Support Intel oneAPI 2025 with configure

[adam-sim-dev]

Since Intel oneAPI 2025, ifort and mpiifort are removed. So simple ./configure won't work because it fails to find the fortran compiler.

17fa100 should be updated by simply moving ifx(mpiifx) ahead of ifort (mpiifort).

hypre/src/configure

Line 5683 in c49433c

for ac_prog in xlf_r ifort ifx gfortran g77 g95 pgf77 pgfortran f77

hypre/src/configure

Line 5731 in c49433c

for ac_prog in xlf ifort ifx gfortran g77 g95 pgf77 pgfortran f77

hypre/src/configure

Line 5782 in c49433c

for ac_prog in mpxlf mpixlf77_r mpiifort mpiifx mpif77 mpipgf77 mpipgifort

hypre/src/configure

Line 5830 in c49433c

for ac_prog in mpxlf mpixlf77 mpiifort mpiifx mpif77 mpipgf77 mpipgifort

config.log

[adam-sim-dev]

I am not clear why ifx is not present in the following three lines and if it has some effects, while the above fix works for me.

hypre/src/configure

Line 4090 in c49433c

for ac_prog in g77 ifort pgf77 xlf

hypre/src/configure

Line 7799 in c49433c

for ac_prog in gfortran g95 xlf95 f95 fort ifort ifc efc pgfortran pgf95 lf95 ftn nagfor xlf90 f90 pgf90 pghpf epcf90 g77 xlf f77 frt pgf77 cf77 fort77 fl32 af77

hypre/src/configure

Line 7848 in c49433c

for ac_prog in gfortran g95 xlf95 f95 fort ifort ifc efc pgfortran pgf95 lf95 ftn nagfor xlf90 f90 pgf90 pghpf epcf90 g77 xlf f77 frt pgf77 cf77 fort77 fl32 af77

[ivikiwi-1111]

Faced the same issues. Replacing ifort to ifx in the configure file seems to work for me.

[adam-sim-dev]

Faced the same issues. Replacing ifort to ifx in the configure file seems to work for me.

See the PR #1193

[ivikiwi-1111]

Did you succeed further? I've tried to compile with following options:

--with-sycl --with-sycl-target=amdgcn-amd-amdhsa --enable-unified-memory

Getting a bunch of errors. Tried to checkout to sycl-cleanup MR, didn't help. Not sure if I'm doing something wrong or the packages are broken for some reason.

[waynemitchell]

Did you succeed further? I've tried to compile with following options:

--with-sycl --with-sycl-target=amdgcn-amd-amdhsa --enable-unified-memory

Getting a bunch of errors. Tried to checkout to sycl-cleanup MR, didn't help. Not sure if I'm doing something wrong or the packages are broken for some reason.

Can you share some of the issues you are having? The master branch should work (the sycl cleanup PR was merged a while ago, so that branch is old).

[ivikiwi-1111]

My setup is:

OS: Ubuntu 24.04.1 LTS x86_64
Tool-Kit: Intel oneAPI 2025
Options: --with-sycl --with-sycl-target=amdgcn-amd-amdhsa --enable-unified-memory

With fixed configure and the options above after running make I get a lot of similar errors like this one:

icpx  -O3 -fsycl -fsycl-unnamed-lambda  -DHAVE_CONFIG_H -I.. -I./.. -I./../struct_mv -I.    -qmkl -I/include -DMKL_ILP64 -I/opt/intel/oneapi/mkl/2025.0/include        -c general.c -o general.obj
icpx: warning: treating 'c' input as 'c++' when -fsycl is used [-Wexpected-file-type]
In file included from general.c:8:
In file included from ./_hypre_utilities.h:648:
In file included from /opt/intel/oneapi/compiler/2025.0/bin/compiler/../../include/sycl/stl_wrappers/assert.h:20:
In file included from /opt/intel/oneapi/compiler/2025.0/bin/compiler/../../include/sycl/CL/__spirv/spirv_vars.hpp:27:
/opt/intel/oneapi/compiler/2025.0/bin/compiler/../../include/sycl/CL/__spirv/spirv_types.hpp:135:1: error: templates must have C++ linkage
  135 | template <typename T, std::size_t R, std::size_t C, MatrixLayout L,
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  136 |           Scope::Flag S = Scope::Flag::Subgroup,
      |           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  137 |           MatrixUse U = MatrixUse::MatrixA>
      |           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./_hypre_utilities.h:14:1: note: extern "C" language linkage specification begins here
   14 | extern "C" {
      | ^
      

Seems like *.c files are preferably compiled with icx? Also it's probably better to move all this into a new issue?

[waynemitchell]

Looks like you are running into the same issue reported here: #1191

I'll address this in a separate PR.

[waynemitchell]

Can you try the branch here: #1197