[FLASH-USERS] Compilation Errors in Debug Mode (v4.8, 4.7, and 4.2.2)

Ryan Farber rjfarber at umich.edu
Tue Jul 23 00:20:36 EDT 2024


Hi FLASH-users,

I am currently debugging a particularly pesky bug.

Sometime post FLASH4.2.2 (at least since FLASH4.7) compiling in debug mode
throws a number of errors (probably more related to updates in compilers
than updates in FLASH I guess). A student of mine patched those (for our
local version of FLASH4.7), but they're still present in a fresh download
of FLASH4.8 (as well as FLASH4.7.1).

Since I am currently working on patching these errors (~again), I wanted to
ask if there is a recommended way to submit a patch for the next version of
FLASH?

More info to possibly reproduce (if interested)

Setup line:
./setup Sedov -auto -2d -debug

make compilation line (though specifically errors are thrown due to "-warn
all" and "-warn error" from what I recall)
/dss/lrzsys/sys/spack/release/23.1.0/opt/x86_64/intel-oneapi-mpi/2021.9.0-gcc-zqpp4cs/mpi/2021.9.0/bin/mpiifort
-DH5Acreate_vers=1 -DH5Aiterate_vers=1 -DH5Dcreate_vers=1 -DH5Dopen_vers=1
-DH5Eclear_vers=1 -DH5Eprint_vers=1 -DH5Epush_vers=1 -DH5Eset_auto_vers=1
-DH5Eget_auto_vers=1 -DH5Ewalk_vers=1 -DH5Gcreate_vers=1 -DH5Gopen_vers=1
-DH5Pget_filter_vers=1 -DH5Pget_filter_by_id_vers=1 -DH5Pinsert_vers=1
-DH5Pregister_vers=1 -DH5Rget_obj_typevers=1 -DH5Tarray_create_vers=1
-DH5Tcommit_vers=1 -DH5Tget_array_dims_vers=1 -DH5Topen_vers=1
-mcmodel=large -c -g -r8 -real_size 64 -i4 -O0 -check bounds -check format
-check output_conversion -warn all -warn error -warn noexternal -real_size
64 -check uninit -traceback -fp-stack-check -diag-disable 10120 -fpe0
-check pointers  -DMAXBLOCKS=1000 -DNXB=8 -DNYB=8 -DNZB=1 -DN_DIM=2
NAME_OF_FILE.F90

Here is one such error (particularly annoying since I'm not using the
Plasma module I don't think with this default Sedov problem...)

PlasmaState_logLambda.F90(39): error #6843: A dummy argument with an
explicit INTENT(OUT) declaration is not given an explicit value.   [LL]

Leaving out "-warn error" generates the following error (potentially
important since my pesky bug occurs during MPI communication)

amr_migrate_tree_data.F90(601): error #8284: If the actual argument is
scalar, the dummy argument shall be scalar unless the actual argument is of
type character or is an element of an array that is not assumed shape,
pointer, or polymorphic.   [BUF]
                       Call MPI_int_SSEND (new_loc(1,i),2,MPI_INTEGER,
     &

See below for intel etc. versions (most relevant is probably intel-mpi,
rest are probably irrelevant).

 1) admin/1.0     3) lrz/1.0        5) intel/2023.1.0       7)
hdf5/1.8.23-intel23-impi   9) libszip/2.1.1
 2) tempdir/1.0   4) spack/23.1.0   6) intel-mpi/2021.9.0   8) zlib/1.2.13

Best wishes,
--------
Ryan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20240722/5c6453ae/attachment.htm>


More information about the flash-users mailing list