On Fri, 11 Jul 2025, Filip Kastl wrote:
Please let me know if there are additional configurations you would like me
to include in the testing.
Seems like work on smtgcc is going nicely. Good to hear!
Have you considered -Ofast or some subset of the flags it enables? Perhaps
that would uncover a lot of new bugs. I'm not familiar with the flags it
enables enough to tell if that wouldn't be more work than it's worth, though.
Most of the optimizations enabled by -Ofast allow the compiler to optimize
floating-point operations in a way that may change the result. This is bad
for smtgcc, since its definition of a compiler bug is an optimization that
changes the result. :)
There are optimizations in -Ofast that do not change the result, but I'm
not sure it's worth the effort to add them as a new test configuration.
/Krister
