[Bug middle-end/114989] Compile time hog when building paml

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114989

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #1)
> > callgraph functions expansion      :   3.16 ( 86%)   0.32 ( 52%)   3.50 ( 81%)   111M ( 75%)
> 
> So expand is usually fast. The only slow part I know of that might be
> hitting here is division by constant where we are trying to find if it is
> better to do multiplies or the division directly.

This is not RTL expansion, it's all optimizations after IPA.  I don't see
anything particularly bad in the time-report.