I am using Fortran for a certified software (DO-178B, DAL C). Rescently the Fortran compiler was changed from 2013 to 2019 (IT issues).
I would like to know the differences between the two compilers:
Intel(R) Visual Fortran Composer XE 2013 SP1 Update 1 Integration for Microsoft Visual Studio* 2013, 14.0.0074.12, Copyright (C) 2002-2013 Intel Corporation
Intel® Parallel Studio XE 2019 Composer Edition for Fortran Windows* Integration for Microsoft Visual Studio* 2017, Version 19.0.0048.15, Copyright © 2002-2018 Intel Corporation. All rights reserved.
If the compiling function did not changed between 2013 and 2019, it will reduced the number of tests.
Over six years, a lot changed in the compilers. New language features, new optimizations, new options. The Release Notes for each version lists feature changes but not bug fixes. You will want to redo all your tests.
I have zero knowledge of the types of tests being discussed here, but should one not expect more, newer, tests to be prepared and run rather than fewer, as computer hardware, operating systems and the critical software itself evolve over the years?
Granted, testing does have costs, but I can imagine that failure of critical software may have costs that are orders of magnitude higher. Boeing 737-MAX comes to mind.
I am curious to know if methods such as those mentioned in this article are being put to use in certifying software for critical applications.