Hello, I encountered different errors when compiling under different environments.

First compilation:
Loaded environment:
module load cmake/3.28
module load intel/2020u1
module load gcc/14.2.0
export CC=which gcc
export CXX=which g++
Error message:
[sx2321035@login1 testmpi]$ make
[ 1%] Building Fortran object src/komega/CMakeFiles/komega.dir/komega_vals.F90.o
[ 2%] Building Fortran object src/komega/CMakeFiles/komega.dir/komega_math.F90.o
f951: Fatal Error: Reading module ‘/fs0/software/intel/2020/compilers_and_libraries_2020.0.139/linux/mpi/intel64/include/mpi.mod’
at line 1 column 2: Unexpected EOF compilation terminated.
make[2]: *** [src/komega/CMakeFiles/komega.dir/komega_math.F90.o] Error 1
make[1]: *** [src/komega/CMakeFiles/komega.dir/all] Error 2
make: *** [all] Error 2

Second compilation:
Loaded environment:
module load cmake/3.28
module load intel/2020u1
module load gcc/14.2.0
export CC=which gcc
export CXX=which g++
export FC=ifort
export CC=icc
Error message:
[ 10%] Building C object src/CMakeFiles/HPhi.dir/check.c.o
/fs0/home/sx2321035/testmpi/HPhi-3.5.2/src/check.c(187): error: "idimmax" must be specified in a variable list at enclosing
OpenMP parallel pragma
#pragma omp parallel for default(none) reduction(+:comb_sum) private(tmp_sz, isite) firstprivate(idimmax, X)
^
compilation aborted for /fs0/home/sx2321035/testmpi/HPhi-3.5.2/src/check.c (code 2)
make[2]: *** [src/CMakeFiles/HPhi.dir/check.c.o] Error 2
make[1]: *** [src/CMakeFiles/HPhi.dir/all] Error 2
make: *** [all] Error 2

Third compilation:
Loaded environment:
module load cmake/3.28
module load intel/2019u3
module load gcc/14.2.0
export CC=which gcc
export CXX=which g++
export FC=ifort
export CC=icc

This time, the compilation succeeded.

[tmisawa]

The error arises because default(none) in OpenMP requires all variables used within the parallel region
to be explicitly declared with appropriate data-sharing attributes.
In this case, the loop variable idim is not declared as private, which may cause a compiler error.
Although idim is part of the for loop header,
it must still be explicitly declared when using default(none).
Thus, I think compilation will succeed if you use private (idim, tmp_sz, isite) in line 187 in check.c.

This stricter behavior may be due to changes introduced in Intel compiler version 2020u1,
which enforces OpenMP specifications more rigorously than previous versions.
Using Intel 2019u3 avoids this issue, as the earlier compiler is more permissive.
However, future releases of HPhi will include fixes to ensure compatibility with Intel 2020u1
and later compilers by properly specifying variable scopes in OpenMP directives.

Also, for your information, export CC=which gcc should be replaced with export CC=$(which gcc)
to correctly assign the path to the gcc compiler.

[tmisawa]

I noticed that you may use `which gcc`. This is a correct description.

Thank you very much for your help. I have another question: when I perform finite-temperature calculations, even if I don't compute the two-body correlation functions, the program still outputs a large number of files named zvo_cisajscktalt_setAstepB.dat. Although these files are all empty, they don't look very tidy.

[tmisawa]

We would prefer to keep each issue focused on a single topic.
So, if possible, could you open a separate issue for this question?
Also, please keep the issue titles as simple as possible—there’s no need to include "Hello" in the title.
For this issue, "Compiling problems under different environments" is good.