[Bug c++/69564] [5/6 Regression] lto and/or C++ make scimark2 LU slower

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69564

--- Comment #16 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Artem S. Tashkinov from comment #15)
> Is this the same bug?
> 
> https://www.phoronix.com/scan.php?page=article&item=ubuntu-1604-
> compilers&num=2
> 
> In Dense LU Matrix Factorization GCC 5.3.1/6.0 is more than 2 times slower
> than Clang.
> 
> G++ options: -O3 -march=native
> CPU: Xeon E3-1280 v5

Maybe.  Though with -Ofast -march=native -funroll-loops is at least on my box
GCC emitted code faster than clang, without -funroll-loops it is the opposite.