[Bug target/103750] [i386] GCC schedules KMOV instructions that destroys performance in loop

[thiago at kde dot org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103750

--- Comment #5 from Thiago Macieira <thiago at kde dot org> ---
Maybe this is running afoul of GCC's thinking that a simple register-register
move is free? I've seen it save a constant in an opmask register, but kmov{d,q}
is not free like mov{l,q} is.