[Bug target/123524] [16 Regression] 6% performance regression in gcc-16 compared to gcc-15 when compiling an interpreter

[liuhongt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123524

--- Comment #10 from Hongtao Liu <liuhongt at gcc dot gnu.org> ---
Created attachment 63586
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=63586&action=edit
Add a heuristic to decide whether X86_CSE_VEC_DUP is profitiable.

The patch can eliminate the vector spill, could you please check if it restore
the performance.