[Wien] ifx status?

[Gavin Abo]

Some additional comments below.

i) -standard-semantics: If Prof. Marks code fix at [1] is applied, then 
that compiler option should not be used.  In the post at [2], I had used 
it as a workaround before that fix was available and that option caused 
WIEN2k to not compile with libxc.  Whereas, the libxc could be used with 
the fix.  Of note, in addition to the fix (charge.patch), I also had 
angle.patch, atom_read.patch, SearchZ.patch, and x_lapw.patch at [3] 
applied before I compiled.

ii) fftw: I was able to compiled it simply (not threaded) with:

username at computername:~$ cd fftw-3.3.10/
username at computername:~/fftw-3.3.10$ ./configure FCC=ifx CC=icx 
MPICC=mpiicx CFLAGS="-gcc-sys" --enable-mpi --prefix=$HOME/fftw-3.3.10
...
username at computername:~/fftw-3.3.10$ make
...
username at computername:~/fftw-3.3.10$ make install

I recall Michael wanting a threaded fftw version and after much effort 
found a set of compiler options that seemed to be the ones at [4].

iii) WIEN2k compile options:  When I ran siteconfig for using ifx, I 
selected the "LI   Linux (Intel ifort compiler (12.0 or 
later)+mkl+intelmpi))".  In the recommended compile options for that, it 
has "-O".  I compiled WIEN2k with ifx version 2025.0.4. According to the 
2024.0 Intel documentation at [5], it would default to "-O2" in my case.

I left all the recommend compiler options that were there for ifort 
without any changes, but used ifx (and icx instead of icc) for the 
compiler,  and it compile fine for me (of note, I was using the patches 
[3]).  I've been using WIEN2k without encountering any issue so far.

Though, I do recall Prof. Fecher preferring to use a set of ifx compiler 
options (with -O2 or -O3) different from that at [6].

[1] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23572.html
[2] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23567.html
[3] https://github.com/gsabo/WIEN2k-Patches/tree/master/24.1
[4] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23584.html
[5] 
https://www.intel.com/content/www/us/en/docs/fortran-compiler/developer-guide-reference/2025-0/o-001.html
[6] 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23579.html

Kind Regards,

Gavin
WIEN2k user

On 3/3/2025 10:25 AM, Peter Blaha wrote:
> I did not install myself with the newest ifx, but from the mailing 
> list comments (sometimes they were a bit unclear) it seems:
>
> i) add  -standard-semantics to the compiler options.
>
> ii) fftw needs a "hack" as described in:
> https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23599.html 
>
>
> Peter
>
> Am 03.03.2025 um 16:01 schrieb Laurence Marks:
>> I am thinking about installing ifx to access the additional registers 
>> etc for Emerald Rapids nodes. What is the current status of ifx 
>> compilation? Can -O2 etc be used or are optimizations (still) broken?
>>
>>
>> -- 
>> Emeritus Professor Laurence Marks (Laurie)
>> Northwestern University
>> Webpage <http://www.numis.northwestern.edu> and Google Scholar link 
>> <http://scholar.google.com/citations?user=zmHhI9gAAAAJ&hl=en>
>> "Research is to see what everybody else has seen, and to think what 
>> nobody else has thought", Albert Szent-Györgyi
>