Umfpack support for local devices (hostnames DARWIN and LINUX) and IPP TOK cluster (ifort and gfortran)

[antonellozito]

Hi @WimVanUytven,
I have successfully managed to compile the code on local devices (hostnames DARWIN and LINUX, using gfortran) with umfpack. As expected, it was quite straightforward. With this, I add the required modifications to your branch, waiting for your branch to be merged to the main one.
Accordingly, I have also added the automatic installation of the library, for both cases, to the install_dependencies script.
I wanted to do it also on the IPP TOK cluster, for both ifort and gfortran, but apparently SuiteSparse is not installed there... @DavidPCoster could you maybe have a look at it? Again, after installing SuiteSparse on the cluster, the linking should be pretty straightforward.

EDIT: implemented successfully now also on IPP TOK cluster, for both ifort and gfortran cases.

[DavidPCoster]

@antonellozito -- On the TOK cluster:

17:22> module avail suite
----------------------------------- /toks/scratch/dpc/easybuild/modules/all ------------------------------------
SuiteSparse/7.7.0-foss-2023b SuiteSparse/7.7.0-iimkl-2023b
SuiteSparse/7.7.0-gfbf-2023b SuiteSparse/7.7.0-intel-2023b

Key:
modulepath
17:22>

[antonellozito]

@antonellozito -- On the TOK cluster:

17:22> module avail suite ----------------------------------- /toks/scratch/dpc/easybuild/modules/all ------------------------------------ SuiteSparse/7.7.0-foss-2023b SuiteSparse/7.7.0-iimkl-2023b SuiteSparse/7.7.0-gfbf-2023b SuiteSparse/7.7.0-intel-2023b

Key: modulepath 17:22>

Ah yes, true... it was visible only after loading all the others. Implemented and working now.

[hrdoktorx]

Reminder to @antonellozito to fix this PR so the CI tests pass.

[antonellozito]

I do not have a working IO account anymore, so I can't see what was the problem... I would assume that whatever broke the CI tests might have been more upstream, since here I just added the options to have SuiteSparse with the IPP/LINUX/DARWIN hosts, without touching any source code nor the ITER setup files.
@WimVanUytven would it be ok for you to implement these changes (same for these) in your SuiteSparse branch(es), so that the IPP/LINUX/DARWIN configuration will be already ready in the moment you will decide to create the PRs towards the wg-release branches?

(FYI, I am also testing the Umpack solver option, and I have not seen anything bad now, everything seems to work as it should. So @WimVanUytven you might also consider to prepare the general PRs).

[WimVanUytven]

I will try to have a look at this in the coming days.