crash… I’m quite confused for these results, and don’t have any explanations more… Maybe these errors could be caused by any int2lm compilation options affecting the north_add_in and east_add_in namelist group specifically? Or there are any other ideas about causes of these errors? Thank you very much! </p>

  @vladimirplatonov in #46cb52f

<p> Dear Burkhardt, thanks a lot for providing <span class="caps"> INPUT </span> file example. I have tried and adapted most of namelists provided in your <span class="caps"> INPUT </span> file. Unfortunately, there are no changes in int2lm crash… I’m quite confused for these results, and don’t have any explanations more… Maybe these errors could be caused by any int2lm compilation options affecting the north_add_in and east_add_in namelist group specifically? Or there are any other ideas about causes of these errors? Thank you very much! </p>
VP replied

Dear Burkhardt, thanks a lot for providing INPUT file example. I have tried and adapted most of namelists provided in your INPUT file. Unfortunately, there are no changes in int2lm crash… I’m quite confused for these results, and don’t have any explanations more… Maybe these errors could be caused by any int2lm compilation options affecting the north_add_in and east_add_in namelist group specifically? Or there are any other ideas about causes of these errors? Thank you very much!

View in channel