More details on the ifort deprecation

[Ron_Green]

I have written this blog article with more details about the deprecation of Intel(R) Fortran Compiler Classic (ifort).   I hope this helps with questions and concerns you may have.

[Steve_Lionel]

Hi Ron - good post. I do wonder, though, about the effects on existing scripts, especially third-party, when (I assume) the "ifort" command becomes unavailable. This is likely to cause a surge in support requests and complaints. If you have not already considered this, may I suggest providing a driver named ifort that 1) warns about ifort options not supported by ifx, 2) invokes ifx automatically.

[Ron_Green]

You have opened an old wound.  When we first created this new compiler, the idea was to KEEP the ifort driver and have a compiler option -qnextgen to invoke what is now ifx. Then when this time came, we simply make that compiler option the default, but have another option to invoke ifort, deprecate that option, and remove that option at some point.  So much easier for Cmake, SPACK, build scripts, etc. 

A good engineering solution, we thought.  but alas, up the management and marketing chains they decided for us that they wanted to truly differentiate the past and have a whole other driver for the new compiler.   We fought with technical arguments and we lost.

But sure, I think with the 2025.0 release when we remove ifort from the package, a thin wrapper driver named 'ifort' that prints a removal message and calls ifx is a good idea.  I'll work on this idea with the team.  

[JohnNichols]

You should have used the engineer's trick and offered them money not an argument, or do a short customer survey, ask us and we will agree with you always.