Skip to content
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

Issue with HPC build #344

Open
jemcquillan opened this issue Mar 8, 2022 · 6 comments
Open

Issue with HPC build #344

jemcquillan opened this issue Mar 8, 2022 · 6 comments

Comments

@jemcquillan
Copy link

Hello,

I just had the HPC guys install/build/make ED2 on our cluster. We have two versions installed. One is the single run, the other is the mpi version. I went to check to see if the build worked. I followed the ED-2.2_StartKit files for validation. Following the readme with all the proper paths set for the ED2IN file and the Santarem_Km83_HEADER path changed.

I loaded in the module and game the -f flag to point the software to the updated ED2IN file. This is the error I received. Path here has been edited for privacy.

`ed2 -f /the path to the startkit/ED-2.2_StartKit/Simulations/S0001_SantaremKm83_Test/ED2IN
+---------------- MPI parallel info: --------------------+

    • Machnum = 0
    • Machsize = 1
      +---------------- OMP parallel info: --------------------+
    • thread use: 8
    • threads max: 8
    • cpu use: 4
    • cpus max: 8
  • Note: Max vals are for node, not sockets.
    +--------------------------------------------------------+
    Reading namelist information
    forrtl: severe (19): invalid reference to variable in NAMELIST input, unit 10, file /the path to the startkit/ED-2.2_StartKit/Simulations/S0001_SantaremKm83_Test/ED2IN, line 514, position 17
    Image PC Routine Line Source
    ed2 0000000001024D8B Unknown Unknown Unknown
    ed2 000000000105827B Unknown Unknown Unknown
    ed_2.2-opt 00000000006B4000 read_nl_ 28 ed_load_namelist.f90
    ed_2.2-opt 000000000040F8F0 ed_1st_master_ 70 ed_1st.F90
    ed_2.2-opt 000000000040F620 MAIN__ 274 edmain.F90
    ed_2.2-opt 000000000040E8E2 Unknown Unknown Unknown
    libc-2.28.so 00007F9DEF208493 __libc_start_main Unknown Unknown
    ed_2.2-opt 000000000040E7EE Unknown Unknown Unknown`

The line in the ED2IN is below:

!---------------------------------------------------------------------------------------! ! ISLCOLFLG -- This controls how to initialise soil colour. This must be a list with ! ! N_ED_REGION+N_POI elements. The first N_ED_REGION elements correspond ! ! to each gridded domain (from first to last). Elements between ! ! N_ED_REGION+1 and N_ED_REGION+N_POI correspond to the polygons of ! ! interest (from 1 to N_POI. Options are: ! ! 1 -- Read in soil colour class from the files set in SLCOL_DATABASE. ! ! 2 -- Assign either the value set by ISOILCOL (see below). ! !---------------------------------------------------------------------------------------! NL%ISLCOLFLG = 2 !---------------------------------------------------------------------------------------!

Specifically, the space separating NL%ISLCOLFLG and the equals sign.

Since the HPC guys installed it for me I am not sure exactly how and where everything is for the build. They are really good in my experience. I am just at a loss as to troubleshoot this message.

Any guidance would be appreciated, and I can add anything more information that might be needed.

Best,
jemcquillan

@jemcquillan jemcquillan changed the title Issue with HPC bulid Issue with HPC build Mar 9, 2022
@xiangtaoxu
Copy link
Contributor

Seems like a ED2IN formatting issue or you have not installed the most recent version. Can you share the ED2IN you used? Maybe the flag before/after ISLCOLFLG got corrupted

@jemcquillan
Copy link
Author

jemcquillan commented Mar 9, 2022

Thank you for getting back to me. Hopefully it's an easy fix. It should be the most up-to-date version. Pulled from the GitHub on Monday.

The ED2IN comes from this path in the StartKit -

'.../ED-2.2_StartKit/Simulations/S0001_SantaremKm83_Test/ED2IN'

The flag was kept default. Here I was just validating the module. The flag was set to = 2.

I attached the ED2IN file below.

Thank you,
jemcquillan

ED2IN.zip

P.S. The two builds installed were -
Single-run: ed2/2.2.0/intel/ED/build/ed2
MPI-run: ed2/2.2.0/mpi/ED/build/ed2

@xiangtaoxu
Copy link
Contributor

I looked over your ED2IN, nothing suspicious comes up at my first glance... Not sure why ISLCOLFLG throws an error.

Can you (1) check the git commit number by running git log in your ED folder to see whether you are using the most up-to-date version? and (2) build the model using KIND=C to enable more debugging output. and rerun your script.

Also, make sure you are using the right executable. In my experience, the compiled executable should has a name of ed2.2-opt. DId you rename it to ed2?

@mpaiao
Copy link
Contributor

mpaiao commented Mar 12, 2022

@jemcquillan it may be worth comparing the ED2IN you are using with the ED2IN available in ED/run/ED2IN. It could be that there are new variables in the most up-to-date version and the start kit is not up-to-date.

@xiangtaoxu
Copy link
Contributor

@mpaiao I just did a diff. Nothing bad showed up. Actually, the ED/run/ED2IN has an extra line for NL%SL_SCALE

ED2/ED/run/ED2IN

Lines 1590 to 1592 in fa80dab

NL%IANTH_DISTURB = 0
NL%SL_SCALE = 1
NL%SL_SCALE = 0

@jemcquillan
Copy link
Author

jemcquillan commented Mar 15, 2022

Thank you all for your help. I am back on it.

A few things. I cannot find the commit. I searched for 'commit' in the install_urc-intel.out. Nothing. But again it would have come from the GitHub project last Monday.

This was not installed by me but by the HPC staff here. So some questions I just don't have permission to access. However, I did grab the intel build install log and have it attached here. I can do the same with the mpi build (but currently this thread is using the single run build.)
install_log.out.zip

I switched the ED2IN from he startkit to the ED/run/ED2IN from the project page, here on GitHub. I updated my paths and ran it. Oddly enough, I get the exact same error, just the line changes (since the files are different.)

ed2 -f /.../ED-2.2_StartKit/Simulations/S0001_SantaremKm83_Test/ED2IN

`
+---------------- MPI parallel info: --------------------+

    • Machnum = 0
    • Machsize = 1
      +---------------- OMP parallel info: --------------------+
    • thread use: 8
    • threads max: 8
    • cpu use: 8
    • cpus max: 8
  • Note: Max vals are for node, not sockets.
    +--------------------------------------------------------+
    Reading namelist information
    forrtl: severe (19): invalid reference to variable in NAMELIST input, unit 10, file /.../ED-2.2_StartKit/Simulations/S0001_SantaremKm83_Test/ED2IN, line 546, position 17
    Image PC Routine Line Source
    ed2 0000000001024D8B Unknown Unknown Unknown
    ed2 000000000105827B Unknown Unknown Unknown
    ed_2.2-opt 00000000006B4000 read_nl_ 28 ed_load_namelist.f90
    ed_2.2-opt 000000000040F8F0 ed_1st_master_ 70 ed_1st.F90
    ed_2.2-opt 000000000040F620 MAIN__ 274 edmain.F90
    ed_2.2-opt 000000000040E8E2 Unknown Unknown Unknown
    libc-2.28.so 00007F88B3B28493 __libc_start_main Unknown Unknown
    ed_2.2-opt 000000000040E7EE Unknown Unknown Unknown
    `

This is again this line:
NL%ISLCOLFLG = 2

I hope the install log helps.

Thank you for your time,
jemcquillan

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants