https://reviews.llvm.org/D15890 introduces a transformation to unpack arrays. This unfortunately causes some fundamental compile time regression, so, let's set a limit on the max size of array we consider for unpacking. Fixes https://llvm.org/bugs/show_bug.cgi?id=30608
Details
Details
Diff Detail
Diff Detail
Event Timeline
Comment Actions
Side note: I'm in general not a huge fan of threshold-based optimisations, but I wasn't able to think of a better solution for now.