Paul Osmialowski

Regarding architecture dependency, string.h functions in glibc were carefully optimized for AArch64 and this can be observed in the results of the above test case. I can't guarantee the same...

I suspect cgmain changes aren't needed after rebase

These test cases are wrongly designed. With too many CHECK-DAG, the right occurrence of `!llvm.loop` is impossible to be found reliably (namely, a slight change in LLVM middle-end will break...

Interesting idea. I wonder if the on-device SRAM could also be used by FUZIX, there is a plenty of it and only a small chunk of it is used by...

> The test is checking that, with default compiler options, `arr` is allocated to match the dimensions of the array returned by `get_array`. In other words, it checks that `-Mallocatable=03`...

Reverting 778933a6ab43214471fb7729d4b592e38853dd49 makes it possible to build MPICH with the new LLVM flang. Fortunately, this still reverts cleanly.

Anything happens here?

Eventually, we'd contacted the Fortran committee and the response was that there is no undefined behaviour here, quite contrary, the behavior is well-defined as per J3/10-007r1 (Fortran 2008). The subclause...

Hi @bryanpkc assuming the problem manifests itself at runtime, I need to ask, what input data set are you using? (test, ref, or something else). Also, with -mcpu=native being used...

Been testing it all night in our CI, with your flags, with O3 and even with Ofast, and it didn't fail. It's time to do it manually, but bear in...